Where the heat goes.
What decides where energy goes when it enters tissue, why does surgery need eleven kinds of it, and what does it take to keep each one where the surgeon intends, and nowhere else?
One apparatus, two currents.
In the autumn of 1926, at the Peter Bent Brigham Hospital in Boston, the neurosurgeon Harvey Cushing operated on a man with a vascular tumor on his head and abandoned the attempt to remove it, because the tumor bled too heavily. For the next operation, days later, he invited William Bovie to bring his apparatus. Bovie was a research fellow at the Huntington Hospital for Cancer Research, on the Harvard Medical School campus, and had been building electrical devices for bloodless biopsies of cancers. He moved his unit, about the size of a desk, some four blocks to Cushing's operating room. During the operation Bovie, who knew little of surgery, adjusted the knobs that set the current and the waveform. The unit delivered high-frequency current through a cutting loop, and with it Cushing removed the tumor.
Bovie described the apparatus in a patent filed in 1928. By then he had moved to Evanston, Illinois, and the patent may record the design of that year rather than the unit of 1926. In it, one apparatus made two currents, both from spark gaps, small air gaps across which each spark set off a burst of high-frequency oscillation. The first came from a transformer of about 1,000 V. It was the less damped and steadier of the two, and it severed tissue while closing the capillaries it cut. The second came from a transformer of about 4,000 V and was heavily damped: each burst of oscillation died away quickly. The patent described it as a current that "readily dries up tissue", used "to stop the flow of blood from large capillaries".
In 2024 Li and Isaacson published a history of the collaboration, and it recorded two hazards of the early apparatus. The original unit lacked an effective return electrode, the contact through which current leaves the patient on its way back to the generator. Cushing also once received a shock that passed to his arm through a metal retractor, an instrument that holds tissue aside; Li and Isaacson took that incident from a 2013 account of Bovie's life.
Cushing wrote in 1928 that when he first saw the loop he foresaw a new tool for surgery. Its first years taught two things. In Bovie's patent, the two currents showed that the form of the current decided what one kind of energy did to tissue. In the operating room, the current went wherever the circuit allowed, once through a retractor into the surgeon's arm. Surgery now uses eleven kinds of energy, and each raises the questions that Bovie's apparatus raised: what the energy does where it enters tissue, which task it suits, and what keeps it in the place the surgeon intends.
Before we start.
This guide puts those questions to all eleven energies at once: What decides where energy goes when it enters tissue, why does surgery need eleven kinds of it, and what does it take to keep each one where the surgeon intends, and nowhere else?
The eleven are monopolar radiofrequency (RF) electrosurgery, conventional bipolar RF, advanced bipolar vessel sealing, ultrasonic dissection, surgical lasers, RF ablation, microwave ablation, cryoablation, pulsed electric fields, thermal focused ultrasound and histotripsy. Pulsed electric fields take two forms, irreversible electroporation for tumors and pulsed field ablation for the heart. Hybrid instruments combine two of the eleven in one jaw. The first five divide and coagulate tissue during an operation. The other six are kinds of ablation, which destroys tissue in place without removing it: usually a tumor, or lines of heart muscle whose scar treats an irregular heart rhythm. Ablation reaches its target through a needle, a catheter or the intact skin.
Surgeons and engineers use several common words in a narrow sense, and the text keeps to it. To cut with current or with a laser is to vaporize a thin line of cells so quickly that the tissue parts; no sharp edge is involved. To coagulate tissue is to heat it until its proteins unfold and set, which kills its cells and stops small vessels bleeding; the word refers to the tissue, and means more than the clotting of blood. A seal is a vessel closed by heat and pressure, its walls fused so that it stays shut when it is divided.
Other words belong to the equipment and its rules. A generator's setting is a power, the rate at which it delivers energy, and the energy the tissue receives depends also on the tissue, the contact and the time. A generator's load is the tissue as the generator sees it, an electrical impedance that changes as the tissue heats and dries. Impedance is the opposition a circuit offers to an alternating current, and it is measured in ohms, like a resistance. The pad is the return electrode of monopolar surgery, the large electrode on the patient's skin through which current leaves the body. A device is cleared in the United States when the Food and Drug Administration finds it substantially equivalent to a device already legally on the market (Chapter 26).
Physics settles where the energy goes. Heat injures tissue according to temperature and time together: ten degrees can change the time needed for the same injury a thousandfold, and at 100 °C the water in cells boils (Chapter 1). Most of the eleven energies end as heat in tissue, and they differ in where they deposit it and how fast. Current heats where it is most concentrated, at the tip of a small electrode or between the jaws of an instrument. A vibrating blade heats tissue at its own face. Light heats a layer whose depth the wavelength sets, from about a micrometer to millimeters. A microwave field heats a volume around an antenna, focused sound heats a small region deep inside the body, and a cryoprobe takes heat away.
How fast matters because heat moves slowly. It spreads through tissue by diffusion, over a distance that grows with the square root of time: about 0.75 mm in one second and about 2.4 mm in ten (Chapter 2). Holding an exposure a hundred times longer carries the heat only ten times as far. Energy delivered faster than heat can spread stays close to where it was put, and energy delivered slowly spreads with it, carrying the injury outward.
Two of the eleven kill cells without relying on heat. Pulsed electric fields open pores in cell membranes that the cells cannot repair, and histotripsy breaks cells apart with bubbles that form and collapse in a focused sound field. Heat destroys the collagen framework of vessels and ducts along with the cells around them, and reviews and preclinical studies report that both of these energies largely spare that framework (Chapters 15 and 16). Each surgical task needs its own geometry and its own time scale, from a line of cells vaporized at the edge of an electrode to a volume destroyed deep inside an organ. No single energy provides every combination of the two, and some tasks also need the framework spared. Together these two reasons explain why surgery uses eleven (Chapter 6).
Keeping each energy where the surgeon intends is a matter of engineering, and then of proof. A generator must hold its output across a hundredfold range of load (Chapter 10). It must also stay safe when any single part of it fails (Chapter 20). Current must leave the patient through the pad or through the instrument's own return path and through nothing else. Cushing's shock through a retractor was current on a path that nobody had designed. Such a path can also be hidden. In laparoscopic surgery the surgeon works through small incisions and watches the tissue on a camera's image, so an instrument's insulation can fail out of the camera's view (Chapter 18). Other hazards come from what the energy releases. Its smoke, sparks, light and electrical interference can contaminate the air, start a fire, harm eyes or disturb monitors and implants (Chapter 19).
The generator, its instruments, its cable and its pad form a system whose parts different people design, buy, connect and clean (Chapter 21). Each requirement on that system has to be traced to a test that shows it is met (Chapter 22). The maker must then prove all of this, against standards and to regulators, before the device may be sold.
How to read this guide
The guide's 27 chapters follow nine parts. Part I builds the physics of energy in tissue, and Part II sets out what open, laparoscopic and robotic access change for every energy. Parts III to V take the energies in turn: Part III electrosurgery, Part IV the ultrasonic instruments and lasers, and Part V ablation. The instruments and lasers of Part IV pass no current through the patient, and the part also covers the hybrid instruments that add bipolar current to an ultrasonic blade. Part VI follows one laparoscopic colectomy step by step; a colectomy is an operation to remove part of the colon. Part VII turns to the system around the energy: the burns, smoke and fire it can cause, the generator built to fail safe, the interfaces between parts, and the path from a requirement to its evidence. The last two parts are dated October 2026. Part VIII describes who makes surgical energy devices and what is changing. Part IX sets out the standards, the compliance tests and the regulation in the United States, the European Union and India, and it reads the public test record of one generator cleared by the US regulator in 2025.
The chapters answer the three parts of the central question in turn. Chapters 1 to 5 answer where energy goes, and Chapter 6 why surgery needs eleven kinds of it. Parts III to V, VII and IX answer what keeps each one in its place, and Chapter 17 shows several of the energies at work in one operation. Chapter 27's closing pages draw the three answers together. A reader on a generator project can start with Chapters 3, 8, 10, 20, 21 and 25, and one working on ablation with Chapters 2, 13 to 16 and 24.
Each chapter opens with the question it answers, and plates are numbered separately so that any one can be cited on its own. Four kinds of box recur:
Blue edge. Carries the structural point of a section, or the chapter's solved example.
Orange edge. Names a common misreading, a trap, or the limit of a claim.
Green edge. Maps the idea onto the reader's own work: in instrument and generator design, at the test bench, or in a regulatory file.
- Each chapter closes with what it established, in four lines.
The mathematics stays at Ohm's law, power, ratios, percentages, squares and square roots, and powers of ten. Where a number is derived, the arithmetic is shown once with real values and the assumptions are stated. Generator circuits are drawn as block diagrams, and a plate that is schematic rather than drawn from data says so.
Inside the plates, color is a legend and never decoration. Electrical current and fields, sound and vibration, light, heat and its effects, and cold each have one color throughout the guide. The measured readout is always blue, and one orange mark in each plate points at the detail that matters most. A key strip under every plate lists only the colors that plate uses.
This is a reference for understanding, not a surgical technique, an engineering specification, a regulatory opinion or clinical advice. It does not cover cost, schedule, make-or-buy decisions or which energy a product should use. Regulatory status, standards editions and product details are stated as of October 2026 and will date. Where a figure is a manufacturer's claim or comes from a manufacturer-authored study, the text says so. Named products are examples of a principle, not recommendations.
+ Part I · Physics
The physics of energy in tissue.
A gram of soft tissue is about three quarters water, and boiling that water off takes about seven and a half times the energy needed to heat the gram from body temperature to 100 °C. Three facts of surgery follow from that one ratio: wet tissue stalls near 100 °C, energy delivered fast into a thin layer cuts, and only tissue that has dried out chars. This part builds the physics that the rest of the book uses. It shows how temperature and time together decide what happens to a cell, how heat spreads and what carries it away, why tissue conducts current and why surgical current must alternate hundreds of thousands of times a second, and what sound and light do when they enter the body. Its last chapter sets the six ways of delivering energy side by side, and nothing in the part assumes more than school physics.
Degrees and seconds.
+ The questionWhy does the same tissue bleed, seal, char or vaporize depending on a few degrees and a few seconds?
Heat unfolds proteins
About a tenth of a second at 60 °C counts as the same dose of heat as four hours at 43 °C, on the rule that researchers use to measure such doses in cancer treatment. The explanation starts with what temperature does to proteins. Temperature is a measure of how violently molecules move, and in tissue at 37 °C that motion is gentle enough for proteins to keep their shapes, which matters because a protein's shape is its function. An enzyme works because its folded chain puts a few chemical groups in exactly the right places, and a collagen fiber is strong because three chains wind around each other like the strands of a rope.
Those shapes are held by a great many weak interactions, among them hydrogen bonds between neighboring parts of the chain and the clustering of oily side chains away from water, while much stronger covalent bonds hold the backbone of each chain. As tissue warms, the added motion shakes the weak bonds loose first, so the protein unfolds while its backbone survives. The unfolded chains then tangle with their neighbors, which is why a heated egg white turns from clear to opaque and does not turn back. In tissue the same change is called coagulation: the tissue blanches, stiffens and stops working, and its cells are dead even if they are still in place. To a surgeon the word means this change caused by heat. The clotting by which blood closes a small wound on its own is a different process.
Collagen matters more than any other protein in this guide, because it gives vessel walls, skin and tendons their strength and behaves in a characteristic way when heated. In workshop notes of 2015 on modeling irreversible thermal damage, John Pearce summarizes measurements made on rat tissues. In those measurements the triple helix unwinds over a band of temperatures between about 60 and 90 °C, depending on the tissue and the test, and the fiber shrinks as it unwinds. If two layers of unwound collagen are pressed together while they cool, their strands entangle across the boundary and the layers fuse. That fusion is how a vessel seal works, and Chapter 9 builds on it.
Damage is a rate, not a threshold
It is tempting to picture a temperature at which tissue dies, but heat damage does not work like a switch. Unfolding a protein is a chemical reaction, and like most reactions it runs at a rate that climbs steeply with temperature. The standard description is the Arrhenius model, in which the rate of damage rises exponentially with temperature and the damage accumulates for as long as the temperature is held. The same injury can therefore come from a lower temperature held for a long time or from a higher one held briefly.
Hyperthermia is the treatment of cancer with mild heat, and researchers in the field express this trade between temperature and time as a thermal dose called CEM43, the number of cumulative equivalent minutes at 43 °C. On this measure each degree above 43 °C halves the time needed for the same effect. A threshold often used for tissue damage is 240 equivalent minutes. In a 2011 review of thresholds for thermal damage in normal tissues, Yarmolenko and colleagues report that in the canine prostate that dose produced complete coagulation. The threshold differs from one tissue to another, but the halving alone shows how fast the time shortens as the temperature rises.
A dose of 240 equivalent minutes takes 240 minutes at 43 °C, and each degree above 43 °C halves the time. At 45 °C the two degrees are two halvings, so the time falls to a quarter, 60 minutes, and at 47 °C four halvings bring it to a sixteenth, 15 minutes. At 50 °C seven halvings divide 240 minutes by 128, which leaves about 1.9 minutes. At 60 °C seventeen halvings divide it by 131,072, which leaves about 0.11 seconds. The ten degrees between 50 and 60 °C are ten halvings, a factor of 1,024, so on the CEM43 line ten degrees shorten the time for the same damage about a thousandfold.
The halving rule describes some tissues well and others poorly. Fitted to skin, the Arrhenius model predicts that each degree roughly halves the time, as CEM43 does. Fitted to liver, the model has a shallower slope, which means that each degree speeds the damage by only 25 to 40 %. Pearce's notes also list a data set for pig liver from the work of Steven Jacques, and with that set the model predicts about 17 minutes at 50 °C, about 47 seconds at 60 °C and about 2.6 seconds at 70 °C. With a skin data set that the same notes recommend, the model predicts about 5.5 minutes at 50 °C. The two predictions differ about threefold at 50 °C and diverge further above it, which shows that the tissue itself changes the rate.
For the rest of this guide, two consequences matter. First, a device that controls temperature within a few degrees controls its effect within a factor of several, so small errors in temperature are large errors in dose. Second, the familiar rule that "proteins coagulate almost instantly at 60 °C" is only approximately true, because at that temperature the same damage takes a fraction of a second in one tissue and tens of seconds in another.
A statement that tissue dies at a given temperature leaves out how long the temperature was held. Published values for the onset of coagulation range from about 50 to 60 °C, because they come from different tissues, species and exposure times. A fair comparison of two devices therefore sets temperature and time side by side, or better still compares the measured zone of damage.
Water turns 100 °C into a wall
Soft tissue is mostly water, about 77 % in human liver, and water has an unusual property that shapes almost everything electrosurgery does: raising its temperature takes a moderate amount of energy, but turning it into steam takes a great deal more. One gram of tissue taken as 75 % water, a round figure close to liver's, has a specific heat of about 3.6 joules per gram per degree, so heating it by 63 degrees, from 37 to 100 °C, takes about 227 J. Each gram of water needs 2,257 J to turn to steam, so boiling off the tissue's 0.75 g of water then takes about 1,693 J, about seven and a half times as much. For pure water the ratio is 8.6.
That ratio has three consequences. The first is that wet tissue holds its own temperature down. While liquid water remains, energy goes into boiling it rather than into raising the temperature, so the temperature stalls near 100 °C. A generator that keeps tissue wet therefore keeps it near 100 °C no matter how much power it delivers, and the extra energy only boils the water faster.
The second consequence is cutting. If energy arrives in a thin layer faster than steam can escape, the water inside each cell flashes to vapor and the cell bursts. This explosive boiling is vaporization, and in electrosurgery it begins once the water in the cells reaches about 100 °C. That is the temperature of the stall in the first consequence, so what separates the two is how fast the energy arrives. A moving line of bursting cells is an incision, which is what surgeons mean by cutting with energy, and Chapter 8 shows how an electrode makes one without being sharp. If the same energy arrives slowly, the water leaves gradually instead, and the tissue dries and shrinks into a firm coagulum. This slow drying is desiccation, and it is how most contact coagulation works.
The third consequence appears when the water is gone. Nothing then holds the temperature down, so dry tissue heats quickly and begins to char, a change called carbonization. In a 2005 article on the physical bases of electrosurgery, Dargent and Schettini put its onset between about 150 and 200 °C once the tissue has dried out, well above the 100 °C at which wet tissue stalls. Char is almost always unwanted. It wastes energy, releases smoke and conducts current poorly, which stalls any further heating by current (Chapter 13).
Every thermal injury has the same anatomy
Because temperature falls with distance from the source, every thermal device leaves the same nested pattern in tissue. Nearest the source the hottest tissue has vaporized and is gone, and around the crater lies a thin layer of char, where the tissue dried and then carbonized. Next comes a desiccated band, heated to about 90 to 100 °C, where the water was driven off. Outside that lies the coagulated zone, where proteins and collagen have denatured and the cells are dead. It spans from about 90 °C down to 50 to 60 °C at its edge, depending on how long the heat was held. The outermost band reached only 43 to 50 °C, and its cells may recover or may die over the following hours.
The devices in this guide differ mainly in how thick they make each band: a short laser pulse can leave a crater with almost no coagulated tissue beneath it, while a slow contact coagulation leaves a thick coagulated band and no crater at all. A vessel sealer aims for a fused band of coagulated collagen and as little else as possible.
The boundary that matters clinically is the outer edge of the coagulated zone, the margin of irreversible injury, and it is not reliably visible at the time of surgery. Coagulated tissue can look almost normal when the instrument is withdrawn, then die and weaken over the following days. That delay is why a thermal injury to bowel can be missed during an operation and present days later, a pattern Chapter 18 returns to.
Four effects, one map
Each task that surgeons ask of energy has its own place on a map of temperature against time. Coagulation stops bleeding from small vessels by heating tissue to 60 to 100 °C for seconds, which shrinks the vessel walls and clots the blood. Sealing does the same to a vessel held closed under pressure, so that its collagen fuses into a band strong enough to replace a suture. Cutting delivers energy so quickly into a thin line that the water boils explosively and the tissue parts. Ablation kills a volume of tissue in place, usually a tumor. Thermal ablation either holds the tissue above about 60 °C for seconds to minutes or freezes it, while pulsed electric fields and histotripsy kill it without relying on heat and so have no place on this map (Chapters 15 and 16).
Char belongs to none of the four, because it marks energy delivered after the water had gone, and a well-controlled device avoids it. The difference between a clean cut and a charred one, or between a seal and a burn, is almost always a difference in how fast the energy arrived and when it stopped.
A specification such as "the jaw reaches 90 °C" states only a temperature, so it says little about the effect. State temperature and time together, or specify the effect directly: the width of the coagulated zone, the pressure at which a seal bursts, the depth of an incision. FDA's guidance for electrosurgical devices recommends the same approach: measuring the thermal damage zone by histology, in stained sections of tissue examined under magnification, and providing graphs of temperature against time, rather than inferring the zone from a temperature alone.
- Heat unfolds proteins by shaking loose their weak bonds, and collagen unwinds and shrinks between about 60 and 90 °C and fuses when pressed together as it cools.
- Thermal damage accumulates at a rate that climbs steeply with temperature: above 43 °C each degree roughly halves the time for the same injury, though liver is less steep than skin.
- Boiling off the water in tissue takes about seven and a half times the energy of heating it from 37 to 100 °C, so wet tissue stalls near 100 °C, fast heating cuts, slow heating desiccates, and only dry tissue chars.
- Every thermal injury leaves the same nested zones, and the margin that matters clinically is the outer edge of the coagulated zone, which is not reliably visible during surgery.
How heat moves.
+ The questionIf energy is deposited in one small place, why does the damage spread beyond it, and what carries the heat away?
Heat flows downhill, slowly
The damage an energy device causes can go on growing for a few seconds after activation ends, although no more energy is going in. Chapter 1 treated each point of tissue as if it kept whatever heat it received, and tissue does not. Heat flows from hotter tissue to cooler tissue, at a rate set by the temperature difference and by the tissue's thermal conductivity. Soft tissue conducts heat about as well as water does, about 0.5 watts per meter per degree for liver against 0.6 for water at room temperature, which makes tissue a poor conductor. Copper conducts heat about 800 times better.
How fast a change of temperature spreads depends on more than conductivity. It depends on the conductivity divided by the amount of heat that each unit of volume can store, and that ratio is the thermal diffusivity. For water and most soft tissues it is about 0.14 square millimeters per second. The number looks abstract, but it answers the practical question of how far heat travels in a given time.
The square-root rule
A pulse of heat deposited at a point does not travel outward at a steady speed, the way sound does. It diffuses, and diffusion covers distance in proportion to the square root of time, so a convenient estimate of how far heat has spread after a given time is the diffusion length: the square root of four times the diffusivity times the time.
The diffusivity of soft tissue is taken as 0.14 square millimeters per second. After a tenth of a second, four times the diffusivity times the time is 0.056 square millimeters, and its square root is about 0.24 mm. After one second the product is 0.56 square millimeters and the distance about 0.75 mm, and after ten seconds the product is 5.6 and the distance about 2.4 mm. After a minute the product is 33.6 and the distance about 5.8 mm. From a tenth of a second to ten seconds the time grows a hundredfold, but the distance grows only tenfold, because ten is the square root of a hundred.
This one relationship explains a great deal of surgical practice. Energy delivered in a fraction of a second stays within a fraction of a millimeter of where it was put, because heat has had no time to move, while energy delivered over ten seconds has had time to reach tissue two or three millimeters away. To make a narrow effect, such as a clean incision or a precise laser crater, a device must therefore deliver its energy quickly. To make a broad effect, such as coagulating a wide area or ablating a tumor, it can take its time and let conduction do part of the work.
The same rule sets how far a vessel seal's damage spreads. In a 2025 study, Chambers and colleagues measured the seals made by Medtronic's current vessel sealers in excised and in living tissue. All three authors were Medtronic employees, and Medtronic funded the work, so the figures are a manufacturer's own. They report seal times of about 2 to 4 seconds, and over that time the square-root rule gives a diffusion length of roughly one to one and a half millimeters. Measured by histology in living tissue, the thermal spread beyond the jaws was about 1.3 to 2.1 mm, the same order as the rule predicts (Chapter 9).
Heat also keeps moving after the energy stops, which is why the damage can grow after activation ends. A hot jaw or blade goes on conducting heat into whatever it touches, and a heated zone goes on spreading until its temperature falls. Chapter 11 shows why this matters most for ultrasonic instruments.
Blood carries heat away
Conduction is not the only way heat leaves a heated zone. Capillaries run through living tissue, and blood entering them at 37 °C picks up heat and carries it away. The standard description is the bioheat equation that Harry Pennes published in 1948, and in words it is a simple heat balance: the temperature of a small volume of tissue rises by the heat the source deposits in it, minus the heat conducted to its neighbors, minus the heat carried off by blood. Metabolic heat is also in the equation, but during surgery it is negligible.
The blood term varies more than any other. Physiologists call the flow of blood through a tissue its perfusion. Resting skeletal muscle receives about 4 milliliters of blood per minute for every 100 grams of tissue, the liver about 95 and the kidney about 400. At the same power, the same device therefore heats kidney less than muscle, and any model or bench test that ignores perfusion will overestimate heating in well-perfused organs. FDA's guidance for electrosurgical devices recommends thermal testing in several tissue types for a device meant for soft tissue in general, and perfusion is one more reason that tests in dead, unperfused tissue need careful interpretation.
Heating also changes the balance as it proceeds. Inside a coagulated zone the blood in the capillaries clots and stops flowing, so the cooling term vanishes there and heating accelerates. The tissue's properties shift too. Its electrical conductivity rises by about 2 % for every degree, which matters for any energy delivered as current, and its thermal properties change as it dries. Berjano's 2006 review of models of RF ablation notes that the 2 % figure comes from measurements made below 42 °C.
The heat sink
Capillaries cool tissue a little everywhere, but a large vessel cools it a lot in one place. Blood flowing through a vein several millimeters wide holds the vessel wall close to 37 °C against all but intense heating, and the tissue immediately around it stays cooler than tissue farther away. Surgeons and engineers call this the heat sink effect.
The effect is strongest in ablation, where the goal is to kill every cell in a volume. Lu and colleagues at the University of California, Los Angeles, measured it in 2002, in an animal study of RF ablation in the livers of 10 pigs. Of the veins wider than 3 mm, 73 % deformed the ablation zone, while none of the smaller veins did, and viable cells remained around every vein wider than 5 mm. In 2003 Lu and colleagues reported a clinical series of 105 liver tumors treated with RF ablation. Treatment was incomplete, or the tumor recurred, in 48 % of the tumors that touched a vessel of 3 mm or more, against 7 % of those that did not. Both studies divided the vessels at 3 mm, so together they support the "3 mm rule" of ablation practice, that a vessel 3 mm or more across beside a tumor acts as a heat sink.
The heat sink can also help. It protects large vessels that run near a heated zone, which is one reason tumors next to major vessels can be ablated at all. It is also one more reason a vessel sealer clamps its vessel shut, because with the flow stopped inside the jaws the vessel's own blood can no longer carry the heat away.
Most bench tests of thermal devices use excised tissue with no blood flow. Such a test shows the zone a device makes when no blood carries heat away, which is usually larger than the zone in living tissue, and it shows nothing of what happens next to a perfused vessel. Chapter 13 returns to the heat sink in ablation, where microwave energy reduces it but does not remove it.
Lateral thermal spread
The practical measure of all of this is lateral thermal spread: how far thermal damage extends beyond the tissue the device was meant to affect. It is measured after the fact, by cutting the tissue, staining it and measuring the damaged zone under magnification. A device's indication is the use its maker claims for it, and a general soft-tissue indication covers soft tissue in general rather than one procedure or organ. For a device with such an indication, FDA's electrosurgical guidance recommends exactly this measurement: the length, width and depth of the thermal damage zone in at least three tissues, such as liver, kidney and muscle, measured in triplicate at the minimum, default and maximum settings.
Spread depends on everything in this chapter: how much energy the device delivers, how fast, into which tissue, with how much blood flow, and how long the device and the tissue stay hot afterward. It is the single number that best summarizes whether a device put its effect where the surgeon intended, and Chapters 9, 17 and 22 return to it for vessel sealers.
When comparing figures for thermal spread, check the conditions under which each was measured. Spread measured in excised tissue after a short activation says little about a long activation in perfused tissue. The comparisons that mean something use the same tissue, the same vessel sizes, the same settings and activation times, and histology rather than surface appearance.
- Soft tissue conducts heat about as well as water, and its thermal diffusivity is about 0.14 square millimeters per second.
- Heat spreads a distance that grows with the square root of time, about 0.24 mm in a tenth of a second, 0.75 mm in one second and 2.4 mm in ten, so energy delivered quickly stays close to where it was put.
- Blood carries heat away, perfusion varies about a hundredfold between resting muscle and kidney, and large vessels act as heat sinks.
- Lateral thermal spread is measured by histology and is the practical summary of where a device's heat actually went.
Tissue as a conductor.
+ The questionWhy does current heat tissue at all, and why must surgical current alternate hundreds of thousands of times a second?
Tissue conducts with ions
In 1891 the French physiologist Jacques-Arsène d'Arsonval found that the response of muscle to alternating current faded as he raised the frequency and had all but disappeared at about 10 kHz, the highest frequency his equipment could reach. The reason lies in how tissue carries current, which also explains why current heats tissue at all. A metal conducts because some of its electrons are free to move, but tissue has no free electrons. It conducts because it is salt water: sodium, potassium and chloride ions dissolved in the fluid inside and between cells carry charge when a voltage pushes them. Anything that changes how many ions there are, or how freely they can move, therefore changes how well a tissue conducts.
Conductivity measures how readily a material carries current, in siemens per meter, and it differs widely between tissues. Gabriel and colleagues published parametric models of the electrical properties of tissues in 1996, and at 500 kHz, a typical electrosurgical frequency, those models give blood a conductivity of about 0.75 siemens per meter, muscle 0.45, kidney 0.23 and liver 0.15. Fat holds little water and conducts at about 0.025, eighteen times less than muscle, and dry skin at about 0.004. Physiological saline at body temperature conducts at about 1.9, roughly thirteen times better than liver. Measurements on living liver are reported to be higher than those on excised tissue, so a fair range for liver is 0.1 to 0.4 S/m, depending on the source and the condition of the tissue.
These values explain where current goes inside a patient, because current offered a choice of paths divides in proportion to how well each path conducts, so it crowds into wet, lean tissue, blood and saline, and avoids fat and dry skin. A pool of saline irrigation around an electrode can carry current away from the tissue it was meant to heat.
How current becomes heat
Current heats any resistor it flows through, and school physics gives the power as the square of the current times the resistance, or equally as the square of the voltage divided by the resistance. Inside tissue it is more useful to think about each small volume and the current density in it, which is the current crossing each unit of area. The heat released in each cubic millimeter is the square of the local current density divided by the local conductivity. That is Joule heating, and it is how every RF device in this guide heats tissue.
Joule heating has a consequence that looks like a contradiction: for a given current density a poor conductor heats more, while for a given electric field a good conductor heats more. Which view applies depends on whether the current or the voltage is held fixed, and Chapter 10 shows that a generator switches between the two as the tissue changes.
Current density decides where the heat appears
The single most important idea in electrosurgery is that heating follows current density, and current density depends on geometry. Current leaving a small electrode must spread out through an ever larger area of tissue. Around a small sphere, the area the current crosses grows with the square of the distance from the center, so the current density falls with the square of the distance. Heating follows the square of the current density, so it falls with the fourth power of the distance: tissue twice as far away heats sixteen times more slowly.
A current of 0.5 A flows out of a small spherical electrode into muscle, whose conductivity is taken as 0.4 siemens per meter. At 1 mm from the center the current crosses the surface of a sphere 1 mm in radius, about 12.6 square millimeters, so the current density is about 4 A per square centimeter. The square of that density divided by the conductivity gives a heating power of about 4 kW in every cubic centimeter. Tissue needs about 3.8 J to warm one cubic centimeter by one degree, so before any heat conducts away the tissue there heats at about 1,000 °C per second. At 2 mm the sphere's area is four times larger, the current density a quarter and the heating a sixteenth, about 65 °C per second. At 4 mm it is about 4 °C per second, and at 10 mm about 0.1 °C per second.
The fall-off is so steep that heat appears only where current is concentrated. At the tip of an electrosurgical pencil, a fraction of a square millimeter carries the whole current, and tissue there boils in milliseconds. The same current leaves the patient through a return pad, a large electrode placed on the skin to carry it back to the generator. Spread over the pad's tens of square centimeters, the current warms the skin by a few degrees. Nothing in the circuit is different except area.
The same principle explains most of the burns that Chapter 18 describes. Any place where current is forced through a small area acts as an electrode: the edge of a return pad that is peeling, a pinhole in an instrument's insulation, the point where a shaft touches bowel. The current heats each of these small areas in the same way, whether or not the surgeon intended it to pass there.
Cells are tiny capacitors
So far tissue has been treated as a uniform conductor, but on the scale of cells it is not one, because each cell is wrapped in a membrane only a few nanometers thick that is an excellent electrical insulator, with conducting fluid on both sides. A thin insulator between two conductors is a capacitor, and so each cell membrane behaves as one.
A capacitor blocks steady current but passes alternating current, and it passes more the faster the current alternates. At low frequency the membranes block current, which then flows only through the fluid between cells. As the frequency rises, the membranes pass more and more of it, and current flows through the cells as well as around them. Gabriel and colleagues' models show the change directly. Between 100 kHz and 1 MHz, muscle's conductivity rises from about 0.36 to 0.50 S/m, because more of the current passes through the cells. Permittivity measures how much charge the membranes store, as a multiple of what empty space would store in the same field, and over the same range muscle's permittivity falls from about 8,000 to 1,800.
Why the current must alternate so fast
The capacitor picture also explains why surgical current alternates at hundreds of kilohertz. A nerve or muscle cell fires when the voltage across its membrane changes by enough, for long enough. Direct current pushes one way all the time, and mains current at 50 or 60 Hz pushes one way for milliseconds at a time, which in both cases is long enough to charge membranes fully and to fire nerves and contract muscles, including the heart.
At a higher frequency, each half-cycle becomes shorter than the time a membrane needs to charge and a nerve needs to respond: nerve cells need a stimulus lasting about 0.1 to 10 milliseconds, while at 400 kHz one half-cycle lasts 1.25 microseconds. The current still flows and still heats, but it no longer stimulates.
That is the fading that d'Arsonval saw as he raised the frequency. When he applied high-frequency current to himself at a high level, he felt nothing but warmth, and at a demonstration in July 1893 a commission recorded about 0.8 A passing through its members' arms. Modern generators work far above the 10 kHz his equipment reached. IEC 60601-2-2 is the safety standard for electrosurgical equipment, and it defines the high-frequency range of such generators as below 5 MHz and generally above 200 kHz. Most of them run between 300 and 500 kHz, near the low end of that range, for a reason given at the end of this section.
When an electrode arcs, the arc conducts better in one direction than the other and partly rectifies the current, which creates a small low-frequency component that can stimulate muscle. Generators place capacitors in series with their outputs to block direct and low-frequency current, and these capacitors limit the effect but do not remove it entirely (Chapter 8).
There is also a reason not to go much higher. The current that flows through a capacitor rises in proportion to frequency, and every cable, instrument shaft and patient forms small, unintended capacitors with whatever lies near it, which engineers call stray capacitance. At higher frequencies more current leaks away through those unintended paths, an effect Chapter 18 quantifies. The working band of a few hundred kilohertz is therefore a compromise between stimulation at the low end and stray coupling at the high end.
Ask two questions of any RF device. First, where is the current density highest, and is that where the heating is wanted? Second, what else in the circuit offers current a small area or a capacitive path, including the cables, the instrument shafts and the patient's contact with the table? Both questions appear in the risk analysis that every electrosurgical device must carry, though in different words.
- Tissue conducts through dissolved ions, and at 500 kHz its conductivity ranges from about 0.75 S/m for blood down to 0.025 S/m for fat and less for dry skin.
- The heat released in each small volume of tissue is the square of the current density divided by the conductivity, and near a small electrode it falls sixteenfold each time the distance doubles.
- Cell membranes act as capacitors, so at high frequency current passes through the cells as well as around them, and each half-cycle is too brief to fire a nerve.
- IEC 60601-2-2 defines the high-frequency range of surgical generators as below 5 MHz and generally above 200 kHz, and most run at 300 to 500 kHz, a compromise between nerve stimulation at the low end and stray coupling at the high end.
Sound in tissue.
+ The questionCurrent heats tissue because it pushes ions through it. How can a vibration heat tissue, cut it or tear it apart?
Sound is a traveling pressure wave
The tip of an ultrasonic instrument is a metal blade that vibrates along its length tens of thousands of times a second, and the tissue pressed against it heats and can part. Where the current of Chapter 3 moves ions through tissue, a vibration moves the tissue itself, because sound is a mechanical wave. A vibrating surface pushes on the material next to it and compresses it, and the compressed region pushes on the next in turn. The disturbance travels outward as a wave of alternating compression and rarefaction, regions where the pressure rises above and falls below its resting value, while each bit of material only moves back and forth about its place.
In soft tissue the wave travels at about 1,540 meters per second, slightly faster than in water at 1,481 and far faster than in air at 343, both at room temperature. The wavelength is the speed divided by the frequency. At 1 MHz, a typical therapeutic frequency, the wavelength in tissue is 1.54 mm, and at 4 MHz it is 0.39 mm. These lengths matter, because both the finest detail a beam of sound can resolve and the smallest spot it can be focused to are of the order of one wavelength.
Surgery uses sound in two quite different regimes. Ultrasonic instruments vibrate a metal tip at 23 to 55.5 kHz, and they act on the tissue that touches the tip: the tip rubs, shears and churns that tissue at its face. Focused ultrasound works at 0.7 to 4 MHz and sends a wave through intact skin and tissue to act at a point centimeters deep. The first regime depends on friction and cavitation at a surface, while the second depends on how waves travel, reflect and are absorbed. This chapter builds both.
What happens at a boundary
Each material has an acoustic impedance, its density multiplied by its speed of sound, and sound crosses a boundary well only when the impedances on the two sides match. Acoustic impedance is measured in rayls, kilograms per square meter per second, so water, at about 1,000 kilograms per cubic meter and 1,481 m/s, has about 1.48 million rayls, or megarayls. Soft tissue has about 1.63 megarayls, a close match to water. Cortical bone is about 7, more than four times the value for tissue, though reported values vary widely. Air is about 0.0004, four thousand times lower than tissue. The fraction of the sound's intensity reflected at a boundary is the square of a ratio: the difference between the two impedances divided by their sum. Between tissue and water the fraction is 0.2 %, and between tissue and bone it is about 39 %. Between tissue and air it is 99.9 %, so essentially all the sound bounces back.
Three practical consequences follow. A transducer turns an electrical drive into sound, and it must be coupled to the skin with gel or water, because even a thin layer of air would reflect almost all of its energy. A beam aimed through gas in the bowel or lung is blocked, and a beam aimed through bone is partly reflected and partly absorbed, so it heats the bone. Focused ultrasound therefore needs an acoustic window, a path to its target free of gas and bone. Inside an ultrasonic instrument, the vibration stays in the metal waveguide that carries it to the tip, because the gas around the waveguide is so poorly matched to the metal that almost nothing leaks out.
Absorption turns sound into heat
As sound travels through tissue, it loses some of its energy in every cycle to internal friction between moving molecules and structures, and the lost energy appears as heat, while a little more of it is scattered out of the beam. Together, absorption and scattering attenuate sound in soft tissue by about 0.54 decibels per centimeter for every megahertz of frequency, and absorption accounts for most of the loss. A loss of 3 dB halves the intensity, so at 1 MHz the intensity halves in about 5.6 cm of tissue, and at 4 MHz in about 1.4 cm.
The attenuation and the wavelength together set the central trade of focused ultrasound. A lower frequency reaches deeper, but its longer wavelength focuses to a larger spot, while a higher frequency focuses more finely but loses its energy sooner. A system that treats the prostate from inside the rectum works a few centimeters from its target and can use 4 MHz, while the histotripsy system that treats liver tumors through the abdominal wall uses 700 kHz. Chapter 16 describes both systems.
A target lies 5.6 cm deep, the depth at which a 1 MHz beam has lost half its intensity. Tissue attenuates sound by about 0.54 dB per centimeter for every megahertz, so at 4 MHz the loss is four times as fast, about 2.2 dB per centimeter, and the intensity halves every 1.4 cm. Over the 5.6 cm to the target, the 4 MHz beam therefore halves four times, to a half, a quarter, an eighth and finally a sixteenth of its starting intensity. That leaves about 6 % of the 4 MHz beam at the target, against about 50 % of the 1 MHz beam. The 4 MHz wavelength in tissue is 0.39 mm against 1.54 mm, a quarter as long. At this depth the 4 MHz beam can therefore focus to a spot about a quarter as wide. Since 6 % is an eighth of 50 %, an eighth as much sound arrives there as at 1 MHz, but a spot a quarter as wide has a sixteenth of the area, so the intensity at the focus is about twice as high. At 5.6 cm the finer focus therefore more than makes up for the greater loss.
Friction at a vibrating tip
Each vibration of an ultrasonic blade has an excursion of 50 to 100 micrometers, and a movement that small is still fast, because it repeats tens of thousands of times a second. If 100 µm is the full stroke at 55.5 kHz, the tip reaches a peak speed of about 17 meters per second and a peak acceleration of about six million meters per second squared, more than 600,000 times the acceleration of gravity.
The blade shears and rubs the tissue pressed against it at high speed, so friction at the interface and internal losses in the deformed tissue turn the mechanical energy into heat, concentrated where the tissue touches the blade. As in Chapter 1, the proteins unfold and form a sticky coagulum that can seal small vessels, and with enough pressure and energy the tissue parts along the blade. No current flows through the patient, because the transducer is driven electrically inside the handpiece and only vibration leaves it. Chapter 11 follows the whole instrument.
Cavitation
The rarefaction half of each acoustic cycle pulls on the liquid in tissue, and if the pull is strong enough, small gas bubbles can grow in it. The growth and motion of such bubbles in a sound field is cavitation, and it takes two forms. In stable cavitation, a bubble oscillates gently in size in step with the wave, swelling a little in each rarefaction and shrinking in each compression. In inertial cavitation, a bubble grows to many times its size, within one cycle or over several, and then collapses violently. The collapse concentrates the bubble's energy into a tiny volume and launches a shock wave, and it is the collapse that damages tissue.
Two surgical tools use cavitation on purpose. An ultrasonic aspirator vibrates a hollow tip at 23 or 36 kHz and creates cavitation just ahead of the tip's rim. Together with direct mechanical cutting, the cavitation fragments tissue with a high water content and weak cohesion, while it spares collagen-rich structures such as vessels, nerves and ducts. Vieira-Coimbra and Ferreira describe this selectivity in a prospective case series of 2025 on the laparoscopic excision of endometriosis with the CUSA aspirator. Integra LifeSciences, which makes the CUSA, supported the publication.
The second tool, histotripsy, generates inertial cavitation deep inside the body with focused pulses (Chapter 16). Whether a pulse starts cavitation depends on whether it pulls hard enough on the liquid, and Vlaisavljevich and colleagues measured that threshold in a 2015 laboratory study. For single-cycle pulses, cavitation starts even without a seed bubble once the peak negative pressure exceeds about 25 to 30 megapascals. The study found this threshold nearly the same in excised tissues of different stiffness and in tissue-like gels, at frequencies from 345 kHz to 3 MHz, so one narrow band of pressure describes the onset across all of them. Three of its authors declared financial interests or other relationships with HistoSonics, which makes histotripsy systems.
Cavitation damages tissue mechanically, with little heating, so a device that uses it can be mistaken for a gentle one. The energy still reaches the tissue, but as pressure rather than heat. Whether a structure survives depends on its mechanical strength, which is why collagen-rich vessels survive and an organ's parenchyma does not. Parenchyma is the soft, water-rich working tissue of an organ.
Sound can act on tissue in three ways, and different devices use each one. When tissue absorbs sound along its path, the sound heats it, which is how thermal focused ultrasound works. When a vibrating surface rubs against tissue, friction heats the tissue at that surface, which is how the ultrasonic shear works. When the negative half of each cycle pulls hard enough on the liquid in tissue, bubbles grow and tear the tissue mechanically, which is how the aspirator and histotripsy work.
Trace the path from the transducer to the target before judging a focused-ultrasound design or its test. List every boundary on that path, because a thin layer of air reflects almost all of the sound, and bone reflects about 39 % of it and absorbs part of the rest, which heats the bone. Work out the loss over the target's depth at the chosen frequency, at about 0.54 dB per centimeter for every megahertz, and set it against the wavelength, which limits how small the focus can be. For a device that relies on cavitation, specify the peak negative pressure it reaches as well as the temperature, because cavitation damages tissue mechanically with little heating.
- Sound travels through soft tissue at about 1,540 m/s, so its wavelength at 1 MHz is about 1.5 mm, and the wavelength limits how finely a beam can focus.
- A boundary reflects more sound the larger the mismatch in acoustic impedance across it: 0.2 % between tissue and water, about 39 % at bone and 99.9 % at air.
- Tissue attenuates sound by about 0.54 dB per centimeter for every megahertz, so a higher frequency focuses more finely but reaches less deep.
- A vibrating tip heats tissue by friction at its surface, while cavitation tears tissue mechanically and, with single-cycle pulses, starts without a seed bubble above a peak negative pressure of about 25 to 30 MPa.
Light in tissue.
+ The questionWhy does light from one laser stop within a hundredth of a millimeter of the surface, while light from another reaches millimeters deep?
Absorbers and scatterers
Tissue is about three quarters water, and the light of a carbon dioxide laser is absorbed within about 16 µm of its surface, yet the light of a diode laser at 980 nm enters the same tissue and spreads over millimeters before it is absorbed. The difference lies in which parts of the tissue absorb each wavelength and which mostly scatter it. Some molecules absorb light, taking each photon's energy and turning it into heat. These absorbers are called chromophores, and three matter most in surgery: water, which is everywhere, hemoglobin in the blood and melanin in pigmented skin. Cell membranes, fibers, organelles and the rest of the tissue mostly scatter light, bending it in new directions without absorbing it.
How strongly a chromophore absorbs depends on the wavelength, and the dependence is often enormous. Water is almost perfectly transparent to green light and almost opaque to much of the infrared, while hemoglobin absorbs green light strongly and near-infrared light weakly. A laser emits one wavelength, so choosing a laser means choosing which chromophore will do the absorbing, and therefore where the heat will appear.
The Beer–Lambert rule
The Beer–Lambert rule describes light that is absorbed and not scattered: its intensity falls by the same fraction over every equal step of depth. How fast it falls is set by the absorption coefficient, written µa, which is the probability of absorption per centimeter. At a depth of one divided by the absorption coefficient, the intensity has fallen to about 37 % of its value at the surface, and that depth is the absorption depth, within which most of the light's energy, and so most of the heat, is deposited.
Across the surgical wavelengths, published absorption coefficients for water span about seven and a half orders of magnitude, so the strongest absorption is tens of millions of times the weakest. A carbon dioxide laser emits at 10.6 µm, where water absorbs about 855 per centimeter, so the absorption depth in water is about 12 µm. A KTP laser emits green light at 532 nm, where water absorbs only about 0.00044 per centimeter, which gives an absorption depth of more than twenty meters.
Water rules the infrared beyond 1.4 µm
Most surgical lasers work in the infrared, where water absorbs strongly, so the depth at which their light is absorbed follows almost directly from water's absorption coefficient. Tissue is about 75 % water, so it absorbs these wavelengths about three quarters as strongly as pure water, and each depth in tissue is about a third longer than in pure water.
| Laser | Wavelength | Water absorption (per cm) | Absorption depth in tissue |
|---|---|---|---|
| Er:YAG | 2.94 µm | about 12,500 | about 1 µm |
| CO₂ | 10.6 µm | about 855 | about 16 µm |
| Thulium fiber | 1.94 µm | about 120 | about 0.11 mm |
| Diode | 1.47 µm | about 29 | about 0.47 mm |
| Ho:YAG | 2.1 µm | about 27 | about 0.5 mm |
Each laser's character follows from its absorption depth. An Er:YAG laser deposits its energy within about a micrometer of the surface, so it removes tissue in thin layers and leaves very little heat behind. A CO₂ laser deposits its energy in a layer about 16 µm deep and vaporizes that layer, which makes a precise incision but seals only the smallest vessels, because very little tissue below the cut is heated. The light of holmium and thulium lasers is absorbed within about half a millimeter and a tenth of a millimeter, and both lasers cut and coagulate, but because water absorbs thulium's light about four to five times more strongly than holmium's, thulium puts the same energy into a shallower layer.
The depths in the table are optical: they show where the light is absorbed. The thermal damage beneath an incision usually reaches deeper, because heat conducts outward from the absorbing layer during and after the pulse. In a 2012 position paper for the Academy of Laser Dentistry on lasers in oral soft tissue, Pang and colleagues give thermally affected layers of 10 to 50 µm for Er:YAG, ten to fifty times its optical depth. How much deeper the damage reaches depends on the length of the pulse, which the last section of this chapter takes up.
Hemoglobin and scattering at shorter wavelengths
Toward the visible end of the spectrum, water absorbs very little, and two other effects take over: absorption by hemoglobin, and scattering. At 532 nm, oxygenated whole blood absorbs about 235 per centimeter, an absorption depth of about 43 µm, while the water around it is almost transparent. Green light therefore passes through water or saline irrigation and is absorbed by tissue filled with blood. That difference is the basis of KTP laser vaporization of the prostate, a procedure that runs under irrigation. Blood absorbs roughly 40 to 200 times more strongly at 532 nm than near 1,000 nm, depending on how much oxygen it carries.
At 980 and 1,064 nm neither water nor hemoglobin absorbs strongly, so scattering dominates, and the light bounces around inside the tissue and spreads over millimeters before it is absorbed. Diode lasers at 980 nm and Nd:YAG lasers at 1,064 nm therefore coagulate deeply rather than cut precisely. Scattering sets their depth, and because scattering varies with the tissue, published depths for these lines are less consistent than those for the lines that water absorbs.
Time: confining the heat
Chapter 2 showed that heat needs time to move. A layer that has absorbed light cools by conducting heat into its surroundings, and the time this cooling takes is the thermal relaxation time. It is roughly the layer's thickness squared divided by four times the thermal diffusivity, which for water and most soft tissues is about 0.14 square millimeters per second.
Four times the thermal diffusivity of tissue is 4 × 0.14, or 0.56 square millimeters per second. The layer that absorbs CO₂ light is about 16 µm, or 0.016 mm, thick. Its thickness squared is 0.000256 square millimeters, and dividing that by 0.56 gives about 0.00046 seconds, a relaxation time of about half a millisecond. The layer that absorbs Er:YAG light is about 1 µm thick, sixteen times thinner, so its relaxation time is sixteen squared, or 256 times, shorter: about 2 microseconds. A layer 0.1 mm thick, the size of a small blood vessel, has a thickness squared of 0.01 square millimeters, and dividing by 0.56 gives about 18 milliseconds. The relaxation time therefore grows with the square of the thickness, so a layer sixteen times thinner loses its heat in a 256th of the time.
When the energy arrives in a pulse shorter than the relaxation time, the absorbing layer heats and vaporizes before it can pass heat to its neighbors, and little tissue beneath the crater is damaged. This is thermal confinement. When the pulse is longer, heat spreads while energy is still arriving, and more tissue is coagulated. Neither choice is right in general, because confinement gives precision while spreading gives hemostasis, the stopping of bleeding.
Power is the rate at which energy arrives, so the length of a pulse also sets its peak power. One joule delivered in 300 µs is a peak of about 3.3 kW, while the same joule spread over 50 ms is 20 W. At very high peak power, especially in liquid, the absorbed energy boils water so fast that the vapor expands explosively and generates pressure waves. The effect then becomes partly mechanical as well as thermal, and Chapter 12 shows how lithotripsy lasers use it to break stones.
Read a surgical laser's specification as a set of choices about where the heat goes. Start with the wavelength, because it decides which chromophore absorbs the light and so how deep the energy is deposited, from about a micrometer for Er:YAG to millimeters for the lines at 980 and 1,064 nm. Then set the pulse length against the thermal relaxation time of the absorbing layer, because a shorter pulse confines the heat and a longer one coagulates the tissue around it. Check the peak power as well as the pulse energy, because at very high peak power the effect becomes partly mechanical. In verification, measure the thermally affected depth in tissue rather than quoting the optical depth, because heat conducts beyond the layer that absorbed the light.
- Tissue absorbs light through chromophores, mainly water and hemoglobin, and scatters it off everything else.
- Without scattering, light loses the same fraction of its intensity over each equal step of depth, and water's absorption spans about seven and a half orders of magnitude across surgical wavelengths, from a depth of about 1 µm for Er:YAG to meters at 532 nm.
- Water sets the depth of the infrared lasers beyond 1.4 µm, hemoglobin that of the 532 nm line, and scattering that of the 980 and 1,064 nm lines, which therefore coagulate deeply.
- A pulse shorter than the thermal relaxation time confines the heat and gives precision, while a longer pulse spreads the heat and gives hemostasis.
Six ways to deliver energy.
+ The questionIf most energies end as heat in tissue, why does it matter what carries it there?
One frequency axis, two kinds of wave
Surgical energy reaches tissue at frequencies from 23 kHz to about 564 THz, a span of more than ten tenfold steps. A carrier is the current, vibration, field or light that brings energy to the tissue, and the previous chapters took the carriers one at a time. Placed on one frequency axis, the carriers fall in this order: ultrasonic instruments vibrate at 23 to 55.5 kHz, most RF generators run at 300 to 500 kHz, focused ultrasound works at 0.7 to 4 MHz and microwave ablation uses 915 MHz or 2.45 GHz. Surgical lasers emit infrared and visible light from about 28 THz, for the 10.6 µm of CO₂, to about 564 THz, for the 532 nm of KTP. Pulsed-field systems deliver bursts of microsecond pulses of current, which have no single carrier frequency but occupy the low end of the same range. That makes six carriers: the vibration of ultrasonic instruments, RF current, focused sound, microwaves, light and pulsed fields. They are the six ways of delivering energy that this chapter compares.
The lowest carrier is the vibration of an ultrasonic instrument at 23 kHz, and multiplying it by ten, step by step, places each of the other carriers on the axis. One tenfold step gives 230 kHz, just below the RF generators at 300 to 500 kHz. Two steps give 2.3 MHz, inside the band of focused ultrasound, and five steps give 2.3 GHz, beside the 2.45 GHz of microwave ablation. Nine steps give 23 THz, a little below the 28 THz of the CO₂ laser, and ten give 230 THz. The KTP laser's 564 THz lies between the tenth step and the eleventh, so the carriers span more than ten tenfold steps, a factor of about 25 billion between the lowest frequency and the highest.
The carriers on the axis are two kinds of wave. Acoustic carriers are mechanical, because in them matter moves back and forth. Electromagnetic carriers are currents, fields and light, in which charges and fields move with no bulk motion of the tissue. Ultrasonic shears and RF generators sit within about a factor of ten of each other in frequency, yet they share nothing in how they heat. Cryoablation has no place on the axis at all, because it carries no energy into the tissue and takes energy out instead.
Seven mechanisms
What matters to tissue is the mechanism by which a carrier deposits its energy, not the carrier's frequency. Every energy in this guide uses one of seven.
- Joule heating. Current flows through tissue and releases heat in each unit of its volume at a rate equal to the square of the current density divided by the conductivity, so the heat is greatest where the current is densest. Monopolar and bipolar electrosurgery, vessel sealing and RF ablation all work this way (Chapter 3).
- Friction and cavitation. A vibrating metal surface heats tissue by friction where the two touch and churns it by cavitation, which is how ultrasonic shears and aspirators work (Chapter 4).
- Photothermal absorption. A chromophore absorbs a laser's light within about one absorption depth of the surface, and the wavelength sets that depth (Chapter 5).
- Dielectric heating. In microwave ablation, an alternating electromagnetic field rotates water molecules and drives ions back and forth, and both motions dissipate energy as heat throughout a volume, with no conducting path to an electrode (Chapter 13).
- Focused acoustic absorption. In thermal focused ultrasound, sound converges on a small focal volume deep in the body and is absorbed there as heat (Chapter 16).
- Heat extraction. In cryoablation, a probe colder than tissue draws heat out, and ice grows outward from it (Chapter 14).
- Non-thermal injury. In irreversible electroporation and pulsed field ablation, strong electric pulses open pores in cell membranes, and in histotripsy, intense acoustic pulses tear cells apart by cavitation. The cells die without being heated much (Chapters 15 and 16).
The first five mechanisms deposit heat, but in five different places: in the tissue where current crowds, at the face of a blade, in a layer at the surface, in a volume around an antenna and at a point deep inside the body. Those places answer the part of this guide's central question that asks why surgery needs eleven kinds of energy. Each surgical task needs heat in a particular geometry: a line for an incision, a band across a clamped vessel, a surface layer for a precise ablation, a sphere several centimeters across for a tumor, a point behind intact skin. No single carrier can produce all of those geometries.
The seventh mechanism completes that answer. Heat that kills cells also destroys the collagen framework of vessels, bile ducts and nerves, and for a tumor wrapped around a major vessel, or for cardiac tissue next to the esophagus, destroying that framework is exactly the wrong effect. The non-thermal energies exist because the task is sometimes to kill cells while leaving their scaffold intact.
What the carrier decides for the device
The carrier also decides almost everything about the device that delivers it, so an engineer or founder who chooses an energy chooses a whole set of consequences at once.
Whether current flows through the patient. RF energies need a complete circuit through the patient, so they need either a return pad or a second electrode, and they bring every stray-current hazard of Part VII. Ultrasonic, laser and microwave energies need no circuit through the patient, which removes those hazards and brings others.
What the instrument must be. RF needs electrodes and insulation rated for kilovolts, and pulsed fields need electrodes that can withstand thousands of volts. Ultrasonic energy needs a titanium waveguide cut to exact half-wavelengths of its vibration. Lasers need optical fibers or, for CO₂, an articulated arm of mirrors. Microwaves need a coaxial cable and a cooled antenna, and cryoablation needs high-pressure gas lines or liquid nitrogen.
What the power electronics must do. An RF generator is an inverter with an isolation transformer, and an ultrasonic generator is a resonant driver that tracks a mechanical resonance. A laser needs a pump source and cooling, and a microwave generator is a gigahertz source. A pulsed-field generator stores energy and releases it as kilovolt pulses, while a cryoablation console controls gas flow rather than electrical power.
Which standards apply. Electrosurgical equipment is tested to IEC 60601-2-2, lasers to IEC 60601-2-22 and IEC 60825-1, and high-intensity therapeutic ultrasound to IEC 60601-2-62. Ultrasonic surgical systems have a standard for their output, IEC 61847, but no particular safety standard of their own was found for this guide. Chapter 25 takes up the standards in full.
The eleven energies at a glance
| Energy | Carrier | Mechanism | Typical task | Chapter |
|---|---|---|---|---|
| Monopolar RF | current, 300–500 kHz | Joule heating at a small electrode | cut, coagulate, dissect planes | 8 |
| Conventional bipolar RF | current between two tips | Joule heating between the tips | coagulate small vessels | 9 |
| Advanced bipolar sealing | current through a clamped vessel | Joule heating under pressure | seal vessels up to 7 mm | 9 |
| Ultrasonic | vibration, 23–55.5 kHz | friction and cavitation at the tip | seal and divide; fragment | 11 |
| Laser | light, 532 nm to 10.6 µm | photothermal absorption | cut, vaporize, coagulate, break stones | 12 |
| RF ablation | current, about 500 kHz | Joule heating around a needle | destroy tumors in place | 13 |
| Microwave ablation | field, 915 MHz or 2.45 GHz | dielectric heating | destroy tumors in place | 13 |
| Cryoablation | none; cold | heat extraction | destroy tumors in place | 14 |
| IRE and PFA | kilovolt pulses | membrane pores, non-thermal | destroy tissue, spare scaffold | 15 |
| Thermal focused ultrasound | focused sound, 0.7–4 MHz | focal absorption | destroy tissue without incision | 16 |
| Histotripsy | focused sound pulses | inertial cavitation, non-thermal | destroy tissue without incision | 16 |
Two devices can reach the same final temperature in the same tissue and still behave completely differently, because they get there through different geometry, speed and side effects. The carrier decides the geometry and the side effects, and the generator's control decides the speed. Each instrument chapter that follows explains its devices through those three properties.
Before comparing products, place each one on this table by its carrier, its mechanism and where it puts the heat. Two devices that put their effect in different places suit different steps of an operation, so many apparent disagreements about which device is "better" are settled once each one is placed. Chapter 17 does this for each step of one colectomy, an operation to remove part of the colon.
- The carriers span more than ten tenfold steps of frequency in two kinds of wave, acoustic and electromagnetic, and cryoablation has no carrier at all.
- Every energy acts through one of seven mechanisms: five deposit heat, each in a different place, one removes heat, and one group kills cells without heating them much.
- Surgery needs several energies because each task needs heat in its own geometry, and some energies are non-thermal because some tasks must spare the collagen scaffold of vessels, ducts and nerves.
- The carrier fixes the instrument, the power electronics, the hazards and the standards a device must meet.
+ Part II · Access
The route to the tissue.
A 5 mm laparoscopic instrument must carry a conductor, its insulation, a jaw mechanism and often a cutting blade, and most of its shaft is out of the camera's view. Laparoscopic and robotic surgery add no new energy, but they change the conditions under which every energy works: what the surgeon can see and feel, how long and thin the instruments must be, where smoke and heat go, and who or what stands between the surgeon's foot and the generator. This part sets out those conditions before the instrument chapters meet them.
Open, laparoscopic, robotic.
+ The questionWhat changes for an energy device when the surgeon can no longer see or touch the tissue it acts on?
Open surgery: the reference case
An energy instrument in open surgery sits in the surgeon's hand a few centimeters from the tissue, and the whole of it is in view. The surgeon can feel the tissue, smoke drifts away into the room, and a burn in the wrong place is usually seen when it happens. Energy devices were designed for these conditions first. The apparatus that made electrosurgery routine was Bovie's generator, and Harvey Cushing first used it in open surgery, in the operation of the opening story, in the autumn of 1926.
Almost everything in open surgery gives the surgeon feedback that needs no thought. The color of the tissue shows that it has coagulated, the resistance it offers the instrument shows that it is being cut, and its smell and sound show when it is drying and charring. Minimally invasive surgery works through small openings in the body wall, and it removes most of that feedback, so energy devices have had to replace it.
Laparoscopy: working through a keyhole
In laparoscopic surgery the abdomen is inflated with carbon dioxide at 12 to 15 mmHg to lift the wall away from the organs. The instruments enter through cannulas, tubes that pass through the wall, and a camera on a rigid telescope shows the field on a monitor. Each of those facts changes how energy behaves. The view is partial, because the camera shows the working ends of the instruments and a few centimeters of their shafts, while the rest of each shaft lies inside the abdomen, out of view and within reach of the bowel.
The instruments are also long and thin, because they must reach the tissue through the cannulas. Cannulas come in a range of sizes, and one maker, Johnson & Johnson, sells them from 5 to 15 mm in diameter. Even at the small end of that range, a 5 mm instrument must carry a conductor, its insulation, a jaw mechanism and often a cutting blade. Its insulation is thin and rubs against the cannulas, and if the instrument is reusable, the insulation is cleaned and sterilized, or reprocessed, again and again. A long insulated conductor inside a metal cannula also forms a capacitor, which means that alternating current can cross the insulation to the metal without any contact, a transfer called capacitive coupling. Where the shaft is out of the camera's view, energy that leaves it through damaged insulation or by capacitive coupling goes unseen, and Chapter 18 is about exactly that.
The space is small and closed, so smoke from a few seconds of cutting fills the gas volume and obscures the camera, and the smoke has to be evacuated while the pressure is maintained. Heat has nowhere to go but into the tissue, because there is no room air to cool a hot jaw between activations, and a jaw that has just sealed a vessel can touch the bowel on its way to the next bite.
Motion is reversed as well: each instrument pivots at the abdominal wall, so its tip moves opposite to the hand. The surgeon also feels force poorly through a long shaft and the valve that keeps the gas in the cannula, so judging by feel alone how hard a jaw is squeezing, or when a seal is complete, becomes unreliable.
Robotic surgery: a controller in the loop
A surgical robot restores some of what laparoscopy removed: the surgeon sits at a console with a three-dimensional view, the instruments have wrists that bend inside the body, and the system scales the surgeon's motion and filters out tremor. The instruments still enter through ports in an abdomen inflated with gas, so every condition above still applies.
The robot also changes how the generator is switched on. In open and laparoscopic surgery, the surgeon's pedal or hand switch activates the generator directly, while in robotic surgery the command passes through the robot's own controller, which decides whether it reaches the generator. Some makers pair the generator with the robot. Intuitive, for example, describes its E-200 generator as integrated with its da Vinci systems, with the energy settings controlled from the surgeon's console, and its vessel sealers and its SynchroSeal instrument work with that generator. Energy delivery then depends on two control systems working together, so the risk analysis must cover both (Chapter 21).
A robot moves the instruments, but it does not decide by itself how much energy to deliver. On robotic systems the surgeon still commands the energy and a generator still controls it, and the same physics, hazards and tests apply as in laparoscopy. What the robot adds is another controller and another set of interfaces that must be verified.
Robotic instruments are reused a set number of times and have complex wristed shafts, and they bring an insulation problem of their own. In a prospective study published in 2011, Espada and colleagues at Mayo Clinic Arizona tested robotic and laparoscopic instruments for insulation failure. At a single test, they found failures in 32 % of the robotic instruments against 13 % of the laparoscopic ones. Among the instruments followed over time, 80 % of the robotic ones failed within ten uses, against 36 % of the laparoscopic ones. Chapter 18 returns to these numbers.
At the single test, the study found failed insulation on 25 of 78 robotic instruments and on 39 of 298 laparoscopic ones. Dividing each count by its total gives 32 % and 13 %. The figures over time rest on other totals. Of the 44 robotic instruments followed, 35 failed within ten uses, which is 80 %, and of the 189 laparoscopic instruments followed, 68 failed, which is 36 %. The first pair is the share found failed at one test, and the second is the share that failed while the instruments were followed, so the two pairs cannot be read as one rate rising with use. Within each pair, robotic instruments failed about 2.4 times as often as laparoscopic ones at one test, and about 2.2 times as often over time.
What the access route asks of energy devices
Taken together, minimally invasive access asks four things of an energy device that open surgery did not. The first is to do more than one job per instrument, because every exchange of instruments takes time and risks losing the view, so instruments that grasp, seal and cut in one device became central to laparoscopy. Advanced bipolar sealers with built-in blades, ultrasonic shears and hybrid instruments all answer this need. A hybrid instrument combines two energies in one jaw, such as bipolar current and ultrasonic vibration. Chapter 9 describes the sealers, and Chapter 11 the ultrasonic and hybrid instruments.
The second is to decide the end point, because when the surgeon cannot see the tissue blanch or feel it change, the generator must decide when a seal is finished. The answer is closed-loop control, in which the generator measures the tissue and stops by itself, and Chapters 9 and 10 describe how it works.
The third is to keep the energy inside the instrument: insulation, shielding and monitoring must stop current from leaving the shaft outside the camera's view. The fourth is to manage what the energy leaves behind. Smoke must be evacuated without losing the pressure that holds the abdomen open, and the heat retained in jaws and blades must be managed. For that heat, at least one current hybrid instrument now includes a thermal shield that slows the transfer of heat to the outer surface of its jaw.
Write the access route into the design inputs of a device for minimally invasive use from the start. List the shaft length and diameter, the insulation material and thickness, compatibility with metal cannulas and with the hybrid cannulas of metal and plastic that Chapter 18 describes, and smoke management. For robotic use, add the interface to the robot's controller and its activation path. Plan the capacitive-coupling test early, because FDA's electrosurgical guidance asks for one against a conductive cannula for monopolar instruments used in minimally invasive surgery.
- Minimally invasive access adds no energy but changes the conditions: a partial view, long thin instruments, a small closed space and reversed motion.
- Most of each instrument shaft is out of the camera's view and within reach of bowel, so energy that leaves it through damaged insulation or by capacitive coupling goes unseen.
- Robotic surgery restores dexterity and depth perception but places the robot's controller between the surgeon's pedal and the generator.
- The access route pushes energy devices toward multifunction instruments, closed-loop end points, containment of stray energy and management of smoke and residual heat.
+ Part III · Current through the patient
Electrosurgery.
A vessel sealer can close an artery up to 7 mm across, and nothing inside its jaws measures the seal: the generator infers it from the voltage and current at its own output. Electrosurgery was in surgical use by 1926, the year of the operation in the opening story, and from Bovie's cutting loop to today's sealer it rests on the rule of Chapter 3, that heat appears where current is densest. This part follows that current from the tip of a monopolar pencil, through the jaws of a sealer, to the generator that drives both, and shows how each is built so that the heat appears only where the surgeon intends.
Monopolar RF.
+ The questionIf current must cross the whole body to complete its circuit, how does it heat only the point the surgeon touches?
The circuit
The pad on a patient's thigh or back carries the same current as the electrode in the surgeon's hand, yet tissue at the tip boils within milliseconds while the skin under the pad warms by a few degrees. The difference comes from the geometry of the circuit alone. A generator drives high-frequency current through a lead to the active electrode, the pencil, hook or scissors in the surgeon's hand. The current enters the patient at the electrode's tip, spreads through the body and leaves through a large return electrode, the pad, which leads it back to the generator. The patient lies in the middle of the circuit, a conductor of irregular shape whose conductivity changes from one tissue to the next.
The same current flows through the tip and the pad, so the difference lies in the area each one offers it (Chapter 3). At the tip the current is squeezed into a fraction of a square millimeter, while at the pad it spreads over tens of square centimeters. Current density at the tip is therefore thousands of times higher, and because heating rises with the square of current density, the heat released in each cubic millimeter there is higher by a factor of millions.
The early history of electrosurgery is a history of learning where else the current could go. William Bovie filed a patent on his apparatus in 1928, and the apparatus it describes made both its cutting and its coagulating currents with spark gaps, as the opening story tells. In the same story, current from Bovie's early apparatus once reached Cushing's arm through a metal retractor. For decades afterward, generators were built with outputs referenced to earth. Current from such an output could return through any earthed contact on the patient, and where that contact was small, it burned. Valleylab's patent filed in 1971 names the electrodes of patient-monitoring equipment as one such path. It describes a generator built from solid-state components, with an isolated output that floats electrically and returns current only through its own return lead, "to prevent patient burns caused by alternate return current paths".
At the tip: from contact to arcs
What the current does at the tip depends mainly on the voltage. Below about 200 V peak, current flows by direct contact between the electrode and moist tissue. The tissue heats, its water boils away gradually, and it dries into a coagulum. This is the desiccation of Chapter 1, or soft coagulation, and it makes no sparks.
Above about 200 V peak, the first burst of heating makes a thin layer of steam between the electrode and the tissue. Steam does not conduct, but at these voltages the field across the thin layer is strong enough to ionize it, and current crosses the gap as tiny arcs. Each arc concentrates its energy into a very small spot, so the cells beneath it vaporize and burst. The electrode rides on its own envelope of steam and moves through the tissue along a line of bursting cells without touching it, which is how it cuts without being sharp. In a 2024 review of the physics of electrosurgery, Koninckx and colleagues put the onset of sparking at about 200 V peak. One of the authors is affiliated with BOWA-electronic, a generator maker. The Technology Committee of the American Society for Gastrointestinal Endoscopy states in a 2013 report that a continuous current with a peak above 200 V is often called a pure-cut current. Both figures put the change from contact heating to cutting at about 200 V peak.
At higher voltages still, in the kilovolts, interrupted bursts of current jump an air gap to the tissue surface. The arcs spread across the surface, dry and char it, and stop oozing from a broad area. This is fulguration, or spray coagulation, and it needs no contact at all.
The waveform decides the effect
A generator produces these effects by changing the shape of its output more than its power. For cutting it delivers a continuous sine wave of a few hundred volts. For coagulation it delivers the same carrier in short bursts, switched on for only a small part of each cycle, at a much higher peak voltage. The fraction of the time the output is on is the duty cycle. The RMS voltage is the steady voltage that would deliver the same power into the same load, and the ratio of the peak voltage to the RMS voltage is the crest factor. A continuous sine wave has a crest factor of 1.41. The lower the duty cycle, the higher the crest factor, and the higher the peak voltage must be to deliver the same power.
Two waveforms are to deliver the same average power into the same tissue: a continuous sine wave, and bursts of the same carrier switched on for 6.25 % of the time, one sixteenth. Power into a fixed load follows the square of the RMS voltage, so the two waveforms need the same RMS voltage. Switching the sine on for one sixteenth of the time divides its mean squared voltage by 16, so its RMS voltage falls to a quarter, the square root of one sixteenth. To restore it, the peak of each burst must be four times the peak of the continuous sine, and the crest factor rises by the same factor, from 1.41 to about 5.7. At equal power, a waveform that is on for one sixteenth of the time must therefore reach four times the peak voltage. Medtronic's specification for its Valleylab FT10 generator lists a typical crest factor of 5.4 for its fulguration mode, which runs at 6.25 %, close to the 5.7 computed here for ideal bursts.
Medtronic's specification sheet shows the same pattern across the FT10's monopolar modes, the preset outputs that the surgeon selects for each task: the lower a mode's duty cycle, the higher its crest factor and its peak voltage. The FT10's pure cut runs at a 100 % duty cycle, with a typical crest factor of 1.6 and a peak voltage of 1,287 V. Its blend runs at 50 %, with a crest factor of 2.2 and 2,178 V, and a third cut mode, which Medtronic names Valleylab, runs at 25 %, with 3.2 and 2,783 V. Its fulguration runs at 6.25 %, with 5.4 and 3,448 V, and its spray at 4.76 %, with 6.2 and 3,932 V.
The high peak voltage of the coagulation waveforms is what lets them arc across a gap, and it is also what makes them dangerous in the ways Part VII describes. A higher voltage breaks through damaged insulation more easily and drives more current through stray capacitance. The same power cuts at a low crest factor, and coagulates and fulgurates at a high one, so the effect on tissue follows the waveform far more than the number on the power dial.
Medtronic's table for the FT10 gives peak voltage, measured from zero to the top of the wave. A reseller's copy of the table for the older Valleylab Force FX gives peak-to-peak voltage, from the bottom of the wave to the top, and Koninckx and colleagues' 2024 review gives both. For a symmetrical wave the peak-to-peak figure is twice as large, so a "cut at 2,000 V" and a "cut at 1,000 V" can describe the same waveform. A comparison of two figures starts by establishing whether each source means peak or peak-to-peak.
The return pad
The return pad works only while its contact area stays large. If part of it peels away, the same current crowds into the remaining contact, and because heating rises with the square of current density, the skin under the edge that still touches heats fastest. A pad that has lost two thirds of its contact carries three times the current density, and each square centimeter of skin still in contact heats about nine times as fast.
Two engineering responses followed. The first is a test. The AAMI HF18 standard required that a return electrode carrying 700 mA for 60 seconds must not raise the temperature of the skin beneath it by more than 6 °C, and IEC 60601-2-2 sets the same 6 °C limit for its own test of return electrodes (Chapter 25). A MegaDyne patent cites the Emergency Care Research Institute for a threshold of necrosis under a return electrode of about 100 mA per square centimeter. An adult pad of about 130 cm² carrying 700 mA has a current density of about 5 mA per square centimeter, some twenty times below that threshold. With two thirds of it lifted, the same pad carries about 16 mA per square centimeter. That is still below the threshold, but each square centimeter in contact heats nine times as fast as in the test that allows a rise of 6 °C.
The second response is a monitor. A split pad has two halves, and the generator passes a small interrogating current from one half to the other through the skin. If the pad starts to lift, the impedance between the halves rises. Valleylab filed a patent on the idea in 1981, and it was granted in 1983. Its monitor uses a 2 mA signal and accepts an impedance between a fixed lower limit of 20 Ω and an adaptive upper limit, set at 120 % of the impedance it records once the pad is in place. When the impedance leaves that window, the generator stops. Later Valleylab generators use the same principle with other limits. The user's guide and service manual of the Force EZ give 5 to 135 Ω as the safe range and describe an alarm when the impedance rises more than 40 % above its starting value. Medtronic's 2023 specification for the FT10 states that its monitor measures contact resistance from 5 to 135 Ω, and gives no alarm rule.
The monitor protects only if the pad is used as designed. Pads come in sizes for neonates, infants, children and adults, and AORN, the association of perioperative nurses, states in a 2016 sample policy that they must not be cut or folded. Cutting reduces the area and can defeat the split-pad monitor. In 2011, 48 leaders of the Society of American Gastrointestinal and Endoscopic Surgeons took a knowledge test, and 10 % of them thought a pad should be cut to fit a child (Chapter 18 returns to the test). The same policy asks for the pad to go over a large, well-perfused muscle, as close to the operative site as practical, so that the current's path through the body is short.
Blocking capacitors
One more component sits in every monopolar output: a capacitor in series with each lead. Chapter 3 noted that an arc conducts slightly better in one direction than the other, so it partly rectifies the current and leaves a small direct-current component. A series capacitor passes the high-frequency current freely but blocks direct and low-frequency current, which could otherwise stimulate nerves and muscles. An ArthroCare patent application filed in 2010 points out a complication: the charge that arcing builds up on such a capacitor can itself discharge through tissue at a low frequency once the arc stops, at 100 Hz in the application's example. The size of the capacitor is therefore a design decision with consequences of its own.
Read a monopolar generator's datasheet for three things from this chapter. First, check the peak voltage and crest factor of each mode, because they govern arcing and coupling, and whether the voltage is quoted as peak or peak-to-peak. Second, check the rated load, because each waveform is specified at that load and nowhere else. Third, check the return-pad monitor's alarm rule, and whether the datasheet states one, because the rule decides how much pad lift the generator tolerates before it stops.
- Monopolar current enters through a small active electrode and leaves through a large return pad, and only the difference in area confines the heat to the tip.
- Below about 200 V peak, contact current desiccates tissue; above it, current arcs across a steam envelope and vaporizes cells, so the electrode cuts without being sharp.
- A lower duty cycle raises the crest factor, so at the same power it needs a higher peak voltage, and the waveform decides cut, blend, fulguration or spray more than the power setting does.
- Isolated outputs return current only through the generator's own return lead, a split-pad monitor stops the generator when the pad lifts, and blocking capacitors stop the direct and low-frequency current that could stimulate nerves and muscles.
Bipolar and vessel sealing.
+ The questionA sealer cannot see inside its own jaws. How does it know when a seal is complete?
Bipolar: the circuit in the instrument
Bipolar forceps held on a vessel for too long dry the tissue until it chars and sticks to their tips, and when the forceps open, the vessel may tear. Held too briefly, they leave the vessel open. Both failures have one cause: with conventional bipolar forceps the surgeon decides how long to apply energy. A bipolar instrument carries both electrodes in its tips, usually the two blades of a forceps. Current flows from one tip to the other through only the tissue held between them, so no return pad is needed and almost no current wanders through the rest of the patient.
A generator offers bipolar modes alongside the monopolar modes of Chapter 8, and the bipolar modes work at low voltages. Medtronic's specification for its Valleylab FT10 generator lists the bipolar modes at 70 W into a rated load of 100 Ω (Chapter 10). Their maximum peak voltages run from 184 to 530 V depending on the mode, a fraction of the kilovolt peaks of monopolar coagulation in Chapter 8. Conventional bipolar forceps coagulate small vessels and bleeding points precisely. For vessels of 3 mm or more, however, conventional bipolar coagulation was not reliable enough to replace a tie, the suture knotted around a vessel to close it.
Advanced bipolar: pressure, current and a closed loop
The advanced bipolar vessel sealer changed that. Valleylab's LigaSure is one such sealer, and FDA cleared it on 28 August 1998. As the introduction explained, a device is cleared when FDA finds it substantially equivalent to one already legally on the US market, its predicate device. The maker makes that case in a premarket notification called a 510(k) (Chapter 26).
A later 510(k) summary for LigaSure describes the mechanism in regulatory language. Collagen and elastin in the tissue are re-formed by heat and pressure to fuse the vessel walls into a permanent seal, while a microprocessor monitors the tissue, ends energy delivery and allows the seal to cool before signaling that the seal is complete. LigaSure is indicated for vessels up to and including 7 mm in diameter, which means that sealing such vessels is the use its maker claims for it (Chapter 2). Many advanced bipolar sealers share the 7 mm limit, and some robotic sealers are limited to 5 mm.
Three things distinguish a sealer from bipolar forceps, and the first is that the jaws apply high, controlled pressure, squeezing the vessel flat so that its two walls touch along their whole width and blood is pushed out of the jaw. The generator delivers high current at low voltage: Medtronic rates the FT10's sealing mode at 20 Ω, with up to 350 W, a maximum current of 5.5 A and a peak voltage of at most 244 V. Into wet, clamped tissue the voltage stays far lower, so the current heats the tissue without arcing. And the generator runs a closed loop, measuring the tissue continuously and deciding for itself when to stop.
The seal itself is the collagen fusion of Chapter 1. Heat unwinds the helices of the collagen in the two opposed walls, and as the tissue cools under pressure, their collagen and elastin entangle into a single translucent band. Heat that escapes sideways from the jaw is the lateral thermal spread of Chapter 2. In a 2025 study funded by Medtronic, Chambers and two other Medtronic employees report a spread in living tissue of about 1.3 to 2.1 mm for Medtronic's current LigaSure sealers, with seal times of about 2 to 4 seconds.
Impedance tells the generator what is happening
Inside the jaw the generator has no camera and no thermometer in the tissue, so all it has is the voltage and current at its own output, and from those it computes the tissue's electrical impedance, many thousands of times a second. Medtronic states that the FT10's tissue-sensing technology monitors changes in tissue impedance 434,000 times a second.
The impedance follows the state of the tissue closely. As the tissue warms, its conductivity rises, so the impedance first falls. Below 42 °C, where it has been measured, the rise is about 2 % per degree, although a seal takes the tissue far above that temperature (Chapter 2). As the tissue approaches 100 °C, its water begins to boil off, the tissue that remains conducts worse and worse, and the impedance turns upward and then climbs steeply. When the climb shows that the water has largely gone, heating can do no more for the seal, and further energy would only char the tissue, so the generator stops, waits while the seal cools in the closed jaws, and sounds a tone.
Different devices apply the same idea in different ways. Some regulate on impedance directly. Ethicon's EnSeal uses a jaw material whose conductivity falls sharply as it warms, so each part of the jaw limits its own temperature, and Olympus's PK system pulses its bipolar energy. All of them manage the same physics, because each must deliver enough energy to fuse the collagen and stop before the tissue dries into char.
A sealer never measures the seal. It measures impedance and infers the seal from the shape of the impedance curve, using an algorithm tuned on bench and animal tests. That is why the strength of a seal must be demonstrated by testing rather than by reading the generator's display. It is also why a seal may not be as strong as the tone suggests when it is made outside the conditions the algorithm was tuned for, on a larger vessel or in unusual tissue.
How strong is a seal?
The bench test for a seal measures its burst pressure. A sealed vessel is filled with liquid and the pressure is raised until the seal fails, and the pressure at that moment is the burst pressure. The pressure in the arteries peaks with each heartbeat, and this systolic pressure is normally about 120 mmHg. A widely used acceptance level for a seal is 360 mmHg, three times systolic pressure. LigaSure's maker uses 360 mmHg as its lower specification limit, and published studies often compare against that level. The level is an industry convention and not a regulatory requirement, because FDA's guidance for bipolar vessel sealers sets out how to test burst pressure but gives no pass value.
Measured burst pressures are usually far above 360 mmHg when a device is used within its indication. Okhunov and colleagues published an independent bench comparison of five devices in 2018. The authors were at the University of California, Irvine, and Duke University, and they declared no competing financial interests. Three of the devices were bipolar sealers: LigaSure, EnSeal and Caiman. The other two were ultrasonic devices, the Harmonic ACE+ and the Harmonic ACE+7. On arteries of 5.1 to 7 mm, the mean burst pressures of the five ranged from about 570 to 1,740 mmHg, the lowest from the Harmonic ACE+, which is indicated for vessels only up to 5 mm. A mean can hide weak seals, though. Even on these vessels, 4 of 10 EnSeal seals and 2 of 10 Harmonic ACE+ seals burst below 300 mmHg, the study's threshold for failure.
The same study also tested all five devices on arteries of 7.1 to 9 mm. Four of them lost about 46 to 60 % of their mean burst pressure there. LigaSure fell from 981 to 530 mmHg beyond its 7 mm indication, EnSeal and the Harmonic ACE+7 both fell from 1,165 to 467 mmHg, and the Harmonic ACE+ fell to 254 mmHg, below the 360 mmHg convention, on vessels beyond its 5 mm indication. Caiman held about 1,680 mmHg, so one device kept its strength while four lost about half of it or more. Beyond the indicated vessel size, then, a seal's strength varies from device to device, and the maker's evidence no longer covers it.
A normal systolic pressure of about 120 mmHg, multiplied by three, gives the acceptance level of 360 mmHg. On arteries of 5.1 to 7 mm, within its 7 mm indication, LigaSure's seals in Okhunov and colleagues' 2018 comparison had a mean burst pressure of 981 mmHg. That is about eight times systolic pressure and about 2.7 times the acceptance level. On arteries of 7.1 to 9 mm the mean fell to 530 mmHg, a fall of 451 mmHg. Dividing 451 by 981 gives about 0.46, so the seals lost about 46 % of their strength. A mean of 530 mmHg is still about 4.4 times systolic pressure and still lies above the 360 mmHg convention, but the margin over that convention fell from about 2.7 times to about 1.5 times on vessels the device was not indicated for.
Burst pressure depends on the vessel's size, type and collagen content, on whether the vessel was tested fresh, and on how the pressure was applied. Numbers from different studies, or from a manufacturer's own study against a competitor, are rarely comparable. A fair comparison needs the same vessels, the same method and an independent laboratory.
FDA issued its guidance for bipolar vessel sealers in 2016, and it describes the evidence FDA expects before such a device reaches the market. Burst pressure is tested for each mode against the predicate device, across the vessel sizes and types claimed, and thermal spread is measured by histology, as Chapter 2 describes. A chronic animal study follows at least five animals for at least three weeks to show that the seals stay closed as the tissue heals. It is run under good laboratory practice, the rules FDA sets for the conduct of such studies. And the indication must state the maximum vessel size, measured in its rounded, perfused state in a subject with normal blood pressure.
Treat the design or evaluation of a sealer as three controlled variables and one inferred outcome. Jaw pressure, energy delivery and end-point detection are designed and verified, while seal strength is inferred and must be demonstrated. Test each variable at its limits: the lowest jaw force, the largest indicated vessel and the thickest tissue. A verification plan built that way tests the inference as well as the hardware.
- Bipolar instruments pass current only through the tissue between their tips, at low voltage and with no return pad.
- Advanced sealers add high jaw pressure, high current at low voltage and a closed loop that ends energy delivery by itself, and many are indicated for vessels up to 7 mm.
- The generator infers the seal from impedance, which falls as the tissue warms and climbs steeply as it dries.
- Burst pressure is the bench test for a seal, its widely used acceptance level of 360 mmHg is an industry convention rather than an FDA requirement, and in one comparison four of five devices lost about half their seal strength or more on vessels beyond their indicated size.
The RF generator.
+ The questionHow can one box seal a vessel at 20 Ω and sustain an arc into a load of two kilohms through the same output?
The load problem
The impedance an electrosurgical generator drives spans a hundredfold range, from about 20 Ω in a sealer's jaws to 2 kΩ or more across an arc. No designer of power supplies would choose such a load. The load is the tissue as the generator sees it, and its impedance depends on the tissue, the electrode and the contact area, and above all on what the energy is doing to the tissue at that instant. A vessel clamped wet and well compressed in a sealer's jaws looks like about 20 Ω, and tissue under a cutting electrode like about 300 Ω. A coagulation arc across a gap, or tissue that has dried, can be 2 kΩ or more. An arc's impedance does not hold steady even for a single cycle of the carrier, which lasts a few microseconds, because it changes as the arc forms and goes out within that cycle.
Manufacturers therefore specify each mode at a rated load, the impedance at which they define its power and waveform. Medtronic's rated loads for the Valleylab FT10 are 20 Ω for sealing, 100 Ω for bipolar and soft coagulation, 300 Ω for monopolar cut and 500 Ω for fulguration and spray. A rated load fixes only one point, while in use the tissue moves the load far from it. FDA's electrosurgical guidance gives 100 to 2,000 Ω, a twentyfold span, as an example of the loads across which a monopolar mode's power is plotted. A single generator must therefore cover a wide range of load with every mode it offers.
Because power is voltage squared over resistance, the RMS voltage is the square root of the product of power and resistance, and the current is voltage over resistance. Sealing at 350 W into 20 Ω needs the square root of 7,000, about 84 V, and about 4.2 A. Cutting at 300 W into 300 Ω needs 300 V, the square root of 90,000, and 1.0 A. Coagulation at 120 W into 500 Ω needs about 245 V and 0.49 A, with a peak of about 1.47 kV at a crest factor of 6. Spray at 80 W into 2,000 Ω needs 400 V and 0.2 A, with a peak of 3.2 kV at a crest factor of 8. The coagulation and spray figures are illustrative assumptions, not a particular generator's specification. Sealing therefore needs the most current, about 4.2 A, and spray the highest peak voltage, about 3.2 kV.
The four modes show the difficulty. The output must deliver more than four amperes in one mode and more than three kilovolts peak in another. The powers of the four modes differ by a factor of only about four, but their load resistances differ by a factor of a hundred, so what makes an electrosurgical output stage hard to design is the range of load, far more than the power.
The power chain
Published patents describe how a generator turns mains power into that output in stages. Covidien's patent granted in 2017 describes the core of the chain, and research designs follow the same structure. The list below combines them with earlier designs.
- The mains input filters and rectifies the incoming supply.
- A DC supply is isolated from the mains and produces a steady internal voltage.
- A DC-DC converter sets the amplitude of the output by adjusting that voltage. It is typically a buck converter, which steps a voltage down.
- An inverter switches the DC at the carrier frequency, usually 300 to 500 kHz, and creates the burst patterns of the coagulation modes by switching on and off. It is often a full H-bridge, four power transistors that can drive current through the load in either direction.
- An output network filters the switched waveform into a sine in many designs, sometimes using a resonant circuit. The Covidien design has no such network and drives the load with a square wave.
- An isolation transformer passes the output across the barrier between the internal circuits and the patient and sets the voltage ratio.
- Blocking capacitors in each output lead stop direct and low-frequency current, as Chapter 8 described.
Splitting the job this way lets each stage do one thing well. The converter controls the amplitude, which is what closed-loop control needs to adjust quickly. The inverter controls frequency and timing, and it can stay at a fixed frequency at which the output network is efficient. The transformer and the internal supply voltage together bound the voltage and current that the stage can reach, and choosing them to span both ends of the load range is the central trade-off of the design.
The earliest solid-state generators used the same ideas in a simpler form. Valleylab's patent filed in 1971 describes coagulation produced near the resonant frequency of an output circuit, at about 400 kHz, with 4,000 V across the output when no load is connected. It describes cutting produced off resonance, at a lower peak voltage.
Holding the output: three limits
A generator cannot deliver its set power at every load, because its power stage has a maximum current and a maximum voltage. Into a very low impedance it reaches the current limit first, and at that fixed current the power rises in proportion to the load. Into a very high impedance it reaches the voltage limit first, and at that fixed voltage the power falls as the load rises, in inverse proportion to it. Only between the two can the generator hold the power the surgeon set, so every electrosurgical mode has three regions: current-limited, constant power and voltage-limited.
The number on the dial is the power the generator delivers at the rated load and across its constant-power band. Outside that band it delivers less, sometimes much less. Two generators set to the same power can deliver different power to the same tissue, which is one reason clinical results do not always transfer between generators.
The closed loop works to hold the output in whichever region the tissue puts it. It measures the voltage and current at the output, computes the power and the impedance, and adjusts the converter, often thousands or hundreds of thousands of times a second. Medtronic states that the FT10's tissue-sensing technology monitors impedance 434,000 times a second. Holding the power steady is difficult, because an arc can change the load faster than a slow loop can react, and control methods for exactly this problem are an active research topic.
FDA's guidance asks for the resulting curves, and so does IEC 60601-2-2, the international standard for the safety of high-frequency surgical equipment (Chapter 25). FDA's electrosurgical guidance asks for graphs of output power at the maximum setting and at half of it, over the range of expected loads, derived from measured data rather than from theory. It also asks for the output waveform of each mode at its rated load, with its amplitude, frequency, duty cycle and crest factor, and it asks for the maximum output voltage in the labeling, the information supplied with the device.
Duty cycle and high current
A generator also has thermal limits of its own, so its output cannot stay on all the time. Medtronic states that the FT10 can work at a duty cycle of 25 %, 10 seconds active and 30 seconds inactive, in any mode for 4 hours. This activation duty cycle, the share of time the output is switched on, is a different quantity from the waveform duty cycle of Chapter 8, the share of each cycle in which the carrier is on. The two can differ widely: pure cut's waveform runs at 100 %, while the FT10's activation rating of 25 % covers every mode, pure cut included.
The activation duty cycle matters to the patient as well, because current that flows for a long time heats the tissue under the return electrode (Chapter 8). IEC 60601-2-2 therefore adds rules for outputs that deliver a lot of current for a long time. It defines a heating factor, current squared multiplied by time, and it treats a monopolar mode as a high current mode when its intended use, at maximum current and maximum activation duty cycle, gives more than 30 A²s in any 60 seconds. For the cut of the solved example, at 1.0 A, 30 A²s is 30 seconds of activation within one minute. Such a mode brings stricter requirements for the return electrode. The standard's 2023 amendment also asks for heating under the return electrode to be considered in the risk analysis for any mode whose intended use has an activation duty cycle above 45 %, even below 30 A²s.
Read a generator's datasheet mode by mode, against the tissue the device will meet. For each mode, find the rated load, the maximum power and the load at which the power-versus-load curve leaves its flat region. Find the carrier frequency and burst repetition rate, and the peak voltage and crest factor, which govern arcing, coupling and insulation stress. Find also the declared high-frequency leakage current, which Chapter 20 explains, the alarm rule of the return-pad monitor from Chapter 8, and any duty-cycle limit. Then compare the rated load with the impedance of the tissue the device will actually meet. A mode rated at 300 Ω and used mostly on tissue at 1,000 Ω may run in its voltage-limited region, at less power than its setting, and the power-versus-load curve shows whether it does.
- A generator must drive loads from about 20 Ω to 2 kΩ or more, and that hundredfold range makes its output stage harder to design than its power does.
- The power chain runs from the mains input through a DC supply, a converter that sets the amplitude and a high-frequency inverter, and in many designs an output network, to an isolation transformer and blocking capacitors.
- Every mode has current-limited, constant-power and voltage-limited regions, and the closed loop holds the output within them as the tissue changes.
- FDA's guidance asks a 510(k) to include power-versus-load curves at the maximum and half settings and each mode's waveform at its rated load, and IEC 60601-2-2 adds rules for outputs that deliver high current for a long time.
+ Part IV · Sound and light
Energy without current.
The blade of an ultrasonic shear has no sharp edge. It sits at the end of a titanium rod made a whole number of half-wavelengths long, and at the 55.5 kHz of Ethicon's Harmonic instruments each half-wavelength is about 46 mm. The blade moves 50 to 100 µm back and forth 55,500 times a second. Ultrasonic instruments and lasers both deliver their energy with no current through the patient. One vibrates a blade at a mechanical resonance, and the other turns electricity into light of a single wavelength. This part follows each of them from the power supply to the tissue, and shows how its physics sets what it does well. Each also brings hazards of its own: Chapter 11 follows the heat an ultrasonic blade keeps after activation, and Chapter 19 takes up the hazards of laser light.
Ultrasonic and hybrid instruments.
+ The questionHow does a blade that is never sharp cut tissue and seal vessels without passing any current through the patient?
From voltage to motion
Ethicon's original Harmonic transducer was a stack of piezoelectric elements held between two aluminum cylinders. Elements of this kind turn electrical power into vibration through the piezoelectric effect: certain crystals and ceramics change shape when a voltage is applied across them, and produce a voltage when they are squeezed. The movement of a single piezoelectric ring is tiny, well under a nanometer per volt, so instruments stack several rings and drive them at their mechanical resonance, where small forces build up large motion.
The standard arrangement is called a Langevin stack. A bolt through the middle clamps a few piezoelectric rings between a heavy back mass and a front mass, and it keeps the rings in compression, because ceramic cracks easily under tension. The front mass tapers into a horn, which amplifies the motion toward its narrow end.
A waveguide cut to the wavelength
From the horn, the vibration travels down a slender titanium-alloy waveguide to the blade. Vibration in a rod is a sound wave, and at the operating frequency the rod carries a standing wave. A standing wave has points that never move, called nodes, half a wavelength apart, and between them points of maximum motion, called antinodes. Patents describe waveguides whose length is a whole number of half-wavelengths. The blade tip sits at an antinode, where its motion is largest, and the supports that hold the waveguide inside its sheath sit at nodes, where there is no motion to lose.
The half-wavelength follows from the speed of the wave and its frequency, with one subtlety. A slender rod vibrates in an extensional mode, whose speed is set by the material's stiffness and density. For titanium and its alloys that speed is about 5,000 to 5,100 m/s, which is lower than the bulk speed of sound that tables quote. Half a wavelength therefore comes to about 46 mm at 55.5 kHz, the operating frequency of Ethicon's Harmonic instruments, and to about 54 mm at 47 kHz, the frequency Olympus lists for its current ultrasonic generator. The bulk speed of about 6,100 m/s would overestimate both by about 20 %.
A titanium waveguide is to vibrate at 55.5 kHz. A slender rod carries the vibration at its extensional speed, about 5,090 m/s in titanium. A wavelength is the distance the wave travels during one cycle, so it is the speed divided by the frequency: 5,090 m/s divided by 55,500 cycles a second is about 0.092 m, or 92 mm. Half a wavelength is therefore about 46 mm, and the waveguide is made a whole number of these half-wavelengths long. At the 47 kHz Olympus lists, the same division gives about 108 mm, and half of it is about 54 mm. The bulk speed of about 6,100 m/s from a table would give about 55 mm at 55.5 kHz, some 20 % too long. So at 55.5 kHz each half-wave section of a titanium waveguide is about 46 mm long.
How far the blade travels in each cycle decides how hard it works the tissue. Homma's 2022 review of the energy devices used in lung cancer surgery puts this excursion at 50 to 100 µm, and Ethicon's own patents give about 40 to 100 µm. Neither says whether the figure is the full stroke, from one end of the motion to the other, or half of it. Chapter 4 took 100 µm as the full stroke. On that reading the tip at 55.5 kHz reaches about 17 m/s and accelerates at more than 600,000 times gravity, and pressed against tissue, that motion becomes heat.
The transducer as a circuit
Electrically, a piezoelectric stack looks like a capacitor in parallel with a resonant circuit. The capacitor is the ceramic's own capacitance, called C0. The resonant branch has an inductance that stands for the moving mass, a capacitance that stands for the stiffness, and a resistance that stands for the mechanical losses, which include the tissue pressed against the blade. This equivalent circuit is the Butterworth–Van Dyke model, and patents on ultrasonic generators describe the transducer in its terms.
The model explains how the generator works. The current in the resonant branch is called the motional current, and it is proportional to the velocity of the blade. The generator therefore regulates that current to hold the blade's motion constant, and it compensates for the current that flows through C0 and does no mechanical work.
The generator also tracks the resonance, because the tissue load and the temperature shift it. The original Harmonic generator used a phase-locked loop, a circuit that adjusted the generator's frequency up or down to stay at resonance, and later patents describe tracking the frequency at which the phase between voltage and current crosses zero. Because the stack works at resonance, the drive it needs is modest. A reseller's copy of Ethicon's specification for its GEN11 generator lists the ultrasonic output as 150 V RMS maximum and 35 W continuous, about half the voltage and a ninth of the power of the 300 W cut in Chapter 10's solved example.
How it cuts and seals
Pressed against tissue, the blade heats it by friction and by internal losses, as Chapter 4 described. Proteins unfold into a sticky coagulum that seals small vessels, and with enough pressure the tissue parts along the blade. Sealing depends on compression, heat and time, the same variables as bipolar sealing (Chapter 9). Standard ultrasonic shears are indicated for vessels up to 5 mm. In October 2013, FDA cleared Ethicon's Harmonic ACE+7 for vessels up to 7 mm, in a mode that modulates the delivery of energy during the seal, and Ethicon called it the first ultrasonic device indicated for vessels that large. Olympus's THUNDERBEAT, which delivers bipolar current together with ultrasonic energy, had already been cleared for 7 mm vessels in January 2012.
Because no current flows through the patient, ultrasonic instruments avoid the stray-current hazards of Chapter 18, and that matters most near structures such as bowel. They carry a different hazard instead: the heat retained in the blade.
Heat left in the blade
Ultrasonic cutting is often described as a low-temperature technique. For the tissue, the maker agrees: Houser's patent for Ethicon, filed in 2008, states that the coagulum seals small vessels at temperatures below 100 °C. The blade is another matter. Govekar and colleagues at the University of Colorado measured it in a bench study, reported as an abstract to a meeting of the Society of American Gastrointestinal and Endoscopic Surgeons. With thermal imaging, they recorded how far the temperature rose at the tips of laparoscopic energy devices fired four times for five seconds each. The rise averaged 172 ± 63 °C for ultrasonic devices, against 81 °C for monopolar and 46 °C for bipolar devices. Bovine liver that the monopolar and bipolar devices touched after activation was within 20 °C of its starting temperature by five seconds, while liver that the ultrasonic devices touched stayed more than 20 °C above it for the full 20 seconds observed.
Both descriptions are true, because they refer to different things. The tissue at the cut stays moderate because its water holds the temperature near 100 °C and the tissue is removed. The metal has no water and a small heat capacity, and it keeps its heat after the vibration stops. A blade fired with no tissue in its jaws has nothing to absorb its energy, so it runs hotter still. A hot blade that touches bowel on its way to the next bite causes a burn as surely as a current would.
The absence of current through the patient removes one class of hazard and leaves another. For ultrasonic instruments, the thermal hazard sits in the blade after activation, not in the tissue during it. Instructions for use, training and jaw design all address it, and so should any comparison of devices.
Hybrid instruments and aspirators
A hybrid instrument combines ultrasonic and bipolar energy in one jaw. Olympus launched THUNDERBEAT in March 2012, after FDA cleared it in January, and describes it as delivering ultrasonic and bipolar energy to the tissue at the same time. The two energies share the work. Bipolar current heats the tissue from within, while friction at the probe heats it from the surface and divides it, so one bite can seal and cut.
The current version illustrates how such devices evolve. FDA cleared THUNDERBEAT II on 7 October 2025. The summary Olympus filed for it lists ultrasonic output at 47 kHz and bipolar output at 380 kHz. It describes three modes: Seal & Cut with both energies, Seal with bipolar only, and an ultrasonic-only mode indicated for vessels up to 3 mm, against 7 mm for the two sealing modes. The generator of THUNDERBEAT II, the USG-410, was first cleared in 2021 and is listed again in the 2025 submission. Olympus calls the new transducer cordless, which here means that the cable has moved to the base of the handle, not that the device runs on a battery. Olympus also describes a thermal shield on the jaw that slows the transfer of heat to its outer surface.
A second family of ultrasonic instruments, the aspirators, does not seal at all. Vieira-Coimbra and Ferreira describe how they work in a prospective case series, published in 2025, on the laparoscopic excision of endometriosis with the CUSA aspirator. Integra LifeSciences, the maker of CUSA, supported its publication. An aspirator vibrates a hollow tip at 23 or 36 kHz and fragments the tissue just in front of it, mainly by the cavitation of Chapter 4 together with direct mechanical impact, while irrigation and suction carry the fragments away. Water-rich, weakly cohesive tissue such as liver parenchyma or tumor breaks up, while collagen-rich vessels, ducts and nerves survive to be clipped or sealed. The selectivity comes from the mechanics of the tissue rather than from heat.
Specify and verify an ultrasonic instrument against three properties from this chapter. First, size the waveguide from the extensional speed of the alloy actually used, because the bulk speed in a table makes every half-wavelength about 20 % too long. Second, measure the blade's temperature after activation and the time it takes to cool, because that retained heat is the hazard an instrument without current through the patient still carries. Third, state the vessel size for each mode separately, because one instrument can carry different limits, as THUNDERBEAT II does with 7 mm in its sealing modes and 3 mm in its ultrasonic-only mode.
- A Langevin stack of piezoelectric rings drives a titanium waveguide at resonance, and the waveguide is a whole number of half-wavelengths, about 46 mm each at 55.5 kHz, with the blade tip at an antinode.
- The generator regulates the motional current, which is proportional to the blade's velocity, and tracks the resonance as the tissue load and the temperature shift it.
- The blade heats tissue by friction and internal losses, seals vessels up to 5 mm, or 7 mm for newer designs, and parts the tissue along its edge with no current through the patient, but it keeps its heat after activation and can burn tissue it touches afterward.
- Hybrid instruments combine bipolar and ultrasonic energy in one jaw, and aspirators fragment tissue mechanically, mainly by cavitation, while sparing vessels.
Surgical lasers.
+ The questionChapter 5 showed that the wavelength decides where light is absorbed. What does it take to make light of one wavelength, and how does that choice shape a laser's whole design?
How a laser makes light
A photon of exactly the right energy passes an excited atom, and in that instant it can trigger the atom to emit a second photon identical to the first, with the same wavelength, the same phase and the same direction. This is stimulated emission, and each such event turns one photon into two. Atoms and molecules hold energy in discrete levels, and they absorb or emit light when they move between them. Normally most of them are in their lowest level, so light passing through a material is more likely to be absorbed than to trigger emission, and a laser has to reverse that balance.
The material that amplifies the light is called the gain medium. A pump lifts its atoms into an excited state faster than they can fall back, until more of them are in the upper laser level than in the level below it. This state is a population inversion. The pump may be a flashlamp, a diode laser or an electrical discharge. Mirrors at either end of the gain medium send the light back and forth through it, so the light is amplified on every pass. One mirror lets a small fraction through, and that fraction is the beam. Lasing begins when the gain on each pass exceeds the losses.
How the pump is made largely sets how much of the mains power ends up as light. A holmium laser pumped by flashlamps converts roughly 1 to 2 % of its electrical input into light, because the lamp's broad spectrum overlaps the absorption of holmium only slightly, and the rest becomes heat that water cooling must remove. A thulium fiber laser pumped by diodes whose output matches thulium's absorption converts about 12 % and can be air-cooled. In carbon dioxide lasers an electrical discharge through the gas does the pumping, and they reach 10 to 20 %.
The difference shows on the nameplate. The manufacturer of one 120 W holmium system lists a single-phase supply of 200 to 240 V at up to 46 A and a weight of 240 to 260 kg, depending on the version. A 60 W thulium fiber laser from another manufacturer runs from 100 to 240 V at 1,200 VA and weighs about 40 kg, about a sixth as much. The optical output divided by the electrical input comes to about 1.1 to 1.3 % for the holmium system and at least 5 % for the thulium laser, which is consistent with the published efficiencies.
Lasers and their jobs
Chapter 5's absorption curves decide what each laser is for.
The CO₂ laser emits at 10.6 µm, and tissue absorbs that light within about 16 µm, so the laser vaporizes a thin layer precisely, with little coagulation beneath it. It is used for cutting and vaporizing soft tissue, and in airway surgery, where precision near delicate structures matters. Tissue absorbs the light of the Ho:YAG laser, at 2.1 µm, within about 0.5 mm, and the light of the thulium fiber laser, at 1.94 µm, within about 0.1 mm. Both lasers therefore cut and coagulate together, thulium in a shallower layer. Both also work well under water, which is why they are used to treat stones and to enucleate an enlarged prostate.
Water also absorbs the light of the 1,470 nm diode laser within about half a millimeter, and this laser is used, for example, to close varicose veins from inside with a radial fiber. The light of the 980 nm diode and of the Nd:YAG laser at 1,064 nm is weakly absorbed and scatters over millimeters, so these lasers coagulate deeply. The 532 nm green laser is made with a KTP or LBO crystal. Hemoglobin absorbs its light and water lets it pass, which suits the vaporization of a blood-rich prostate under irrigation.
Getting the light there
The wavelength also decides how light reaches the tissue. Silica optical fibers carry light from the visible to about 2 µm with little loss, so every laser above except CO₂ can be delivered through a flexible fiber small enough to pass down an endoscope. The size of the fiber's core matters too. Holmium lasers need fibers with a core of 200 µm or larger, because their beam quality limits how tightly they can be focused into a fiber, while thulium fiber lasers couple into fibers as small as 50 to 150 µm. The GreenLight XPS system for the prostate uses the MoXy fiber, a side-firing fiber with a 750 µm core that directs the beam sideways at the tissue.
Silica absorbs light at 10.6 µm strongly, so it cannot carry a high-power CO₂ beam. CO₂ lasers are therefore delivered by an articulated arm of mirrors, or through hollow waveguides made for the purpose.
Pulses, bubbles and stones
Chapter 5 showed that the length of a pulse sets how far its heat spreads and how high its peak power climbs, and lasers for lithotripsy, the breaking of stones, use both effects. A holmium pulse delivers about 1 J in roughly 300 µs, a peak of about 3.3 kW. A thulium fiber laser works at a peak of about 500 W, so the same joule takes about 2 ms, and the laser compensates with far higher repetition rates. The manufacturer of one current thulium system states that it offers pulse energies from 0.025 to 6 J at up to 2,400 Hz.
Under water, a laser meets an obstacle before it reaches the stone, because the water itself absorbs the light. The front of each pulse therefore boils a vapor bubble at the fiber tip, and the light that reaches the stone is the light that passes through that bubble. Since 2017, pulse-shaping modes have exploited the bubble by firing a short, low-energy sub-pulse to open it and then the main pulse through it, so more of the energy reaches the stone. The manufacturer of one such mode claims 50 % less stone movement and a 33 % higher ablation rate.
Figures such as the claimed 50 % less stone movement come from manufacturers' own testing, under conditions they chose. They describe the mode under those conditions and are not a measure of clinical benefit in a particular patient or stone type. Independent comparisons, with the same stones, fibers and settings, are the evidence to look for.
Pulse settings then decide how a stone breaks. Two published sources give the settings: a 2017 review by Aldoukhi and colleagues and a 2018 doctoral thesis by Hardy on thulium fiber laser lithotripsy. The review's settings are for a 120 W Lumenis holmium laser, and one of its authors, Ghani, consulted for Lumenis and Boston Scientific when it was published. For dusting, between them, the two sources give low pulse energies within about 0.2 to 0.5 J at 40 to 80 Hz, which erode the stone into fine particles that can be washed out. For fragmentation, they give higher energies within about 0.5 to 2 J at 5 to 30 Hz, which break the stone into pieces to be removed with a basket. Longer pulses, lower energies and smaller fibers all reduce the recoil that pushes a stone away from the fiber.
Between them, the review and the thesis above give settings within about 0.5 to 2 J at 5 to 30 Hz for fragmentation and about 0.2 to 0.5 J at 40 to 80 Hz for dusting. The average power is the energy of each pulse multiplied by the number of pulses each second. One setting inside each range shows the difference. A fragmentation setting of 1 J at 20 Hz gives 1 J times 20, or 20 W, and a dusting setting of 0.5 J at 80 Hz gives 0.5 J times 80, or 40 W. The peak power is the energy of one pulse divided by the time it lasts. If the fragmentation pulses are holmium pulses of 1 J lasting roughly 300 µs, or 0.0003 seconds, each one peaks at about 3,300 W, about 170 times the 20 W average. At these two settings, dusting therefore runs at twice the average power of fragmentation with pulses of half the energy, and the average power of a setting says little about the peak each pulse reaches.
A laser's wavelength decides what absorbs it, how deep the energy goes, which fibers can carry it, and, through the pump that makes it, how much mains power it draws and how it must be cooled. Choosing a wavelength is choosing most of the system.
Treat the wavelength as the first requirement of a laser system, and derive most of the rest from it. Carry the pump's efficiency into the mains supply, the cooling and the weight, because a flashlamp-pumped holmium laser turns less than 2 % of its input into light and needs water cooling, while a diode-pumped thulium fiber laser can be cooled by air. Check that silica fiber can carry the wavelength, and which core sizes the beam can enter. Specify each pulse setting by its energy, rate and duration together, because the average power follows from the energy and the rate, and the peak power from the energy and the duration.
- A laser pumps a gain medium into population inversion and amplifies stimulated emission between mirrors, and the way it is pumped sets its efficiency: under 2 % for flashlamp-pumped holmium, about 12 % for thulium fiber and 10 to 20 % for CO₂.
- The wavelength decides the job: CO₂ for precise vaporization, holmium and thulium for cutting under water and for stones, 980 nm diode and Nd:YAG for deep coagulation, 1,470 nm diode for vein closure, 532 nm for blood-rich tissue.
- Silica fibers carry every surgical wavelength up to about 2 µm, while CO₂ needs mirrors or hollow waveguides.
- Under water each pulse must first boil a vapor channel, and pulse shaping and pulse settings decide how much energy reaches a stone and whether it dusts or fragments.
+ Part V · Destroying tissue in place
Ablation.
A 2025 study of current cryoablation needles found that, around a single needle, the region cooled to −40 °C or below spans only 44 to 60 % of the width of the visible ice ball. Part of what the operator sees freezing may therefore survive. Resection removes a tumor, while ablation kills it where it lies, through a needle or from outside the body, and leaves the dead tissue for the body to clear. This part covers heating with current and microwaves, freezing, pulsed electric fields and focused sound. Each energy has its own answer to the same two questions: how large a volume it can kill, and what happens at the edge of that volume.
RF and microwave ablation.
+ The questionAn RF needle and a microwave antenna of similar size both heat tumors. Why does the microwave antenna keep heating where the RF needle stalls, and what decides the size of the zone each one makes?
Killing a volume in place
In ablation, every cell in a volume a few centimeters across must be held above about 60 °C long enough to die, while the tissue outside it is spared. The volume is a sphere or an ellipsoid that contains the tumor and a margin of normal tissue around it, and the volume a treatment kills is called its ablation zone, or simply its zone. The device is usually a needle placed through the skin under ultrasound or CT guidance, so the heat must spread outward from a thin probe into a volume many times its size.
Ablation therefore asks for something different from cutting or sealing, and Chapter 2's physics sets the difficulty: heat conducts slowly, blood carries it away, and large vessels act as heat sinks. RF and microwave ablation both heat, but they take different approaches to the problem, and the difference comes down to where each one deposits its power.
RF ablation: heat in a thin shell
An RF ablation needle is an electrosurgical electrode used in a different way. The generator passes current from the needle's exposed tip through the tumor to return pads on the patient's skin. It typically operates at about 500 kHz. Makers have specified maximum outputs of 200 W for Boston Scientific's RF 3000 generator and 250 W for AngioDynamics' Model 1500X, part of the RF ablation line that AngioDynamics discontinued in 2024 (Chapter 23). The heating follows Chapter 3: current density falls steeply away from the needle, so almost all the power is deposited in a shell a few millimeters thick around the tip. Conduction from that shell heats the rest of the zone, over minutes.
The approach has a built-in limit. Tissue touching the needle is the hottest, and as it reaches 100 °C it dries and chars. Dry, charred tissue conducts poorly, so the impedance at the electrode rises sharply, the current falls and the generator can no longer deliver power. This is impedance roll-off, and when it happens the zone stops growing.
Most of the engineering of RF ablation is about delaying roll-off. Internally cooled electrodes circulate chilled saline inside the needle and keep the tissue at its surface below boiling, so that it does not char. The hottest point then moves a few millimeters into the tissue. An isotherm is a surface on which the temperature is the same everywhere, and the lethal isotherm is the one at the temperature that kills tissue, about 60 °C here, which makes it the edge of the zone. The lethal isotherm moves outward with the hottest point.
Multi-tine electrodes take another approach: they deploy an array of curved prongs from the needle, which spreads the current over a larger volume. Boston Scientific's 2015 brochure lists multi-tine arrays that open to diameters of 2 to 5 cm, and AngioDynamics states that its generator and electrodes ablate zones of 1 to 7 cm. Generators control the output by temperature or by impedance, and the impedance method uses roll-off itself as its signal. AngioDynamics describes automatic temperature control for its Model 1500X, while Boston Scientific's RF 3000 reads the end of an ablation from a rise in impedance. The solved example below takes zones of 2.5, 4 and 5 cm, all within these ranges, to show how quickly the volume grows with the diameter.
If an ablation zone is treated as a sphere, its volume is the diameter times itself times itself again, then times about 0.52, which is π divided by 6. A 2.5 cm zone holds 2.5 times 2.5 times 2.5, or about 15.6, times 0.52, which is about 8 cm³. A 4 cm zone holds 64 times 0.52, about 33 cm³, and a 5 cm zone holds 125 times 0.52, about 65 cm³. A zone 25 % wider, 5 cm against 4, therefore holds almost twice the tissue, and doubling the diameter from 2.5 to 5 cm multiplies the volume by eight. Going from 4 to 5 cm adds about 32 cm³, almost as much as the whole 4 cm zone holds.
Microwave ablation: heat in a volume
A microwave antenna does not need current to flow through tissue. It radiates an electromagnetic field at 915 MHz or 2.45 GHz, and the field acts directly on water molecules. Each water molecule is a small electric dipole, slightly positive at one end and slightly negative at the other, so in an alternating field it turns back and forth to follow the field, billions of times a second. The molecule lags behind the field and jostles its neighbors as it turns, and the energy lost to that lag appears as heat. The same field pushes ions in the tissue fluid back and forth, and they add losses of their own. Together these make up dielectric heating, the physics of a microwave oven (Chapter 6).
The field penetrates tissue far beyond the antenna's surface. In liver at 2.45 GHz, the wavelength is about 1.85 cm, against 12.2 cm in free space, and the power falls to about 37 % of its value within about 1 cm. At 915 MHz the wavelength is about 4.7 cm, and the power takes about 2 cm to fall the same way. These figures are for a plane wave, the flat wave found far from a source, and close to a thin antenna the power falls faster as the field spreads. Power is therefore deposited through a volume a centimeter or two around the antenna rather than in a millimeter-thick shell, so the tissue heats faster.
Two reviews describe microwave zones as faster and hotter than RF zones, often above 100 °C, and some of the authors of each held equity in NeuWave Medical, the original maker of the NEUWAVE microwave system described below. The field also continues to heat tissue that has dried. Dry tissue absorbs microwaves less well, because it contains less water, but no conduction path is needed, so there is no roll-off in the RF sense. Microwaves also work in tissues that conduct poorly, such as lung and bone, and they need no return pad. Several antennas can run at once without competing for current, and one of those reviews reports that simultaneous antennas produced zones nearly six times the volume of a single one.
Current systems show the engineering choices. Medtronic states in its summary for FDA that its Emprint system runs at 2.45 GHz with 5 to 150 W and cools the shaft of its antenna with circulating saline. NeuWave Medical, the original maker of what is now Ethicon's NEUWAVE system, described it to FDA in 2013 as having three independent amplifiers of up to 140 W each and cooling its thin probes with carbon dioxide. Both systems are cooled because the coaxial cable and the antenna themselves heat as they carry the power, and that heat would otherwise damage tissue along the needle track. Hui and colleagues measured what such a system can make in a 2023 bench study that compared Emprint HP with other current microwave devices in excised pig liver, and Medtronic funded the study. At 150 W for ten minutes, Emprint HP produced zones about 5.3 cm long and 4.8 cm wide.
The choice between the two frequencies is a trade. Simo and colleagues compared 915 MHz and 2.45 GHz systems in a 2013 clinical study, a retrospective analysis of 48 patients with 124 tumors treated at one center. They found similar results, but the 2.45 GHz system needed a mean of 4.0 minutes per ablation against 8.1, and its ablation times followed tumor size more closely, which the authors read as more predictable ablation. The two systems also differed in power and antenna design: the 915 MHz system used one to three antennas at 45 W, and the 2.45 GHz system a single antenna at 100 W. Not all of the difference is therefore due to frequency. The higher frequency allows shorter antennas and more spherical zones, but its cable losses are higher.
The heat sink, revisited
Chapter 2's heat sink affects RF ablation most, because RF relies on conduction for most of its zone. Microwave heating is less affected, because it is deposited directly through a volume, but it is not immune. A 2015 study in excised pig liver measured the effect. A glass tube 5 mm across stood in for a vessel, and saline flowed through it at 700 mL per minute. With the antenna 10 mm from the tube, the radius of the microwave zone was 37 % of its size without flow, and with the antenna 15 mm away it was 80 %. For both energies, tumors against large vessels remain the hardest to ablate completely.
The heat-sink advantage of microwave ablation is real but partial. The study above used a 915 MHz system at 45 W, and no data for current high-power 2.45 GHz systems near perfused vessels were found for this guide. Planning should still allow for vessels near the target.
Judging the result
Neither RF nor microwave zones can be seen reliably while they form, because gas from boiling tissue obscures ultrasound and the heated tissue contracts, which distorts the zone's apparent size. The result is judged afterward on contrast-enhanced CT or MRI, where the dead zone no longer takes up contrast. The margin of dead tissue around the tumor on that scan predicts whether the tumor will recur locally, which is why ablation reports record it.
Read a manufacturer's zone chart as a measurement in excised tissue without perfusion, at a stated power and time. It gives the largest zone a system can make, not the zone it will make next to a vessel in a living liver. Compare systems in the same tissue, at the same power and time, and ideally in perfused models. Then check that the zone covers the tumor plus a margin, not the tumor alone.
- RF ablation deposits power in a shell a few millimeters thick around the needle and relies on conduction for the rest, and charring at the needle raises the impedance and stops the zone growing.
- Cooled and multi-tine electrodes delay roll-off and let the zone grow to a few centimeters.
- Microwaves heat water directly through a volume about a centimeter deep, keep heating dried tissue and need no return pad, and reviews describe their zones as faster and hotter.
- Both energies make smaller zones next to large vessels, and the zones are judged after treatment on contrast-enhanced imaging.
Cryoablation.
+ The questionEvery energy so far adds heat. How does taking heat away kill tissue, and why is the edge you can see not the edge that dies?
Cooling by expansion
A cryoablation probe is a needle a couple of millimeters wide, and its tip must get far colder than anything else in the operating room without freezing the needle's shaft or the skin it passes through. Most systems do this with a property of real gases. When a gas under high pressure expands through a narrow orifice without doing work on anything, its temperature changes, and this is the Joule–Thomson effect. In most gases at room temperature, including argon and nitrogen, attraction between the molecules dominates, so pulling them apart as the gas expands takes energy from their motion, and the gas cools. In helium at room temperature repulsion dominates, and the gas warms.
In an argon cryoprobe, gas from a high-pressure supply flows down a thin capillary inside the needle to its tip. There it expands through an orifice into the closed chamber of the tip and cools sharply. The cold gas absorbs heat from the surrounding tissue through the wall of the tip, then flows back along the space between the capillary and the outer wall and vents outside the patient. Argon boils at about −186 °C, which sets the floor. Hinshaw and colleagues report that argon systems cool the probe to −160 °C or colder. Their 2014 review compares microwave, RF and cryoablation tools, and three of its authors held shares in NeuWave Medical, the original maker of the NEUWAVE microwave system (Chapter 13).
The same effect can also run the other way. Because helium warms when it expands, some systems thaw the tissue by switching the probe from argon to helium. Liquid nitrogen gives a second way of cooling, one that avoids high-pressure gas altogether. Liquid-nitrogen systems circulate the liquid under pressure to the tip, where it boils, at −196 °C at atmospheric pressure and somewhat warmer under pressure. IceCure makes one such system, ProSense, and FDA's decision summary for it gives a minimum of −170 °C at the center of the probe's cooling zone.
How cold kills
Freezing kills cells in several ways at once. As tissue cools, ice forms first in the fluid between the cells. The ice excludes salt, so the remaining fluid becomes more concentrated and draws water out of the cells by osmosis, and the cells shrink and their contents are damaged. If the cooling is fast, water has no time to leave, and ice forms inside the cells themselves. This intracellular ice is far more damaging than the extracellular ice between the cells, and at the probe, where cooling is fastest, it is the main cause of death.
Farther from the probe, cooling is slower and the temperature higher, and cells may survive. There, injury depends on what happens next. During thawing, small ice crystals recrystallize into larger, more damaging ones, and water rushes back into the shrunken cells. The small vessels of the frozen zone are injured, and in the hours after the procedure the tissue loses its blood supply. A 2010 review of cryoablation by Erinjeri and Clark lists four factors that together decide the injury: the cooling rate, the lowest temperature reached, the time held there, and the thawing rate.
The ice ball and its isotherms
A frozen zone has one great advantage over RF and microwave zones: it can be seen. Ice shows on ultrasound, CT and MRI, so the operator watches the zone grow in real time. The difficulty is that the visible edge of the ice ball is a surface of equal temperature, the 0 °C isotherm, and cells at 0 °C survive. Both reviews place reliable cell death below about −20 to −40 °C. Hinshaw and colleagues estimate the lethal temperature at −20 to −35 °C and suggest that cell death may not occur until about 8 mm inside the edge of the visible ice ball.
The ice ball's edge is the easiest thing in ablation to see and the easiest to misread. It marks 0 °C, not death. A plan that covers the tumor with visible ice, but not with the lethal isotherm of Chapter 13, leaves a rim of living tumor inside what looked like a successful freeze.
A 2025 study measured the −40 °C isotherm of current needles against the visible ice. With a single needle, the isotherm spanned 44 to 60 % of the width of the visible ice ball, depending on the system, and clustering several needles raised the fraction. The practical rule follows: the visible ice must extend well beyond the tumor, by at least the distance between the visible edge and the lethal isotherm, if the whole tumor is to die.
A tumor 15 mm across, the largest breast cancer ProSense is authorized to treat, is to lie inside the −40 °C isotherm, before any margin of normal tissue is added. If that isotherm spans 60 % of the visible width, the best case the 2025 study found for single needles, the visible ice must be 15 divided by 0.60, or 25 mm wide. At 44 %, the worst case, it must be 15 divided by 0.44, or about 34 mm wide. Hinshaw and colleagues' figure of about 8 mm inside the visible edge gives a similar width by another route: 8 mm on each side of the tumor makes 15 plus 16, or 31 mm. Either way the visible ice must be roughly twice as wide as the tumor. The ProSense breast protocol targets more than 35 mm after the first freeze and more than 40 mm after the second.
Protocols usually freeze twice with a thaw between, because the second freeze meets tissue already damaged by the first and extends the lethal zone toward the edge of the ice. The whole sequence takes time: Hinshaw and colleagues give typical ablation times of about 25 to 30 minutes. For the breast, FDA's decision summary for ProSense gives a default of nine minutes of freezing, eight of thawing and nine more of freezing, 26 minutes in all, with target ice-ball widths of more than 35 mm after the first freeze and more than 40 mm after the second.
What cold gives and takes
Cryoablation's strengths follow from its physics: the zone is visible as it forms, there is no current through the patient and no return pad, and several probes can be combined to shape the ice around an irregular tumor. Its weaknesses follow too. Cold does not coagulate blood or cauterize injured vessels, so frozen tissue has no built-in hemostasis, the stopping of bleeding, and Erinjeri and Clark note that cryoablation can therefore worsen bleeding. RF ablation, by contrast, coagulates damaged vessels as it goes. Blood flow also works against cold as it works against heat. Large vessels warm the ice near them, acting as a heat source rather than as the heat sink of Chapter 2, so they can change the size and shape of the ice ball. Argon systems also need high-pressure cylinders and gas lines, which the makers of liquid-nitrogen systems cite as part of the case for their own.
Cryoablation reached a new indication in the US in 2025. On 3 October 2025, FDA granted IceCure's ProSense a De Novo authorization for the local treatment of biologically low-risk breast cancers of 1.5 cm or less in patients aged 70 or over who receive adjuvant endocrine therapy. A De Novo is the route FDA uses for a new kind of device of low or moderate risk when no device already on the market can serve as its comparison.
The supporting trial was IceCure's own ICE3 study, a single-arm study with no comparison group. FDA's analysis of it reported ipsilateral recurrence rates of 2.3 to 8.7 %, depending on the analysis population, while IceCure reports a five-year estimate of 3.1 %. An ipsilateral recurrence is the cancer's return in the same breast. The figures differ, and each is correct for its own population and method.
Ask of every zone size in a cryoablation system's specification which isotherm it refers to, because a "4 cm ice ball" and a "4 cm lethal zone" are very different claims. Plan from zone charts that give the −20 °C and −40 °C isotherms, measured for single and for clustered needles, because a chart of the visible ice alone does not show which tissue will die. On the bench, measure temperatures inside the ice as well as at its edge, so that a stated lethal zone rests on a measured isotherm.
- Cryoprobes cool by the Joule–Thomson expansion of high-pressure argon or by circulating liquid nitrogen, and helium can thaw the tissue.
- Freezing kills by extracellular ice and osmotic injury, by intracellular ice where cooling is fast, and by injury to the small vessels after thawing.
- The visible edge of the ice is at 0 °C and reliable cell death lies below about −20 to −40 °C, several millimeters inside it, so the ice must extend beyond the tumor by at least that distance.
- Cryoablation gives a visible zone and no current through the patient, but no hemostasis, and its newest US indication is low-risk breast cancer in patients aged 70 or over.
Pulsed fields: IRE and PFA.
+ The questionHow can kilovolt pulses kill cells without heating them, and why would anyone want an ablation that does not use heat?
The problem heat cannot solve
On 13 December 2023, FDA approved Medtronic's PulseSelect, the first system in the US that treats atrial fibrillation, a disorder of the heart's rhythm, with strong electric pulses instead of heat or cold. Such a system answers a problem that heat cannot solve, because every energy so far kills by temperature, and temperature does not discriminate. Three targets show the problem: a tumor wrapped around the portal vein, a prostate cancer beside the nerves and sphincter that control erection and continence, and the wall of the left atrium lying against the esophagus. In each, the cells to be killed sit next to structures that must survive, and heat destroys both alike. The collagen framework of a vessel or duct survives many other insults, but the temperatures that ablation reaches destroy it along with the cells around it.
Pulsed electric fields offer a different mechanism, in which very short, very strong electric pulses open holes in cell membranes. If the holes are large and numerous enough, the cell cannot repair them and dies, while the scaffold of collagen and elastin between the cells is left intact, because it has no membrane to open. In tumor ablation the technique is called irreversible electroporation, or IRE, and in cardiology it is called pulsed field ablation, or PFA.
A voltage across the membrane
Chapter 3 described a cell as a conducting interior wrapped in an insulating membrane. In an external electric field, charge piles up on that membrane, positive on one side and negative on the other. Within about a microsecond, the voltage across the membrane settles at a value that depends on the field, the cell's size and the position on the cell. For a spherical cell, the voltage is largest at the two poles that face the field, where it equals 1.5 times the field strength times the cell's radius. It falls away from the poles and is zero at the equator, midway between them.
A spherical cell with a radius of 5 µm sits in a field of 400 V/cm. In meters, the field is 40,000 V per meter and the radius is 0.000005 m, five millionths of a meter. The voltage at each pole is 1.5 times the field times the radius, and 1.5 times 40,000 V/m times 0.000005 m gives 0.30 V. The voltage rises in step with the field, so 600 V/cm gives 0.45 V and 1,000 V/cm gives 0.75 V. Membranes are reported to open pores once the voltage exceeds about 0.5 to 1 V, and reaching 0.5 V at this cell's poles takes 0.5 V divided by 1.5 and by 0.000005 m, about 67,000 V per meter, or about 670 V/cm. A cell with twice the radius reaches 0.5 V at half that field, so a larger cell crosses the threshold first.
Membranes are reported to open pores once the voltage across them exceeds about 0.5 to 1 V. Above that threshold, the membrane's lipid layer rearranges into nanometer-scale pores. After a brief, mild exposure the pores reseal and the cell survives: this is reversible electroporation, which is used to push drugs or DNA into cells. After stronger or more numerous pulses the damage is beyond repair, and the cell dies over the following hours. Because the voltage at the poles grows with the cell's radius, a larger cell reaches the threshold at a lower field than a smaller one.
Thresholds and the shape of the zone
The field around a pair of needle electrodes is strongest between and immediately around them and falls off with distance, so the ablation zone is the region where the field exceeds the irreversible threshold for that tissue. That threshold differs from one tissue to another. In a 2023 laboratory study funded by Medtronic, Kos and colleagues compiled the mean irreversible thresholds that earlier studies had measured in living tissue: 539 V/cm for liver, 538 for kidney, 585 for brain, 700 for prostate and 1,050 for bone. Three of the study's authors were Medtronic employees. Below the irreversible threshold lies the reversible one, at which pores open and then reseal, and a 2005 modeling study by Davalos and colleagues used about 360 V/cm for it in liver. The thresholds fall as more and longer pulses are delivered.
Clinical settings keep the field well above these values. In 2016, Martin and colleagues called for a standard way of delivering IRE energy in locally advanced pancreatic cancer, and they report that surgeons typically apply 1,400 to 2,000 V for each centimeter between electrodes. The electrodes are usually placed no more than 2 cm apart, so that the field across the target stays above threshold with margin. The first author was a paid consultant for AngioDynamics, the maker of the NanoKnife IRE system, when the study was published.
Because the mechanism is electrical rather than thermal, flowing blood does not carry the effect away. There is no heat sink of the kind Chapter 2 described, and the zone reaches right up to the wall of a vessel crossing it. Animal studies report that vessels, bile ducts and bronchi within the zone keep their structure. The mechanism predicts this, because their framework of collagen and elastin has no membrane to open. That preservation is the reason to use IRE, because it lets IRE treat tissue that heat-based ablation cannot safely reach.
IRE pulses carry a lot of energy: one maker's manual for its IRE generator lists outputs of 500 to 3,000 V and up to 50 A, in pulses of 20 to 100 µs carrying up to about 15 J each. Every pulse deposits some Joule heat (Chapter 3). Near the electrodes, after many pulses, tissue can heat enough to add thermal injury. Non-thermal describes the mechanism that kills most of the zone, not an absence of heating.
Pulses and the heart
Pulses of thousands of volts stimulate every nerve and muscle they reach and can trigger the heart, so IRE needs two precautions. The pulses are synchronized with the ECG, the trace of the heart's electrical activity, and delivered in the refractory period just after each heartbeat, when the heart cannot be triggered. The generator whose manual is cited above uses this synchronized mode by default and also offers fixed rates of 90 pulses a minute, for soft tissue below the waist, and 240, for the prostate. And because the pulses make skeletal muscle contract violently, patients are treated under general anesthesia with complete neuromuscular block, in which drugs stop the skeletal muscles from contracting.
Cardiac PFA addresses the same problem differently, with short pulses of a different shape. Each pulse can make nearby skeletal muscle twitch, so pulse shape matters in practice. Pulses that reverse direction, called biphasic, appear to reduce this. In a 2022 preclinical study, Bi and colleagues found less muscle contraction with biphasic pulses than with long monophasic pulses, which drive current in one direction only. One of the authors worked for a company, Tianjin Intelligent Health Medical. The pulses remain strong. One present-day system is Boston Scientific's FARAPULSE, and its maker states that it offers outputs of 1.8 to 2 kV in applications of 2.5 seconds each.
Pulsed field ablation for atrial fibrillation
Pulsed fields are being adopted rapidly in cardiology. Atrial fibrillation is treated by creating lines of scar around the pulmonary veins, and for decades that was done with RF heat or with cold. Both carry a small risk of injuring the esophagus, which lies against the back of the left atrium, and the phrenic nerve, which runs near the right-sided veins. The phrenic nerve drives the diaphragm, and an injured one can leave the diaphragm paralyzed, a phrenic palsy. Pulsed fields can spare such structures if heart muscle dies at a lower field than they do.
Heart muscle has a comparatively low irreversible threshold, below that of every tissue in the list above, although the list includes neither the esophagus nor the phrenic nerve. In the same Medtronic-funded study that compiled that list, Kos and colleagues measured a median of 416 V/cm in three human donor hearts, kept perfused outside the body and treated with Medtronic's biphasic pulses. A proposed explanation for the selectivity of PFA is the size of the cells. The solved example showed that a cell with a larger radius reaches the pore-opening voltage at a lower field, and heart muscle cells have a larger radius than the smooth-muscle cells of the esophagus, so heart muscle crosses the threshold first.
After PulseSelect, FDA approved Boston Scientific's FARAPULSE on 30 January 2024, Medtronic's Affera Sphere-9 on 24 October 2024 and Johnson & Johnson's Varipulse on 6 November 2024. Sphere-9 delivers both pulsed field and RF energy. In January 2025, Johnson & Johnson paused US cases with Varipulse after neurovascular events, which affect the blood supply of the brain, and in February it announced a limited restart with updated instructions for use (Chapter 26). Later in 2025, FDA approved two more pulsed field systems, Kardium's Globe on 27 August and Abbott's Volt on 19 December.
MANIFEST-17K is a retrospective registry, compiled from the records of patients already treated, and many of its authors report ties to Boston Scientific. In it, researchers at 106 centers in Europe and Israel reported on 17,642 patients treated with one system, Boston Scientific's FARAPULSE. They recorded no esophageal complications and no persistent phrenic nerve palsy, and major complications in 0.98 % of patients.
Measures of adoption depend on what is counted. A market-intelligence firm modeled the purchasing data of more than 4,500 US facilities and estimated that PFA accounted for 69 % of spending on atrial fibrillation ablation catheters in 2025 and 78 % in early 2026, but for 48 % and 57 % of the catheters bought. A physician survey projected that PFA would be used in 49 % of procedures in 2025. All three figures can be accurate, because they measure different things.
IRE for tumors has moved more slowly, and its US change of December 2024 concerns its indication. An indication is the use a maker claims for its device, and a clearance or approval covers the indication on the label (Chapter 2). AngioDynamics' NanoKnife had long been cleared by 510(k) for the surgical ablation of soft tissue. On 9 December 2024, AngioDynamics announced a clearance that extended that indication to name prostate tissue. NanoKnife's US indication names the tissue rather than a disease, a distinction that matters in reading any device's label.
Treat the field map as the design of a pulsed-field device. The electrode geometry, the pulse amplitude, width, number and polarity, and the tissue's threshold together decide where the zone ends, and the heating near the electrodes decides whether the zone stays purely non-thermal. Verify the system by measuring zone size and thermal effects across the range of pulse settings, and not only the generator's electrical output.
- At a cell's poles, an external field induces a membrane voltage of about 1.5 times the field times the radius, pores open above about 0.5 to 1 V, and after enough pulses the cell cannot repair them.
- Irreversible thresholds are several hundred volts per centimeter and vary by tissue, and the zone is wherever the field exceeds the threshold, with no heat sink to carry the effect away.
- In animal studies, the collagen scaffold of vessels and ducts survives, which is the reason to use pulsed fields, although every pulse still deposits some heat near the electrodes.
- IRE pulses are timed to the heartbeat and given under complete neuromuscular block, while cardiac PFA uses biphasic pulses and has been adopted rapidly since FDA first approved a PFA system in December 2023.
Focused ultrasound: HIFU and histotripsy.
+ The questionEvery energy so far reaches tissue through a needle, an instrument or an incision. How can sound destroy tissue deep inside a closed body without breaking the skin?
Bringing sound to a point
On 6 October 2023, FDA authorized HistoSonics' Edison system to destroy liver tumors by a non-thermal, mechanical process, in which pulses of focused sound sent in through the skin break the tumor's cells apart. The pulses can reach the tumor because sound travels through soft tissue with little reflection and moderate absorption, as Chapter 4 showed. A focused transducer uses both facts. Its surface is curved like the inside of a bowl, or made of many elements driven in a carefully timed sequence, so that sound leaving every part of it arrives at one point in phase. Spread over the large area of the transducer and the skin beneath it, the intensity is modest, but concentrated at the focus it is many times higher. That difference is what lets the sound destroy tissue at the focus while sparing the skin it crosses.
The focal zone is an elongated spot rather than a point, narrow across the beam and longer along it, and the wavelength and the geometry set its dimensions. Sonablate treats the prostate at 4 MHz, and FDA's decision summary for it describes a single lesion about 10 to 12 mm long and 1 to 2 mm wide, made at focal intensities of 1,200 to 2,000 W/cm². Larger volumes are treated by moving the focus and laying lesions side by side, under imaging.
Everything in the beam's path matters, and Chapter 4's reflection figures become rules for treatment. Gas in the bowel or lung reflects 99.9 % of the sound and blocks it, while bone reflects about two fifths, absorbs much of the rest and heats. Fat attenuates the sound. Perinephric fat surrounds the kidney, and in their 2024 review of HIFU, Turner and Cranston report a 2020 study in which that fat reduced the energy delivered from 58 % at a depth of 2 cm to 26 % at 5 cm. Treatment therefore needs the acoustic window of Chapter 4, a path of soft tissue free of gas and bone, and planning centers on finding one.
Thermal HIFU
In high-intensity focused ultrasound, or HIFU, the sound absorbed at the focus heats it during exposures that last seconds, and FDA's summary for Sonablate describes three seconds on and six off. Focal temperatures rise well above the 56 to 60 °C often quoted for tissue death within about a second, and Turner and Cranston report temperatures above 80 °C. Chapter 1 shows how much the time to tissue death at a given temperature varies from one tissue to another. Because the focus is small and heating is rapid, the lesion boundary is sharp, and tissue a few millimeters away is spared.
The advantages of a treatment without an incision are balanced by constraints. Each lesion is small, so large volumes take time. The heat is still heat, so perfusion and the heat sink of Chapter 2 apply, and nearby sensitive structures still need protection. The imaging that guides the treatment matters too, because systems guided by MRI can map temperature in the tissue during treatment, while ultrasound-guided systems see the anatomy in real time but have no direct thermometry.
In the US, FDA first approved a focused-ultrasound device on 30 June 1988, when it approved Sonocare's system for treating glaucoma by premarket approval, FDA's route for devices of the highest risk (Chapter 26). An MRI-guided system that treats uterine fibroids followed in 2004, also by premarket approval. For the prostate, FDA granted the Sonablate 450 a De Novo authorization on 9 October 2015. Chapter 14 introduced the De Novo route, which serves a new kind of device that has no predicate, no device already legally on the market to compare it with. With Sonablate's authorization, FDA created a new regulation for high-intensity ultrasound systems that ablate prostate tissue.
EDAP made a competing system called Ablatherm, and after Sonablate's authorization it withdrew its own De Novo request. EDAP filed a 510(k) instead, the route by which FDA clears a device by comparing it with a predicate, and named Sonablate as that predicate (Chapter 9). FDA cleared Ablatherm on 6 November 2015, and EDAP's later system Focal One received 510(k) clearance on 7 June 2018. The sequence shows how the US system works: the first device of a new kind creates a classification, and later devices of that kind follow it (Chapter 26). The international safety standard for such equipment is IEC 60601-2-62, which IEC first published in 2013. FDA's decision summary for Sonablate cites the diagnostic ultrasound standard, IEC 60601-2-37, instead.
Histotripsy: sound without heat
Histotripsy uses the same focused sound in a completely different way: instead of long exposures that let heat accumulate, it fires pulses lasting microseconds at duty cycles below 1 %, which means that the sound is on for less than a hundredth of the time. The peak negative pressure of each pulse is high enough to produce inertial cavitation at the focus. Chapter 4 described inertial cavitation as the violent growth and collapse of bubbles, and it put the threshold for single-cycle pulses at about 25 to 30 MPa of negative pressure, with no seed bubble already present. A cloud of bubbles forms, grows and collapses with each pulse, and the repeated collapses break the cells apart, after which the body clears the debris. With so little time on, the tissue barely warms.
FDA's decision summary for Sonablate describes HIFU exposures of three seconds on and six seconds off. Each cycle therefore lasts nine seconds, and the transducer is on for three of them, a third of the time, or about 33 %. A histotripsy system runs at a duty cycle below 1 %, so over the same nine seconds its pulses add up to less than a hundredth of nine seconds, under 0.09 seconds of sound. Three seconds divided by 0.09 seconds is about 33, so the HIFU transducer is on more than 33 times as long as the histotripsy transducer, which is silent for more than 99 % of the time. HIFU therefore keeps its sound on long enough for heat to build at the focus, while histotripsy leaves the focus almost no time in which to warm.
The mechanism has three useful consequences. There is no heat sink, because heat is not the agent. Collagen-rich structures such as vessels and bile ducts are more resistant to the mechanical stress than the cells around them, and in a 2026 review of preclinical studies, Yu and colleagues report that vessels larger than about 1 mm are preserved. And the bubble cloud is bright on ultrasound, so the treatment can be watched as it happens.
The same review describes Edison as operating at 700 kHz, with pulses shorter than 20 µs and duty cycles below 1 %. FDA's De Novo authorization of October 2023 covers the destruction of liver tumors by a non-thermal, mechanical process. The tumors it covers include unresectable ones, which surgery cannot remove. FDA's decision summary reports the results of HistoSonics' own trial, which supported the request: technical success was 95.5 %, in 42 of 44 tumors, and major complications occurred in 3 of 44 patients at 30 days. The trial treated tumors smaller than 3 cm, required an adequate acoustic window, and kept the main portal vein, main bile ducts, gallbladder, stomach and bowel out of the treatment volume.
HistoSonics announced on 11 May 2026 that it had submitted a De Novo request for kidney tumors, and as of October 2026 no FDA decision has been announced. On 13 July 2026, HistoSonics also announced CE marking for Edison in Europe, for the destruction of liver tumors. CE marking is the mark a manufacturer affixes before a device can be sold in the European Union, once its conformity assessment is complete (Chapter 27).
Press coverage of new energies often describes a new use as available before any regulator has authorized it. A De Novo request or a 510(k) submission is an application, and only FDA's decision makes a device legally marketable for that use. Histotripsy for liver is authorized in the US, while the kidney request had no announced decision as of October 2026.
HIFU and histotripsy can use similar transducers, focused through the same windows, and differ mainly in timing and peak pressure. Long exposures let heat build, and the focus coagulates. Microsecond pulses at a duty cycle under 1 % leave no time for heat to build but pull hard enough to cavitate, and the focus is fractionated mechanically. In principle, one transducer driven in two different ways becomes two different energies.
Treat the path to the focus as part of a focused-ultrasound device. Write the acoustic window into the requirements: the gas and bone the beam must avoid, the depth of fat it may cross, and how the user confirms a clear path before treatment. Specify each exposure by its timing and its peak pressure together, because, in principle, one transducer can coagulate or fractionate depending on both. State how the user watches the treatment, by MRI thermometry, by the bubble cloud on ultrasound or by the anatomy alone, and verify that view against lesions measured afterward. Before choosing a US route, check whether a device of the same kind has already created a classification, because a later device can follow it by 510(k), as Ablatherm followed Sonablate.
- A focused transducer concentrates sound at a small, elongated focus deep in the body while keeping the intensity low along the path, which must avoid gas and bone.
- Thermal HIFU heats the focus above about 60 to 80 °C in seconds, and because each lesion is small, volumes are treated lesion by lesion.
- Histotripsy fires microsecond pulses above the cavitation threshold at a duty cycle under 1 %, fractionating cells mechanically with little heat and no heat sink.
- Sonablate in 2015 and Edison in 2023 were authorized by De Novo, and later prostate systems followed by 510(k), while the kidney request that HistoSonics announced in May 2026 had no announced FDA decision as of October 2026.
+ Part VI · The case
One operation.
Partway through a laparoscopic right hemicolectomy, the surgeon puts down a monopolar hook that has just opened the thin membrane over the colon's blood supply and picks up a sealer for the vessels in the fat beneath it. That change of instrument is a change of energy, made because sealing vessels needs different physics from opening a thin membrane. This part follows one common operation step by step, as an illustration rather than a surgical technique, and puts the physics and the instruments of Parts I to V to work on one patient. It shows why one operation calls for several energies, and what each one does that the others could not.
One operation, several energies.
+ The questionIn a single laparoscopic colectomy, which energy does each step use, and why that one rather than another?
The operation
A right hemicolectomy removes the right side of the colon, usually for cancer, together with the fat, lymph nodes and blood vessels that supply it. Done laparoscopically, through the ports and cannulas of Chapter 7, it follows a sequence that varies with the surgeon and the anatomy but has a common shape. This chapter follows one version, a medial-to-lateral approach, as its illustration of where energy is used, and it does not describe surgical technique. The energy it names for each step is the chapter's example of a choice that suits the physics, and surgeons may choose otherwise.
The surgeon first places the ports and inspects the abdomen, and then opens the peritoneum, the thin membrane covering the mesentery, which is the sheet of fat that carries the colon's blood vessels and lymph nodes. The cut runs along the ileocolic vessels, which supply the end of the small bowel and the start of the colon, and those vessels are divided at their origin.
The mesentery is then lifted off the duodenum and the pancreas behind it, in a natural tissue plane. The duodenum is the first part of the small bowel. The rest of the mesentery, including the right branch of the middle colic vessels, is divided, and so is the omentum, the apron of fatty tissue that hangs from the stomach. The hepatic flexure, where the colon turns beneath the liver, and the attachments along the side of the abdomen are freed. Finally the bowel is divided on either side of the tumor, and the two ends are joined, usually with staplers.
The version followed here has seven steps and uses two or three energies and one or two kinds of mechanical closure. The choice at each step applies the physics of the earlier chapters to a particular tissue.
Opening planes: monopolar
Opening the peritoneum and developing the plane behind the mesentery means dissecting thin tissue that carries few blood vessels. The work needs precision, speed and a clear view, and it rarely needs to seal anything larger than a capillary. In the version followed here, the surgeon does it with a monopolar hook or scissors, an instrument that is thin, cheap and precise. At cutting voltages, the arc of Chapter 8 divides the membrane in a clean line, and when the surgeon switches to a coagulation waveform, the same instrument stops oozing from small vessels.
The costs of monopolar energy are the ones Part VII describes: the current crosses the patient to the return pad, and energy can leave the instrument anywhere along its shaft where the insulation is damaged. Where the shaft runs through a cannula, current couples across even intact insulation, and if the cannula is a hybrid, a metal tube held in a plastic anchor, it can discharge that current into the bowel (Chapter 18). In this step the shaft works close to the duodenum and the small bowel. The waveform matters too: the setting most likely to couple energy where it is not wanted is a coagulation waveform at several kilovolts peak, activated near the bowel or in open air.
Chapter 8 describes the modes of Medtronic's Valleylab FT10 generator. Medtronic's specification lists a rated output of 300 W and a peak voltage of 1,287 V for pure cut, and a rated output of 120 W and a peak voltage of 3,448 V for fulguration. Dividing 120 W by 300 W gives 0.4, so fulguration is rated at two fifths of the cut's power. Dividing 3,448 V by 1,287 V gives about 2.7, so fulguration puts about 2.7 times the cut's peak voltage across the instrument's insulation. The spray mode, at 3,932 V, puts a little over three times. A higher voltage breaks through damaged insulation more easily (Chapter 8), so a flaw that holds at the cut's peak may fail at fulguration's peak, even though fulguration's rated power is lower.
Dividing vascular tissue: sealing and ultrasonic energy
The mesentery and the omentum carry vessels a few millimeters across, too many to clip one by one and too large to trust to monopolar coagulation. In the version followed here, this is the work of advanced bipolar sealers and ultrasonic shears. Each grasps a bite of tissue, seals the vessels in it and divides it, so the surgeon can work along the mesentery bite by bite without changing instruments.
The two devices approach the same job through different physics. A bipolar sealer fuses collagen with current under high jaw pressure and ends each seal by itself, at an end point it reads from the tissue's impedance. Sealers are typically indicated for vessels up to 7 mm, and a manufacturer-authored study reports a thermal spread of about 1.3 to 2.1 mm for that maker's sealers (Chapter 9). An ultrasonic shear seals and cuts by friction, with no current through the patient, and dissects precisely, but its blade stays hot after each activation. Standard shears are indicated for vessels up to 5 mm, and newer designs for vessels up to 7 mm (Chapter 11).
In this chapter's account, the surgeon chooses between the two, or uses a hybrid, according to the tissue, the step and experience, and the choice turns less on which device is better than on which physics suits that moment. The robotic version of the operation makes the same choice with different instruments. On the sealing side, Intuitive states that its SynchroSeal seals and cuts at the same time, for vessels up to 5 mm in diameter. Ibanez Jimenez and colleagues tested it in a 2022 preclinical study that Intuitive funded, and two of the authors were employed by the company. SynchroSeal produced seals with burst pressures of about 1,160 mmHg, about three times the 360 mmHg acceptance level of Chapter 9, and a thermal spread of 1.2 to 1.5 mm. Its jaws reached a peak temperature of about 110 °C, against about 247 °C for the ultrasonic comparator in the same study, Ethicon's Harmonic ACE+7 (Chapter 11).
Randomized comparisons of energy devices in laparoscopic colorectal surgery exist, and this guide has not reviewed them. It does not rank devices or recommend a technique. The chapter explains why each energy suits the steps it is used for, and the choices it names illustrate that physics.
The largest vessels: a decision
The ileocolic artery and vein are usually the largest vessels divided in a standard right hemicolectomy. There the surgeon has a real choice between a sealer used within its labeled vessel size and a mechanical closure with clips or a vascular stapler. Chapter 9 frames the decision: beyond the indicated size a seal's strength varies from device to device and the maker's evidence no longer covers it, and a failed seal on a main vessel is a serious bleed.
A clip or staple does not depend on the tissue's impedance or collagen content, but it adds an instrument and leaves a device in the body. The surgeon in this version uses energy up to a size and a mechanical closure above it, and the size at which to switch is a clinical judgment that this guide does not make.
In the version followed here, the surgeon divides and joins the bowel itself with staplers rather than energy, because the bowel is thick, it contains bacteria and gas, and its ends must be joined reliably. Energy may open the small holes through which the stapler enters. Otherwise, near the bowel, energy is a hazard to manage rather than a tool.
What the generator saw
When an advanced bipolar device seals one bite of mesentery, the generator's whole record of the seal is a few seconds of electrical measurements. The generator drives power at its setting while the impedance dips as the tissue warms. As water boils off, the impedance climbs, the output reaches its voltage limit and the delivered power falls. When the impedance curve shows that the water has gone, the generator stops, waits while the seal cools in the closed jaws, and sounds a tone.
Nothing in that record measures the seal. The generator knows voltage, current and time, and it infers the rest through an algorithm tuned on bench and animal tests. The surgeon hears a tone and opens the jaws. Whether seals of this kind hold against arterial pressure while the vessel heals was tested before the device was sold, with the burst tests, histology and long-term animal studies of Chapters 9 and 25.
Why surgery needs eleven energies
With Parts III to V behind it, this operation shows why surgery needs several kinds of energy, and the table below extends the reasoning to all eleven. Surgery's tasks need heat in different geometries and at different speeds, and some tasks need the tissue's scaffold to survive and so need no heat at all. Where two energies meet the same need, each trades a different weakness.
| Task | What the physics needs | Energy that fits |
|---|---|---|
| Open a thin plane | a moving line of vaporization, fast and narrow | monopolar cut |
| Coagulate a bleeding surface | shallow coagulation over an area | monopolar coag or spray |
| Coagulate a small vessel precisely | current confined between two tips | conventional bipolar |
| Seal and divide vascular tissue | collagen fusion under pressure, with an end point | advanced bipolar, ultrasonic or hybrid |
| Close the largest vessels | a mechanical closure that does not rely on heating the tissue | clips or staplers, or a sealer within its size |
| Vaporize a thin layer precisely | shallow absorption, short pulses | CO₂ laser |
| Break a stone under water | pulsed energy through a vapor channel | holmium or thulium laser |
| Kill a tumor in place | a lethal volume several centimeters across | RF, microwave or cryoablation |
| Kill cells but spare vessels and ducts | a non-thermal mechanism | IRE, PFA or histotripsy |
| Treat without an incision | a focus deep inside intact tissue | HIFU or histotripsy |
Each row asks for a geometry, a speed or a mechanism that the others cannot provide. Where a row names more than one energy, the energies differ in a second property. Microwave keeps heating past the impedance roll-off that stops RF, and the heat sink affects it less (Chapter 13). Cryoablation shows its zone as it forms but does nothing to stop bleeding (Chapter 14). Ultrasonic energy keeps current out of the patient.
The last part of the question asks what it takes to keep each energy where the surgeon intends, and nowhere else, and the rest of the guide answers it. Parts VII and IX cover the energy that leaves the instrument before it reaches the tip, the smoke and light that leave the patient, and the evidence a device must produce before it is sold.
Map a device's intended use onto the steps of a real operation before writing its requirements. For each step, record the tissue, the largest vessel the device will meet, the structures that lie close to its jaws or shaft, and the energy or mechanical closure a surgeon could use instead. Test at each step's worst case: the largest vessel within the indication, the mode with the highest peak voltage working near the bowel, and a jaw or blade still hot from its last activation. Because the generator infers each seal and never measures it, prove seal strength at those worst cases with burst tests (Chapter 9).
- In the version this chapter follows, a laparoscopic right hemicolectomy uses monopolar energy for tissue planes, advanced bipolar or ultrasonic devices for vascular tissue, either a sealer within its size or clips or a stapler for the largest vessels, and staplers for the bowel.
- Each choice follows the physics: fast, narrow vaporization for dissection, collagen fusion for sealing, and mechanical closure where sealing that depends on the tissue is least reliable.
- The generator infers each seal from voltage, current and impedance, and the strength of such seals is established by testing before the device is sold.
- One operation shows why surgery needs several energies: its tasks need heat in different geometries and at different speeds, some need no heat at all, and where two energies meet one need, each trades a different weakness.
+ Part VII · The patient and the room
The system around the energy.
A metal cannula held in the abdominal wall by a plastic anchor can collect current from an instrument whose insulation is intact and discharge it into bowel that no instrument touched. An energy device is designed to act at its tip, and this part first follows the energy that goes elsewhere: current that leaves an instrument before it reaches the tissue, and the smoke, sparks, laser light and electrical noise that leave the patient altogether. It then turns to the system that keeps the energy in its place: a generator built to fail safe, and the interfaces where parts from different makers meet, each with an owner. The part ends with the process by which a team traces each hazard to its control and to the test that shows the control works.
Burns away from the tip.
+ The questionHow can electrosurgical energy burn tissue that no instrument ever touched?
Three stray paths
A bowel wall coagulated by electrosurgical current can look almost normal when the instrument leaves it, then die and perforate days later (Chapter 1). Electrosurgical injuries in laparoscopy are uncommon but serious. A review published in 2000 put them at two to five per thousand laparoscopic procedures in the 1990s, and a 2021 review gives 3.6 per thousand. Many of the injuries are not recognized during the operation, and the current that causes them need not come from the tip. Chapter 3 established that heat appears wherever current is concentrated, and Chapter 7 that most of a laparoscopic instrument lies out of the camera's view, which means that current leaving a monopolar instrument before it reaches the tip can burn bowel where nobody is looking. It can leave in three ways.
Direct coupling. The active electrode touches another conductor, such as a metal grasper or the laparoscope, and that conductor touches tissue. Current flows through the second instrument and heats the tissue at its point of contact, wherever that is.
Insulation failure. Wear, reprocessing or rough handling can break the insulation of the shaft and expose the conductor. If tissue lies against the defect, current flows out through a tiny area, and the current density there can be as high as at the tip.
Capacitive coupling. Even with perfect insulation, a conductor inside a metal cannula forms a capacitor, and alternating current flows across it without any contact at all.
Every burn in this chapter happens away from the tip, and most of them out of view, so a device that looks perfectly controlled on the monitor can still be delivering energy somewhere else. Testing, monitoring and design must therefore cover the whole path from the generator to the return, not only the working end.
Insulation that fails out of view
Insulation defects are not rare. In a 2023 study at one center, Homma and colleagues tested 69 reusable laparoscopic instruments and found defects in 8 of them, 11.6 %. In 2010, Montero and colleagues found insulation failure in 31 of 165 reusable laparoscopic instruments, or 19 %. Robotic instruments fare worse, as Chapter 7 showed. Espada and colleagues found an insulation failure in 32 % of the robotic instruments they examined, against 13 % of the laparoscopic ones. Among the instruments that the same study followed over time, 80 % of the robotic ones failed within ten uses, against 36 % of the laparoscopic ones. A later phase of the study found failures in 81.7 % of the robotic and 19.5 % of the laparoscopic instruments.
A defect needs voltage to become a burn. High-voltage coagulation waveforms reach thousands of volts peak, several times the peak voltage of a cutting waveform (Chapter 8). They drive current through small defects far more readily than a cutting waveform, and even through thin insulation that is intact but stressed. That is why IEC 60601-2-2 requires the insulation of active accessories to pass a high-frequency test at 120 % of the accessory's rated voltage for 30 seconds. An accessory's rated voltage is the highest peak voltage it is designed to withstand. The standard ties the crest factor of the test waveform to that rating, so an accessory rated for a high voltage must survive a waveform like the one it will carry (Chapter 25).
The cannula as a capacitor
Capacitive coupling can be estimated from school physics, because an insulated conductor inside a metal tube is a cylindrical capacitor. Its capacitance depends on the length of shaft inside the tube and on the ratio of the tube's radius to the conductor's. When the insulation is thin compared with the shaft, the cylinder behaves almost like a flat capacitor: the insulation is the gap, and the surface of the shaft inside the tube is the area of the plates. The capacitance then grows with the length of shaft in the tube and falls as the insulation gets thicker.
The example assumes a 5 mm instrument with a conductor of radius 2.25 mm in insulation 0.25 mm thick, of relative permittivity about 2.1, pressed against a metal cannula. Over 10 cm of shaft the insulation covers about 15 square centimeters. The permittivity of free space, 8.85 pF per meter, times 2.1 and that area, divided by the thickness, gives about 110 pF. An air gap of 0.25 mm lowers this to about 40 pF, 30 cm of shaft near metal raises it to about 330 pF, and the example takes 50 to 200 pF as typical. At 500 kHz, 50 and 200 pF present reactances of about 6.4 and 1.6 kΩ, and the coupled current is the voltage divided by the reactance. A cut at 300 W into 300 Ω runs at 300 V RMS and drives about 47 to 188 mA through these capacitances. The coagulation modes run at a higher voltage. Chapter 8 gave the FT10's fulguration and spray as 3,448 and 3,932 V peak, with crest factors of 5.4 and 6.2, from Medtronic's specification. Each peak divided by its crest factor is about 635 V RMS, twice the cut's voltage, so the current roughly doubles.
The estimate is an upper bound, because a generator limits its own output, and it agrees with a company's own measurement. Electroscope, now Encision, measured normal capacitive currents of 50 to 250 mA RMS in the shield of an instrument of its own design, and reported them in a patent filed in 1991 and granted in 1994. Coupled current rises with voltage, with frequency and with capacitance, so coagulation waveforms drive the most current through the cannula, and so does activation in open air, because with no tissue to load the generator, its output voltage rises to its maximum.
Where that current goes depends on the cannula. An all-metal cannula held in the abdominal wall passes the coupled current into the wall over a broad contact, where its density is low and it returns harmlessly. The dangerous combination is a hybrid cannula, a metal tube held in the wall by a plastic anchor. The metal collects the charge and the plastic insulates the tube from the wall, so the charge discharges into whatever touches the metal, which may be bowel. FDA's electrosurgical guidance recommends that manufacturers test capacitive coupling to a conductive cannula and warn in their labeling against hybrid trocar systems, which combine metal and plastic.
Engineering answers
Active electrode monitoring addresses both insulation failure and capacitive coupling. A conductive shield surrounds the insulated conductor of the instrument's shaft and connects back to the generator's return, and a second layer of insulation covers the shield. The shield collects any current that escapes the conductor, through a defect or by capacitance, and returns it safely, while a monitor measures that current. Electroscope's patent describes a monitor that tells the two kinds of current apart by their timing, because normal capacitive current is 90° out of phase with the voltage, a quarter of a cycle apart, while the current of a fault is in phase with it. When a fault appears, the monitor stops the generator.
The other answers are procedural, and they belong in the instructions for use and in training. They include inspecting the insulation, using the lowest effective voltage, preferring a cutting waveform to a coagulation waveform wherever the cutting waveform works, never activating in open air, keeping the whole active electrode in view where possible, and avoiding hybrid cannulas. Chapter 19 lists some that have become design features, such as low-voltage modes and activation that cannot occur accidentally.
Pad burns, and what surgeons know
The return pad is the other place where monopolar current can burn away from the tip, and Chapter 8 described the physics: a pad that lifts concentrates current at its remaining edge. The heating under a pad grows with the square of the current and with the time it flows, which is the heating factor of Chapter 10. IEC 60601-2-2 therefore adds a safeguard for the high current modes defined there, monopolar modes whose intended use can exceed 30 A²s in any 60 seconds, with stricter requirements for the return electrode. Its 2023 amendment asks for heating under the return electrode to be considered for any mode whose intended use has an activation duty cycle above 45 %, even below that level (Chapters 10 and 25).
Knowledge is part of the system. In 2011, 48 leaders of the Society of American Gastrointestinal and Endoscopic Surgeons took the knowledge test that Chapter 8 mentioned. Feldman and colleagues reported the results in 2012: a median score of 59 %, and 31 % of the leaders did not know how to handle a fire on the patient. The authors wrote for the society's task force on the Fundamental Use of Surgical Energy, FUSE, and the society's answer was training. In 2012 the society published the manual and online curriculum of its FUSE program, and in 2014 it certified its first cohort of FUSE surgeons.
Treat the whole path from the generator to the return as the device under test. For a monopolar laparoscopic instrument, estimate the capacitance of its shaft to a metal cannula from the shaft's diameter, the length inside the cannula and the thickness and permittivity of its insulation. Then divide the generator's highest RMS voltage by the reactance of that capacitance at the working frequency, to bound the coupled current. Test the insulation at high frequency against the accessory's rated voltage, and write insulation inspection into the instructions for use. Warn in the labeling against hybrid cannulas and activation in open air, and where the instrument has active electrode monitoring, verify that the monitor stops the generator when an in-phase fault current appears.
- Current can leave a monopolar instrument by direct coupling, insulation failure or capacitive coupling and burn tissue out of view, and a coagulated bowel wall can look almost normal when the instrument leaves it and perforate days later.
- Insulation defects are common in reusable instruments and more common in robotic ones, and high-voltage coagulation waveforms, at several times the peak voltage of a cut, drive current through them most readily.
- On this chapter's assumptions, a shaft inside a metal cannula forms a capacitor of roughly 50 to 200 pF that passes tens to hundreds of milliamperes, and a hybrid cannula can discharge that current into bowel.
- Active electrode monitoring, insulation testing, lower voltages and the correct choice of cannula are the engineering and procedural answers.
Smoke, fire, light and implants.
+ The questionBeyond the patient's tissue, what does surgical energy do to the air, the eyes, the monitors and the implants around it?
Smoke
OSHA estimates that about 500,000 US workers are exposed to laser or electrosurgical smoke each year. Chapter 1 showed that cutting boils the water in cells until they burst and that dried tissue chars. Both release the contents of the cells into the air as a plume, and the plume is not harmless steam. Reviews of surgical smoke report more than 80 chemical compounds in it, including benzene, formaldehyde, acrylonitrile and polycyclic aromatic hydrocarbons, and viral DNA has been detected in plume from the treatment of papillomavirus lesions. A 1981 study compared the smoke with cigarettes for mutagenicity, the power to cause mutations, and found that the smoke from one gram of tissue matched about six cigarettes for electrocautery and three for a CO₂ laser.
The particles are small. A 2019 review by Liu and colleagues reports a mean aerodynamic size of about 0.07 µm for electrosurgical smoke particles and about 0.31 µm for laser plume particles, well under the size of a bacterium. Ultrasonic devices produce an aerosol mechanically rather than by combustion, and Casey and colleagues compared the two kinds of particle in a 2021 laboratory study, sizing the particles by laser diffraction. They cut liver with electrosurgery and with an ultrasonic device, and bone with an ultrasonic device and a high-speed saw. In all four, 90 % of the particles were 6.27 µm or smaller. Half were smaller than 2.5 µm in all but ultrasonic cutting of bone, where half were smaller than 2.72 µm. Particles this small follow the air and reach deep into the lungs.
Because the particles follow the air, the effective control is to capture the plume at its source. NIOSH's guidance of 1996 recommends local exhaust ventilation with high-efficiency filtration, drawing air at a capture velocity of 100 to 150 feet per minute into a nozzle held within 2 inches of the site. In laparoscopy the same problem appears inside the abdomen, where smoke obscures the camera and must be removed without losing the gas pressure that holds the abdominal wall up. FDA's electrosurgical guidance recommends that the labeling warn of the hazards of smoke and call for effective evacuation. As of October 2026, the Association of periOperative Registered Nurses lists 22 US states that have enacted laws requiring surgical smoke evacuation.
Fire
A fire needs three things at once, often drawn as the three sides of a triangle: an oxidizer, a fuel and a source of ignition. An operating room supplies all three in abundance. The oxidizers are oxygen and nitrous oxide, and oxygen is often concentrated under surgical drapes or around the airway. The fuels include alcohol-based skin preparations, drapes, gauze, hair, the airway tube itself and the gas in bowel. The ignition source is usually an energy device.
In a 2018 analysis of Pennsylvania's patient-safety reports, Bruley and colleagues counted 28 surgical fires reported between July 2011 and June 2016. Electrosurgery was the ignition source in 79 % of them, 22 of the 28, and about two thirds occurred in the head or neck, where oxygen is delivered close to the surgical site. The 28 fires amount to 5.6 a year, 44 % below the 10 a year found in an earlier analysis published in 2012. In that analysis, Clarke and Bruley had counted 70 fires from July 2004 to June 2011. Bruley and colleagues also cite a 2009 guide by ECRI, the Emergency Care Research Institute, for the finding that oxygen-enriched atmospheres contribute to about 70 % of surgical fires.
Each side of the triangle has a control. On the oxidizer side, the anesthesia team can keep open oxygen delivery near the surgical site at 30 % or less and avoid nitrous oxide, and Clarke and Bruley call keeping oxygen below 30 % desirable to prevent rapidly spreading fires. On the fuel side, alcohol preparations must dry completely before the drapes go on, and laser surgery in the airway uses laser-resistant tubes whose cuffs are filled with saline. On the ignition side, energy devices can be used at the lowest effective setting and kept in holsters when not in use. In endoscopy of the colon the gas in the bowel is itself a fuel, and the 2010 guideline of the European Society of Gastrointestinal Endoscopy notes the risk of a gas explosion when methane exceeds 5 %, hydrogen 4 % and oxygen 5 %.
Light
Surgical lasers are Class 4 lasers, the highest hazard class: hazardous to view directly or by reflection, and able to ignite materials. A beam weakens as it spreads, so its hazard to the eye ends at some distance from the laser. The eye's maximum permissible exposure is the highest exposure it can receive without injury, and the nominal ocular hazard distance is the distance along the direct beam over which the exposure stays above that maximum. Reflections and focusing optics define a wider nominal hazard zone around the laser.
That distance can be large. The manufacturer of one 180 W, 532 nm system states a nominal ocular hazard distance of 33.9 m and requires eyewear with an optical density of at least 6 at 532 nm, that is, eyewear that passes no more than one millionth of the light at that wavelength. For one holmium system and its compatible fibers, the manufacturer states a much shorter worst-case distance, 1.6 m, and eyewear with an optical density of 3, which passes one thousandth. Where the distance exceeds the room, as it often does, the whole room is the hazard zone and everyone in it wears eyewear rated for the wavelength.
The distance also depends on how long an exposure the calculation assumes. In its technical manual, OSHA works through an example of a 100 W Nd:YAG laser. For an exposure criterion of 8 hours, the manual gives hazard distances of 1.4 m for a diffuse reflection, 11.3 m for the beam through a lens and 1,410 m for the direct beam. For a criterion of 10 seconds, the same three distances are 0.8 m, 6.3 m and 792 m.
In OSHA's example, the direct beam of a 100 W Nd:YAG laser has a hazard distance of 792 m for a 10-second exposure and 1,410 m for an 8-hour one. Far from the laser the beam spreads at a fixed angle, so its diameter grows in proportion to the distance and its irradiance, the power on each square centimeter, falls with the square of the distance. At each hazard distance the irradiance has fallen to the maximum permissible exposure for that exposure time. The ratio of the two distances is 1,410 divided by 792, about 1.78, and the square of 1.78 is about 3.2. The two distances therefore imply an 8-hour maximum permissible exposure about 3.2 times lower than the 10-second one. The diffuse-reflection distances, 1.4 m and 0.8 m, stand in nearly the same ratio, 1.75. An exposure limit lowered about threefold lengthens the hazard distance by less than twofold, because the distance grows only with the square root of the ratio of the limits.
Two standards govern laser safety. IEC 60825-1 classifies laser products and sets requirements for them, and its current edition is from 2014. IEC 60601-2-22 sets the particular requirements for medical lasers. Its fourth edition dates from 2019, and IEC published an amendment to it in April 2026. In the US, laser products must also meet the federal laser performance standard, 21 CFR 1040.10 and 1040.11. In Laser Notice 56, FDA states that it does not intend to enforce most of those requirements for products that conform to the comparable clauses of the 2014 edition of IEC 60825-1 and, for medical lasers, of edition 3.1 of IEC 60601-2-22, the edition before the fourth. FDA first issued the notice in May 2019 and reissued it in February 2023, and as of October 2026 the 2023 version is the current one.
Implants and monitors
Current flowing through the patient is also electrical noise, and it can disturb devices that sense the body's own electrical signals. Pacemakers and implantable defibrillators are the most important of these devices. Their leads pick up voltages, and electrosurgical current passing near the heart can be misread as cardiac activity, which can inhibit pacing or trigger a defibrillator.
The 2010 guideline of the European Society of Gastrointestinal Endoscopy also sets out the usual precautions for patients with these implants. Bipolar energy is preferred where it will do. If monopolar energy is needed, the return pad is placed so that the current path does not cross the pacemaker and heart, and a pure cutting waveform may be preferable. Qualified staff may reprogram the device, or use a magnet where the device supports it, and continuous ECG monitoring and a defibrillator are kept ready. The Heart Rhythm Society and the American Society of Anesthesiologists give more detailed recommendations in a joint consensus statement. Ultrasonic energy, which passes no current through the patient, avoids this interference altogether.
The operating room's own monitors are affected too, most visibly the ECG, which shows interference whenever electrosurgery is active. IEC 60601-1-2, the standard for electromagnetic compatibility, limits a device's unintended emissions and sets how much interference the device must tolerate, and IEC 60601-2-2 modifies those requirements for electrosurgical equipment. The generator's intended output is itself a strong high-frequency current in the patient, so it cannot be designed away, only kept where it belongs.
Treat smoke, fire, laser light and electromagnetic interference as hazards with a device side, although they are usually handled as the operating room's problems. On the device side, decide on evacuation ports and filters, activation that cannot occur accidentally, holsters, low-voltage modes, and fiber and beam interlocks. Put the specification of the eyewear the laser requires in the labeling, keep emissions within the limits of IEC 60601-1-2, and record each of these design and labeling decisions in the risk file.
- Surgical smoke contains more than 80 chemical compounds and particles with a mean size of about 0.07 µm for electrosurgery, and capture at the source is the effective control, which 22 US states had enacted laws to require as of October 2026.
- A surgical fire needs an oxidizer, a fuel and an ignition source at once: electrosurgery was the ignition source in 79 % of the fires reported in Pennsylvania from July 2011 to June 2016, and limiting open oxygen to 30 % or less is the main control on the oxidizer side.
- Surgical lasers are Class 4, and their nominal ocular hazard distance depends on the exposure criterion and can exceed the room, so everyone present wears eyewear rated for the wavelength.
- Electrosurgical current can interfere with pacemakers, implantable defibrillators and monitors, and bipolar or ultrasonic energy, pad placement and management of the implanted device reduce the risk.
Built to fail safe.
+ The questionWhat has to be true of a surgical generator, beyond delivering the right energy, before it can safely touch a patient?
Basic safety and essential performance
A surgical generator that never gives anyone an electric shock can still injure a patient, by delivering too much energy or by delivering any energy at the wrong moment. The general standard for medical electrical equipment, IEC 60601-1, organizes safety around two ideas. Basic safety is freedom from unacceptable risk caused directly by physical hazards when the equipment is used in normal condition or with a single fault: electric shock, fire, mechanical injury, excessive temperature. A generator could meet that and still harm a patient through its output, which is where the second idea applies. Essential performance is the performance the device must maintain for the patient to be safe, as identified by the manufacturer's risk analysis.
For a surgical generator, essential performance is largely a statement about energy. The output must match the setting within its declared limits, it must not appear when it has not been commanded, and it must stop when commanded. The rest of this chapter follows from making those three statements true under fault conditions.
A second channel that can stop everything
IEC 60601-1 requires equipment to remain safe after any single fault, such as a failed component, a stuck switch or a software error. A generator's controller cannot be the only thing standing between a fault and an unintended output, because the controller may be what has failed. Generators are therefore built with two channels.
The control channel is the one described in Chapter 10: the user interface, the controller, the power stage and the sensors that close the loop on the output. The safety channel runs alongside it, with its own sensors on the output and its own logic, independent of the controller. It checks three things: that the output matches what was commanded, that nothing is delivered without activation, and that the return pad monitor and the other interlocks are satisfied. If any check fails, the safety channel disables the power stage directly, without going through the controller. Valleylab filed a patent on its split-pad monitor in 1981, and the patent already describes a monitor that disables the generator (Chapter 8). Modern designs extend the same principle to the output itself.
Around the channels sit the indicators and alarms the operator relies on. Activation produces a tone and a light. IEC 60601-2-2 specifies the colors of the lights for warnings and for the cut, coagulation and ready states in a table, and its 2023 amendment replaced that table with a new one. Alarms follow IEC 60601-1-8, the collateral standard for alarm systems.
Insulation barriers and leakage
A generator connects to the mains on one side and to the patient on the other, and it must keep the two apart even after a fault. IEC 60601-1 calls the barriers that do this means of protection: insulation, clearances through air and creepage distances along surfaces, each rated for a test voltage. The protection of the operator and the protection of the patient are rated separately and differently. For mains up to about 250 V, one means of patient protection is commonly quoted as a test at 1,500 V AC with 4 mm of creepage and 2.5 mm of clearance, and two means of patient protection as a test at 4,000 V AC with 8 mm and 5 mm.
The standard then limits the current that may leak through those barriers. IEC 60601-1 calls the parts that contact the patient applied parts and classifies them as type B, BF or CF, from the least to the most isolated, and CF is required for anything applied directly to the heart. Under normal conditions, patient leakage current is limited to 100 µA for types B and BF and 10 µA for type CF, rising to 500 µA and 50 µA with a single fault. Earth leakage from the equipment is limited to 5 mA normally and 10 mA with a fault.
Those limits concern low-frequency current, which is dangerous because it can disturb the heart. A surgical generator also has a high-frequency limit of its own, for a different harm. Its output is meant to deliver current to the patient, so the relevant question is how much high-frequency current escapes by paths other than the intended return, through stray capacitance to earth. For monopolar outputs, IEC 60601-2-2 limits this high-frequency leakage current to 150 mA, measured through a 200 Ω resistor that stands in for a person. That limit is three to four orders of magnitude above the patient-leakage limits, because at several hundred kilohertz the risk is a burn, not fibrillation. Medtronic declares under 100 mA for the monopolar outputs of its FT10 generator, measured directly at the generator's terminals. The standard's limit of 150 mA applies with the leads laid out as the standard specifies, so the declared figure comes from a different set-up and cannot be compared directly with it (Chapter 25).
IEC 60601-2-2 allows 150 mA of high-frequency leakage from a monopolar output, while IEC 60601-1 allows a patient leakage current of 100 µA from a type B or BF applied part under normal conditions and 10 µA from a type CF part. A milliampere is 1,000 microamperes, so 150 mA is 150,000 µA. Divided by 100 µA, that gives 1,500, and divided by 10 µA it gives 15,000. A factor of 1,500 is three tenfold steps and a further factor of one and a half, and 15,000 is four tenfold steps and the same one and a half. Even the single-fault limits of 500 µA and 50 µA sit 300 and 3,000 times below 150 mA. The high-frequency limit therefore sits a little more than three tenfold steps above the type B and BF limit, and a little more than four above the type CF limit.
Software and cybersecurity
Modern generators are defined as much by software as by hardware: the closed-loop algorithms of Chapters 9 and 10, the end-point detection that decides when a seal is finished, the safety checks and the user interface. Software development follows IEC 62304, the standard for medical device software life-cycle processes. In its 2023 guidance on software in premarket submissions, FDA sorts devices into two documentation levels, Basic and Enhanced, which replaced the earlier Levels of Concern. FDA's electrosurgical guidance states that electrosurgical devices for general surgery should generally address the Enhanced level.
Connectivity adds a newer obligation. Section 524B of the US Federal Food, Drug, and Cosmetic Act covers "cyber devices", devices with software that can connect to the internet and could be vulnerable to cybersecurity threats. For premarket submissions made on or after 29 March 2023, it has required the sponsors of such devices to submit a plan for monitoring and addressing vulnerabilities, to maintain processes for updates and patches, and to provide a software bill of materials, a list of the software components in the device. A generator with a network or service interface is likely to qualify. The public summary of one generator's clearance in 2025 notes an added service interface and a cybersecurity assessment performed according to FDA guidance (Chapter 25).
Risk and use
Risk management under ISO 14971 organizes all of this. The standard requires the manufacturer to identify hazards, estimate and control their risks, and show that what remains is acceptable. Many of the hazards in Chapters 8, 18 and 19 appear in a generator's risk file: burns at the pad and away from the tip, fire, interference with implants and unintended activation. Some are controlled in the hardware and software, such as by the pad monitor and the safety channel, and others by labeling and training. Chapter 22 follows one hazard from its analysis to the test that shows its control works.
Usability engineering under IEC 62366-1 connects the controls built into the device with those left to labeling and training. The knowledge-test results of Chapter 18 show that trained surgeons do not always know how their energy devices fail, so a design that tolerates predictable errors controls risk more reliably than a warning in the instructions. A generator can, for instance, refuse to operate with a single-plate pad in a mode that needs a split one.
The safety architecture of a generator follows one rule: no single failure may deliver energy that was not intended, whether a component, the software or the user fails. The second channel, the insulation barriers, the pad monitor, the software process and the usability work are that rule applied to different kinds of failure.
Write the essential performance of a generator as statements about energy that a test can check: the output matches the setting within its declared limits, no output appears without a command, and the output stops when commanded. For each statement, list the single faults that could break it, such as a failed component, a stuck switch or a software error, and show which channel catches each one without relying on the controller that may have failed. Measure patient leakage against the limit for each applied part's type, and the high-frequency leakage of each monopolar output against 150 mA through 200 Ω. Record each control and each result in the risk file under ISO 14971.
- IEC 60601-1 organizes safety around basic safety and essential performance, and for a generator, essential performance means an output that matches its setting, appears only when commanded and stops when commanded.
- Generators use an independent safety channel that can disable the power stage even if the controller fails.
- Means of protection and patient-leakage limits in microamperes guard against low-frequency current, which can disturb the heart, and a separate limit of 150 mA through 200 Ω governs high-frequency leakage from monopolar outputs, a current whose risk is a burn.
- Software follows IEC 62304, the sponsors of cyber devices carry the duties of section 524B in premarket submissions made on or after 29 March 2023, and risk management under ISO 14971 and usability engineering under IEC 62366-1 organize the whole design.
Where one maker's part meets another's.
+ The questionIf the parts of an energy system are designed, bought, connected and cleaned by different people, which interfaces decide whether the energy goes where it should, and who owns each one?
Inside and outside the energy system
A reusable Valleylab pencil, Covidien's E2100, carries three statements in its instructions for use that its maker cannot keep true alone. Covidien rates the pencil for a maximum peak voltage of 5,600 V, but the generator it plugs into sets the voltage it meets. The instructions state a validated life of 50 uses with steam sterilization, but the hospital cleans the pencil and counts the uses. And they limit it to the generators listed on their cover, the FT10 among them. Each statement marks where the pencil meets a part or person that another party controls.
The pencil is one part of a larger system. Beyond the generator of Chapter 20, the energy reaches the tissue through the instrument and its cable, the return pad, the switches, often a smoke evacuator, and in robotic surgery a robot (Chapter 7). The system must tolerate three things it does not control: the tissue, a load whose impedance ranges from about 20 Ω to 2 kΩ or more and changes as the energy acts, as Chapter 10 showed; the people who connect, use, clean and service its parts; and the other equipment in the room. The tissue differs from the other two, because it also lies inside the control loop: the generator reads it through its own output (Chapters 9 and 10).
Each place where two parts meet is an interface, and an energy system has four kinds. High-frequency power crosses from the generator to the instrument and returns through the pad. Signals cross from the switches and the pad monitor. Data cross between a robot and its generator, and through service ports. Procedures cross from the maker to the hospital, as instructions for reprocessing and service. Each interface needs a specification, and the party that writes it owns the interface, even where another party must meet it and a third verifies it.
The generator and its instruments
A generator first has to know what is plugged in. Medtronic describes the FT10 as "a plug-and-play system with automatic device detection and power setting", and its specification lists an RFID function at 13.56 MHz without stating what the function is used for. Erbe states that the VIO 3 recognizes an instrument and loads preprogrammed starting settings and instrument-specific parameters. Neither maker's document says how recognition works or what it restricts, and Medtronic's does not link the RFID function to recognition. Olympus does state what its ESG-410 generator restricts: the generator recognizes only Olympus instruments on its universal socket, and no monopolar or bipolar instrument automatically. Recognition therefore belongs to particular pairings, and the generator's maker owns it.
Voltage, by contrast, is the interface every pairing has. IEC 60601-2-2 defines a generator's maximum output voltage, for each mode, as the highest peak voltage that can appear between its patient connections. An active instrument carries a rated accessory voltage, which MEDTEQ, a maker of test equipment, describes as the rated peak voltage stated in the instructions for use. The instrument's insulation must pass a high-frequency test at 120 % of that rating for 30 seconds (Chapters 18 and 25). The generator's maker controls one number and the instrument's maker the other.
FDA links the two numbers through the labeling, in an electrosurgical guidance whose version of 9 March 2020 is current as of October 2026. The guidance asks for the rated voltage on the accessory's labeling, the maximum output voltage on the generator's, and instructions that "direct the user to verify that the generator's output voltage does not exceed the rated accessory voltage". Makers write the same rule into their own instructions. Kirwan, for one, tells users of its AURA 70 bipolar generator, in operating instructions revised in December 2019, to use only accessories whose voltage rating is greater than the generator's maximum output voltage of 210 V peak. On the generator's side of that comparison, Medtronic lists peak voltages for the five FT10 monopolar modes of Chapter 8 from 1,287 V for pure cut to 3,932 V for spray, and the FT10's bipolar and sealing outputs stay at 530 V or below (Chapter 9).
Covidien rates its E2100 pencil for 5,600 V peak and lists the FT10 as compatible, and Medtronic lists the FT10's highest monopolar peak, in spray, as 3,932 V. Taking each listed peak as the highest its mode reaches, the rating exceeds the spray peak by 5,600 less 3,932, or 1,668 V, and 5,600 divided by 3,932 is about 1.42, a margin of about 42 %. The insulation is tested at 1.2 times 5,600, or 6,720 V peak, about 1.7 times the spray peak, and above a rating of 4,000 V the test waveform's crest factor is 6, so the test runs at 6,720 divided by 6, or 1,120 V RMS, for 30 seconds. A hypothetical pairing that no maker lists shows the trap: the FT10 with Cook Medical's ACU-1 active cord, which Cook rates at "5 kVp-p (2.5 kVp)". Against 2,500 V peak, the Valleylab, fulguration and spray modes exceed the rating by 283, 948 and 1,432 V, while pure cut and blend stay within it. Read as 5,000 V peak, the same rating would seem to clear every mode by more than 1,000 V. A peak-to-peak rating must therefore be halved, for a symmetrical wave, before it meets peak voltages, and here the halving puts three of the five modes above the cord's rating.
Cook names no generator and warns against any setting that may exceed the cord's rating, which leaves the comparison to the user. A pairing that the generator's maker never tested still needs evidence, and FDA and IEC place the duty for it with different parties. FDA places the duty on the accessory's maker, asking the maker of an active electrode to address its use with other makers' generators, and to name a compatible generator or give criteria for identifying one. IEC divides the duty instead. In its clause on electromagnetic disturbances, the 2023 amendment to IEC 60601-2-2 states that ensuring compatibility between generator and accessories is the responsibility of the operator, the person who uses the equipment. IEC clarified the point in an interpretation sheet of November 2025: providing the information for it, along with each product's compliance on electromagnetic compatibility, is each manufacturer's responsibility. Under the two IEC texts, the operator is responsible for the combination and each maker for the information it rests on. AORN, the association of perioperative nurses, turns the operator's duty into practice: its summary of its 2026 guideline, published in December 2025, asks staff who use several generators to verify compatibility and to "label accessories to correspond with each generator".
Evidence for a pairing exists only where someone tested that pair, and a plug that fits the socket is no evidence. Covidien offers a variant of its pencil, the E2100E, for Erbe's generators. Intuitive states in its 2024 summary that the E-200 generator is compatible with "open and laparoscopic third-party handheld monopolar and bipolar instruments". Its 2023 summary reports tests with other makers' products: compatibility with an Erbe split pad, a capacitive-coupling evaluation with third-party laparoscopic monopolar accessories, and a validation in pigs that included third-party instruments, evidence the generator's maker carried itself.
The pad, the robot and the smoke evacuator
The split pad of Chapter 8 and the generator's monitor that watches it can come from two makers. Since its sixth edition of 2017, IEC 60601-2-2 has required monitored pads for adults, so the two must work as one circuit (Chapter 25). A recall shows what follows when they do not. In May 2016 Conmed recalled 6,640 pediatric split pads because a version "may not be compatible with some electrosurgical generators", which put patients at risk of undetected pad lift and a burn. FDA classed the recall as Class II, a grade below its most serious, and recorded the cause as device design (Chapter 26).
A robot puts its own controller between the surgeon's command and the generator's output (Chapter 7). Intuitive's summary for the 2024 clearance of its E-200 describes a separate generator that exchanges "control and status messages" with the da Vinci system through an Ethernet cable. Medtronic's summary for the December 2025 clearance of its Hugo robot shows the FT10 on the robot's tower, with one console pedal for each output, monopolar and bipolar. Neither summary says how a press of the pedal becomes an output.
The surgeon still commands the energy, as Chapter 7 showed, but the command crosses two products and their data link, and in each system one company makes both products. Such a link is what FDA's 2017 guidance on interoperable devices addresses, since its scope includes "exerting command and control over one or more medical devices". FDA asks a submission to discuss "each externally-facing electronic interface" and to verify that the interface meets its design specifications, and it asks the labeling to give "specifications for each interface". Intuitive's summaries list that verification and a human factors evaluation: a study of how users work with the generator in its integrated, backup and standalone set-ups.
A smoke evacuator meets the generator at the activation signal. Medtronic states that its RapidVac, once synchronized, evacuates only while the pencil is activated. It names two ways to synchronize: an interlink cable to its ForceTriad, Force FX-C and Force EZ-C generators, and an activator that works "by recognizing the cut and coag current that is being emitted", which its 2020 information sheet lists for generators other than Valleylab's. The cable is an interface one maker specifies at both ends, while the activator senses the current from outside. The activation path can also fail at the instrument. In October 2024 Stryker recalled 2,620 of its Neptune SafeAir smoke evacuation pencils, which could activate without manual input. FDA classed the recall as Class II and recorded its cause as under investigation by the firm, and Stryker extended the recall by 47 lots in September 2025.
The room: electromagnetic compatibility
The room holds equipment that the generator's maker never sees: monitors, other generators, and pacemakers and defibrillators inside the patient (Chapter 19). Their shared interface is electromagnetic compatibility, or EMC, which standards set: IEC 60601-2-2 applies the EMC standard, IEC 60601-1-2, with modifications of its own. Emission limits raise a particular question for a surgical generator, because its output is radio-frequency energy delivered on purpose. FDA's guidance calls surgical generators "intentional emitters" and asks for emissions testing only in the idle state. IEC's interpretation sheet of November 2025 states that a generator belongs to Group 2 of CISPR 11, an emissions standard, the group for equipment that generates radio-frequency energy on purpose. The sheet asks that an idle generator meet the limits of Group 1, the group for other equipment. In both documents, the emission limits therefore apply to the idle generator and not to its output during activation.
The accessories enter the whole evaluation, emissions and immunity alike. The 2023 amendment states that EMC is evaluated with active accessories in their least favorable configuration, and that this configuration is considered in setting their maximum permissible length. The amendment's rationale gives the reason: accessories act as antennas. IEC's interpretation sheet adds that neither the generator nor its accessories are exempt from IEC 60601-1-2. The amendment itself leaves the maker to specify how compliance is checked during the immunity tests. Kirwan's instructions of December 2019 already warned users of its AURA 70 that longer handpieces or cables, "or those produced by other manufacturers, could adversely affect the EMC performance of the unit".
The hospital owns part of this interface, because it decides where the current flows. Schulman and colleagues measured the effect of that choice in a prospective study reported in 2019, in which the return pad was placed by protocol in 144 patients with implantable defibrillators. Boston Scientific, a device maker, supported the study with an unrestricted grant. Monopolar electrosurgery above the umbilicus caused electromagnetic interference in 14 of 70 patients, 20 %, and clinically meaningful interference in 5. At or below the umbilicus it caused interference in 1 of 40 and meaningful interference in none. In cardiac surgery, with the pad under the body, it caused interference in 23 of 34. The surgical site and the pad's position therefore decided how close the current ran to the implant.
Reuse and service: specifications the hospital executes
Reprocessing instructions are a specification that the maker writes and the hospital executes. FDA sets out what that specification must contain in its reprocessing guidance, issued on 17 March 2015 and final as of October 2026. The guidance's six criteria ask that the instructions reflect the device's intended use, advise thorough cleaning, name an appropriate microbicidal process, be technically feasible with legally marketed equipment, and be comprehensive and understandable. FDA also asks the maker to state "how many times the device can be reused, based on testing", the instrument's use life, and to give users a way to tell when it is exceeded. For electrosurgical instruments, FDA's electrosurgical guidance adds what that testing must show: insulation integrity and mechanical strength "after multiple reuse and reprocessing cycles".
Use lives differ by instrument. Beside Covidien's 50 uses, Intuitive's instrument listing of November 2023 gives 10 uses for its da Vinci monopolar scissors, hook and spatula, 14 for its bipolar forceps, graspers and dissector, and single use for its vessel sealers. Between uses, inspection is the hospital's control, and FDA's recommended warning states its limit: "visual inspection alone may not be sufficient to ensure that the insulation is intact". AORN's summary of its 2026 guideline describes active electrode monitoring in laparoscopic procedures as an early warning of insulation failure and stray energy, a check made during the procedure rather than between uses. Chapter 7 gave the failure rates when reused instruments were tested, and Chapter 18 the burns that follow.
Biomedical engineers test a generator in service. IEC 62353, the standard for the recurrent test and the test after repair, stands at edition 2.0 of 2014 as of October 2026, and its abstract describes safety tests, such as leakage, and no test of output. Output follows the maker's procedure: Medtronic's training page for the FT10 lists test loads of 50, 100, 200, 300 and 500 Ω and an electrosurgical analyzer. Kirwan's instructions for the AURA 70 ask for an inspection at least once a year, with the measured output within 20 % of the set power and the RF leakage current at or below 30 mA. Fluke Biomedical, which sells such analyzers, states that generators are typically tested every six months. Erbe states that its staff and distributors can update a VIO 3 on site over a secured wireless link, so the service port is an interface too, with the cybersecurity duties of Chapter 20.
Draw up an interface table before the first design review, one row for each place where the generator meets another part, product or person. Name the party that writes each specification and those that must meet and verify it, and write each specification so that a test can check it. Add a row for every accessory, robot or evacuator the labeling will list.
| Interface | Owner of the specification | Specification | How it is verified |
|---|---|---|---|
| Generator output to active accessory | Generator maker (maximum output voltage); accessory maker (rated voltage) | Every mode's peak at or below the rating, both as peak values | Dielectric test at 120 % of the rating; each mode's peak set against it |
| Generator to return pad | Generator maker (monitor); pad maker (pad) | Contact quality monitor and monitored pad, IEC 60601-2-2 | Monitor study with each pad the labeling lists |
| Robot console to generator | Robot maker and generator maker | Interface specification for activation and settings | System interface verification; human factors on every set-up |
| Generator to smoke evacuator | Evacuator maker | Synchronizing cable or current sensor | Evacuation starts and stops with activation, for each listed generator |
| Equipment to the room | Each maker for its product; operator for the combination | IEC 60601-1-2 as modified by IEC 60601-2-2 | Idle emissions and immunity, with accessories in their least favorable configuration |
| Reusable instrument to reprocessing | Instrument maker writes; hospital executes | Reprocessing instructions and use life | Validated cleaning and sterilization; insulation and strength after the stated cycles |
| Generator to biomedical engineering | Generator maker | Service manual for output; IEC 62353 for safety | Output into the stated loads; leakage; recorded interval |
- An energy system holds the generator, instruments, pad, switches, evacuator and robot, and must tolerate a changing tissue load, its users and the room's other equipment.
- Peak voltage and rated voltage come from two makers, and a pairing holds only if every mode's peak stays at or below the rating, both read as peak values.
- The operator makes each combination, each maker supplies the information for it, and evidence for a pairing exists only where someone tested that pair.
- Pads, robots, evacuators, the room, reprocessing and service are interfaces that one party specifies and another executes, each needing a named owner and a test.
From requirement to evidence.
+ The questionHow does a team turn "seal vessels up to 7 mm" into requirements, tests and records that prove the device does it, and keep that proof true after every change?
From a claim to design inputs
The claim that a sealer seals vessels up to and including 7 mm cannot fail a test as it stands. It names no vessel type, no state in which the diameter is measured, no pressure the seal must hold and no number of seals to try. Valleylab's 510(k) summary for a LigaSure clearance of January 2005 describes the sealer as "capable of sealing vessels up to, and including, 7mm" (Chapter 9). Twenty-one years later Medtronic states the same limit for the LigaSure sealer of its Hugo robot (Chapter 23). FDA's guidance for bipolar vessel sealers of 15 August 2016 supplies one missing piece, asking the indication to state "a maximum vessel size measured in the rounded, perfused state in a normotensive subject", one with normal blood pressure. The vessel types, the pressure and the number of seals remain the team's work.
Design controls are the process that does that work. FDA's design control guidance of 11 March 1997 draws them as a waterfall. User needs become design inputs, the requirements the design must meet, and the design process turns these into design outputs. Verification checks each output against its input. Validation goes further: the guidance describes it as asking "whether devices produced in accordance with the design actually satisfy user needs and intended uses". FDA still lists the guidance as final as of October 2026, although it was written for the design-control section of the former US rule (Chapter 26). FDA's quality management system regulation replaced that rule on 2 February 2026 and rests on ISO 13485, the international standard for a device maker's quality management system (Chapter 26). ISO 13485 sets out the same steps, so the waterfall of 1997 still describes the work, and FDA's slides on the standard state that outputs must "meet the input requirements".
Requirements at four levels
A requirement is useful only if a test can fail it, and the claim becomes one first for the whole system. FDA's sealer guidance names the three system tests of Chapter 9. Burst pressure is measured for each mode and representative instrument, against the predicate, across the vessel sizes of the indication and the vessel types the team justifies. Thermal spread is measured by histology under magnification, two dimensions being "minimally acceptable". A chronic animal study follows the seals for at least three weeks in at least five animals, against success criteria "pre-defined and clinically justified", such as no vessel leaks and normal healing.
No single part makes the seal, so each system requirement is divided among the parts that produce it. The jaws must apply the controlled pressure of Chapter 9, and the instrument's rating must cover every peak the generator can deliver (Chapter 21). The generator must deliver its sealing mode and meet the essential performance of Chapter 20. For the FT10, Medtronic rates that mode at 20 Ω and up to 350 W, with a peak voltage of at most 244 V (Chapter 9). The software must infer the end of the seal from the impedance curve of Chapter 9, and the interfaces of Chapter 21 become requirements of their own. Each part can then be verified on its own, and the system test shows whether together they meet the claim.
Risk controls in their order, on one hazard
Risk management runs inside the design, because every control it chooses becomes a requirement with a test. As of October 2026, ISO lists ISO 14971:2019 as current, confirmed on 7 March 2025, and FDA's training slides quote its definitions. A hazard is a "potential source of harm", a hazardous situation the "circumstances in which people, property or environment is/are exposed to one or more hazards", and harm "injury or damage to health of people, or damage to property or environment". Every hazard of Chapters 18 and 19 enters the risk file in these terms. On the return pad of Chapter 8, the hazard is the monopolar current that leaves through the pad. The hazardous situation is a pad partly lifted during activation, so that the current crowds into the edge still in contact, and the harm is a burn there.
ISO 14971 ranks the ways to control a risk, and BSI gives the order in a white paper on the 2019 edition. BSI is a notified body, one of the organizations designated to assess devices under EU law (Chapter 27). In BSI's words, "The first and preferred option is to eliminate the risk", by inherently safe design; the second is protective measures; and the third, where those reduce the risk too little, is information for safety. UL Solutions, a certification body, gives barriers and alarms as examples of protective measures, and a written warning as an example of information for safety, in a summary of the 2019 edition published in 2020. On the lifting pad the three options fall on three parts of the system. The design is the pad's area, since Chapter 8 showed that an adult pad stays below the threshold of necrosis even with two thirds lifted, although the skin still in contact then heats nine times as fast. The protective measure is the split pad and its monitor, which stops the generator when the impedance between the halves rises. The information is the instruction not to cut or fold a pad.
Each control needs evidence of its own. The pad's area is tested by the return-electrode heating test, which allows the skin a rise of no more than 6 °C (Chapter 25). The monitor needs a study with each pad the labeling lists, and Intuitive's 2023 summary, for example, reports a contact-quality-monitor study and compatibility with an Erbe split pad (Chapter 21). Information works only if people act on it, and the knowledge test of Chapters 8 and 18 illustrates its weakness: 10 % of the 48 leaders thought a pad should be cut to fit a child, the opposite of the instruction not to cut or fold a pad. BSI adds that each control is verified once for its implementation and once for its effectiveness, and the remaining risk is then estimated again.
Verification, validation and use
Each level of test shows something the others cannot. Bench tests into resistors show that the generator delivers its declared output across the load range (Chapters 10 and 25), and say nothing about tissue. Burst and histology tests show what a seal does when it is tested, and only the chronic animal study shows whether it survives healing. Intuitive's summary for the 2024 clearance of its E-200 generator sorts its evidence the same way: bench verification of software, cybersecurity, EMC, electrical safety, system interface and instrument compatibility, then validation by simulated clinical use in pigs and by a human factors evaluation.
Use is the level that bench and animal tests miss. FDA's human factors guidance of 3 February 2016 defines a use error as a "user action or lack of action that was different from that expected by the manufacturer", and a critical task as one that, "if performed incorrectly or not performed at all, would or could cause serious harm". Formative evaluations test the user interface while it is still changing. Human factors validation testing comes at the end, on a user interface that "represents the final design", and the guidance states that "in general, the minimum number of participants should be 15" for each distinct user population. The Johner Institute, a consultancy, notes that IEC 62366-1 calls the same test a summative evaluation. On a robot one design has several set-ups to validate: Intuitive's 2023 summary reports human factors work on the E-200's integrated, backup and standalone set-ups, and Medtronic's Hugo summary lists a human factors validation, although neither gives results (Chapter 21).
One requirement traced from claim to label
Traceability links each claim to the requirements that serve it, the tests that verify them and the label line that limits it. FDA's sealer guidance builds the lowest link into its burst-pressure report: a table listing each vessel sealed with its burst pressure, vessel type, diameter and device sample number, so that each seal points to the instrument that made it. The top link is the label warning the guidance recommends: "DO NOT USE this instrument on vessels larger than __mm in diameter." The record that holds the chain is the design and development file, which ISO 13485 requires for each device type or family.
FDA's guidance sets no pass value, so the team chooses 360 mmHg, three times a systolic pressure of 120 mmHg, the industry convention of Chapter 9. The design input reads that seals of vessels up to 7 mm, measured rounded and perfused, hold at least 360 mmHg in each mode with each representative instrument. FDA sets no sample size either, so the team picks a confidence and a reliability and plans a test that allows no failure. A team that wants 95 % confidence that at least 90 % of seals hold, and treats each seal as an independent trial, reasons from the chance of a clean run. If only 90 % held, the chance that 28 seals in a row all passed would be 0.9 multiplied by itself 28 times, about 0.052, just above 5 %. For 29 seals it is about 0.047, below 5 %, so 29 seals without a failure are the smallest test that supports the requirement at that confidence. For 95 % of seals at 95 % confidence the count is 59, since 0.95 multiplied by itself 59 times is about 0.049, and for 90 % at 90 % confidence it is 22, since 0.9 multiplied by itself 22 times is about 0.098, below 10 %. The count also multiplies with the configurations the claim covers: a hypothetical sealer with one mode and two jaw sizes needs 29 seals of each size, 58 in all, with the predicate sealed beside them. The counts are the team's choices of confidence and reliability turned into seals, and a single failure leaves the requirement unproven under this plan.
Twenty-nine seals that held support the requirement for the vessels, sizes, modes and instruments sampled, and for nothing outside them. Chapter 9 showed what can happen outside them. In an independent bench comparison that Okhunov and colleagues published in 2018, four of five devices lost about half their seal strength or more on vessels larger than 7 mm. The label must therefore carry the limit the test carried.
The loop from the field
Proof made before launch must stay true afterward. A complaint is evidence of a possible failure to meet a specification, and FDA notes that ISO 13485 addresses the investigation of complaints and the corrective action that follows (Chapter 26). An investigation that finds a cause leads to a correction, a design change or both, and every change reopens the trace. The team finds the requirements the change touches, repeats the tests behind them, updates the risk file and the label, and assesses whether the change needs a new submission, by the US and EU criteria of Chapters 26 and 27.
The Varipulse pulsed field catheter shows the loop running. In February 2025 Johnson & Johnson's unit Biosense Webster sent a correction that updated the instructions for use, after 4 of 132 patients in the US evaluation had a stroke or a transient ischemic attack, a brief stroke-like episode (Chapters 15, 24 and 26). In ISO 14971's terms, the update was information for safety, the last of the three options. The catheter had reached the market by premarket approval, and FDA's records for that approval then show two later changes (Chapter 26). A label change to the recommended irrigation flow rate was approved on 26 June 2025. A software update to the TRUPULSE generator was approved on 23 December 2025 "to facilitate automatic delivery of irrigation at 30 mL/min upon initiation of energy delivery", which moved a control from the instructions into the design. FDA's records do not link the irrigation changes to the strokes, although Johnson & Johnson reported a neurovascular event rate from its complaint data in its release on the flow rate (Chapter 24).
A correction can come before its cause is known. When Megadyne corrected its reusable return electrodes in 2023 after reports of burns, it asked hospitals to confirm that staff follow the instructions for use and to post its visual aids, while it continued to evaluate the root cause (Chapter 26). Both corrections began as information for safety, and the trace then shows which requirement a later fix must change and which tests it reopens.
Keep one matrix per claim, one row per requirement, with an identifier in every cell, so that any test report, risk control or label line leads back to its claim. Fill the columns below for the 7 mm claim, then add rows for thermal spread, the chronic study and each interface. When a change touches a cell, rerun every test that depends on it before the change is released.
| Column | The 7 mm sealing claim, burst-pressure row |
|---|---|
| Claim | Seals vessels up to and including 7 mm |
| User need | Vessels up to 7 mm, sealed and divided, stay closed under arterial pressure |
| Design input | Seals up to 7 mm, measured rounded and perfused, hold at least 360 mmHg in each mode with each representative instrument |
| Design output | Jaw pressure, sealing mode, end-point algorithm, instructions for use |
| Verification | Zero-failure burst test, 29 seals per mode and instrument, beside the predicate |
| Validation | Chronic animal study, at least 3 weeks in at least 5 animals; human factors validation |
| Risk control | End-point detection in the generator; warning on vessel size |
| Label line | DO NOT USE this instrument on vessels larger than 7 mm in diameter |
- Design controls turn a claim into design inputs a test can fail, divided among the parts; verification shows that outputs meet inputs, and validation that the device meets its users' needs.
- FDA's sealer guidance sets no pass value or sample size, so the team chooses them: a zero-failure burst test at 95 % confidence and 90 % reliability needs 29 seals.
- ISO 14971 prefers inherently safe design, then protective measures, then information for safety, and on the lifting pad these are the pad's area, the monitor and the instructions, each with its own test.
- A trace from claim to test to label shows which tests a change must repeat, and the field feeds changes back, as Varipulse's instruction update was followed by new generator software.
+ Part VIII · The field in 2026
Who makes it, and what is changing.
On three consecutive days in September 2026, FDA cleared three energy devices from three makers: a LigaSure vessel sealer for Medtronic's Hugo robot on 15 September, Olympus's SOLTIVE thulium fiber laser systems on 16 September and Ethicon's HARMONIC 1100 ultrasonic shears on 17 September. Each energy in this book has its own makers and its own public record, and both change from month to month. This part is dated October 2026. It sets out who makes each kind of device and what the regulator's record shows for it, then what is changing at the edge of the field and what is likely to stay hard.
The makers, energy by energy.
+ The questionWhich companies make each kind of surgical energy device in October 2026, and what does the regulator's record show for each?
The record counts decisions rather than products
Between 1 January 2024 and 20 October 2026, FDA recorded decisions on 153 premarket notifications under the product code for electrosurgical devices. A product code is the three-letter code under which FDA files every record for one type of device. The electrosurgical code, GEI, holds more than surgical generators, because aesthetic RF devices, endovenous RF devices, electrodes and pencils share it. The 153 records were filed under 98 forms of applicant name, and one company can appear under several of them: Olympus, for example, under four name forms of three legal entities. One record can cover a generator together with many instruments. FDA publishes these data through its openFDA service and states that they are unvalidated.
Eight makers whose surgical generators, sealers, pencils or instruments this chapter describes appear among the applicants. Erbe holds 6 of the 153 records, Stryker's two units 5, Covidien 3, Olympus 4 under four name forms, LivSmed 3, Intuitive 2, Ethicon 2 and Megadyne 1. Adding them gives 26 records. Dividing 26 by 153 gives about 0.17, so these makers hold about one record in six. The other 127 records belong to the remaining applicants, many of them makers of aesthetic RF devices. A count under one product code therefore measures what the code holds, and the part of it that belongs to surgical energy has to be worked out applicant by applicant.
The makers are therefore best followed energy by energy, through the records of their current products and what each company states about them. The record's words are kept throughout: a 510(k) is cleared, a De Novo request is granted, and a premarket approval, or PMA, is approved (Chapter 26). A figure, a count or a "first" that comes from a company is given as that company's statement.
Generators, sealers and instruments that pass current
Medtronic's annual report for its fiscal year to 24 April 2026 lists its electrosurgical products as the Valleylab FT10 and FX8 energy platforms and its LigaSure vessel-sealing instruments. Chapter 8 read the FT10's monopolar modes, and Chapter 9 its sealing. Two of its records in the period are a pencil and a sealer. FDA cleared a smoke-management pencil from the company's Covidien unit on 1 July 2024, and Medtronic states that the LigaSure sealer cleared for its Hugo robot on 15 September 2026 seals vessels up to and including 7 mm. Erbe holds the most records of the eight makers, and its generators also drive other makers' instruments. FDA cleared Erbe's VIO 3n generator with its accessories on 29 August 2025, and Chapter 25 reads its public summary. FDA also cleared Erbe's HybridTherm System on 5 August 2024, and Intuitive lists an Erbe generator, the VIO dV, among those that drive its Vessel Sealer Extend.
FDA cleared a generator from Johnson & Johnson's Ethicon, the ETHICON Total Energy System, on 17 January 2025. The company states that it has since renamed the generator DUALTO with no change to its intended use or regulatory status, and that DUALTO powers several energy modalities and runs any two instruments at once whose maximum power draw is 150 W or less. The same company also lists two other generators as separate products: GEN11, for its HARMONIC ultrasonic and ENSEAL bipolar devices, and a MEGADYNE generator. FDA cleared LivSmed's ArtiSeal, a vessel sealing system with its own generator, on 15 November 2024. Stryker's records since 2024 include PhotonBlade 3, an electrosurgical blade with smoke evacuation cleared on 21 October 2024, and a remanufactured LigaSure Maryland sealer from its reprocessing unit, cleared on 18 June 2026.
BOWA's latest generator record under the electrosurgical code is the ARC 400, cleared in 2020. KLS Martin's latest is a generator sold under another company's name, the Codman Electrosurgical Generator, cleared in 2019, and the latest records of CONMED, for its CleanSeal bipolar sealer, and of Apyx Medical, for its Renuvion helium-plasma handpieces, date from 2023. None of the four holds a record under these names and this code with a decision date from January 2024 to October 2026. A product cleared earlier can stay on the US market, and a company can file under another code or another legal name, so a quiet record shows only that nothing new was cleared there.
Ultrasonic instruments and lasers
Ethicon, Olympus and Medtronic each sell ultrasonic instruments of the kind Chapter 11 describes. FDA cleared Ethicon's HARMONIC 1100 Shears on 17 September 2026. FDA cleared Olympus's THUNDERBEAT II, with a transducer that Olympus calls cordless, on 7 October 2025, and its front-actuated THUNDERBEAT shears on 1 July 2026. Olympus's word means that the cable has moved to the base of the handle, not that the device runs on a battery (Chapter 11). Medtronic lists Sonicision 7, which it calls a cordless ultrasonic dissection system, in its 2026 annual report.
The product code for ultrasonic surgical instruments, LFL, holds eight 510(k)s decided between January 2024 and October 2026, and FDA still lists the code as unclassified as of October 2026 (Chapter 26). All eight are aspirators or similar devices: two each for Integra's CUSA Clarity and Stryker's Sonopet iQ, and one each for Misonix's neXus, Trice Medical's Tenex, Hunan Handlike and Shanghai Saints Sages. None of them comes from Ethicon, Medtronic or Olympus: the clearances of HARMONIC 1100 and of both THUNDERBEAT shears appear under the electrosurgical code, GEI, so a search under LFL alone would miss them.
The surgical-laser code, GEX, lists more than 100 applicants over the same period, most of them makers of aesthetic lasers. Among makers of surgical lasers, El.En holds seven records, Quanta System four, and Boston Scientific, Olympus and IPG Medical three each. In 2021 Boston Scientific bought Lumenis's surgical business, with its urology and ENT lasers, its fibers and the MOSES pulse technology of Chapter 12. Lumenis Be also holds three records under the code, but for aesthetic devices, because it kept Lumenis's aesthetic and vision businesses after that sale. FDA cleared Boston Scientific's Flexiva Pulse laser fibers on 9 July 2025 and its MOSES Raydar holmium laser system on 18 May 2026. FDA cleared Olympus's SOLTIVE thulium fiber laser systems three times, on 23 August 2024, 24 July 2026 and 16 September 2026.
Ablation by heat, cold, pulses and sound
Seven applicants hold the 510(k)s for microwave ablation decided between January 2024 and October 2026. Among them is Varian, whose IntelliBlate system FDA cleared on 19 July 2024. Medtronic's annual report lists the Emprint microwave system of Chapter 13. AngioDynamics states in its annual report that it discontinued its RF ablation product line as of 29 February 2024. It keeps its Solero microwave system, indicated for the ablation of soft tissue during open procedures, and its NanoKnife IRE system (Chapter 15). The cryosurgical code, GEH, also covers devices for skin and warts; its applicants in the period include IceCure, Boston Scientific and Erbe, and IceCure's ProSense holds the De Novo grant for low-risk breast cancer described in Chapter 14.
Pulsed field ablation first reached US patients through four premarket approvals, from Medtronic's PulseSelect on 13 December 2023 to Johnson & Johnson's Varipulse on 6 November 2024, with Boston Scientific's FARAPULSE and Medtronic's Sphere-9 between them, and the plate below dates each. Approvals continued after those four, and FDA approved further systems from Kardium and Abbott in 2025 (Chapter 15). Johnson & Johnson announced on 8 July 2026 that FDA had approved its Dual Energy THERMOCOOL SMARTTOUCH SF, which delivers RF and pulsed field energy through a single catheter. A maker's own figure gives a sense of the scale of use: Boston Scientific states in its 2025 annual report that more than 500,000 patients have been treated with FARAPULSE. A European marking can also come before a US approval. Medtronic announced on 23 January 2026 that its Affera Sphere-360 catheter had received CE marking in Europe, and as of October 2026 no FDA approval of it had been found. Chapter 24 follows Varipulse's first year in the US and its newer European version.
HistoSonics' Edison, the histotripsy system of Chapter 16, holds a De Novo grant for liver tumors, and Chapter 16 follows its kidney request. The company's own announcements give a sense of its scale. HistoSonics announced on 7 August 2025 that a consortium of investors had agreed a management-led acquisition of a majority stake, valuing the company at about US$2.25 billion, and in June 2026 it announced financing at a valuation of US$3.75 billion. Among the makers of other ultrasound ablation systems, Insightec announced on 8 July 2025 that FDA had approved its Exablate Neuro for staged treatment of both sides of the brain in advanced Parkinson's disease. Profound Medical states that its TULSA-PRO is cleared by 510(k) and reports 78 installed systems at the end of 2025. EDAP states that FDA cleared its Focal One i System in November 2025.
Robots and the energy they carry
Intuitive announced FDA's clearance of its da Vinci 5 system on 14 March 2024 and states that the system integrates an electrosurgical unit and insufflation. Its E-200 generator holds a clearance dated 5 August 2024, and Intuitive states that SynchroSeal, the 5 mm sealer of Chapter 17, runs on the E-100 and the E-200. The robots that followed show the De Novo route of Chapter 26 at work, in which the first device of a kind creates a product code that later devices follow. FDA granted CMR Surgical's Versius a De Novo on 11 October 2024, creating the product code SCV for modular electromechanical surgical systems. Medtronic's Hugo was cleared by 510(k) under that code on 3 December 2025, with Versius as its predicate, and as of October 2026 its US indication remains urologic procedures (Chapter 24). FDA cleared Versius Plus on 16 December 2025, and CMR's investor Cambridge Innovation Capital reports that the clearance covers gallbladder removal.
On 12 October 2026 Stryker announced that it would invest about US$400 million in CMR Surgical, with an exclusive option to buy all of the company and the role of exclusive US sales agent for Versius, subject to regulatory approvals and closing conditions. SS Innovations, an Indian company, states that its 510(k) for the SSi Mantra robot was under FDA review for gallbladder removal as of October 2026. FDA asked for more information, the company responded on 5 October 2026, and it expects the review to be completed in the first quarter of 2027. Neither CMR's nor SS Innovations' announcements name an energy instrument or a generator.
India's makers and the record abroad
In FDA's open data, the only applicant with an Indian address under the surgical-laser code is Allengers Global Healthcare, of Derabassi in Punjab. It holds two clearances, for a holmium laser on 18 August 2023 and for a thulium fiber laser on 17 December 2024. The only Indian-address applicant ever recorded under the electrosurgical code is Ace Medical Devices of Pune, for resection and vaporization electrodes cleared in 2016, and as of October 2026 no Indian electrosurgical generator or vessel sealer appears in those data. India's Department of Pharmaceuticals publishes the approved applicants of its Production Linked Incentive scheme for the domestic manufacture of medical devices. Its list as on 6 November 2023 names one surgical-energy system: a thulium fiber laser, listed for Indovasive Private Limited together with laser probes and loop and cutting electrodes for prostate resection.
Medtronic launched its Valleylab FT10 electrosurgical generator and its FT10 vessel sealing generator in India on 30 September 2025 and described both as "USFDA-approved", although a device that reaches the US market by 510(k) is cleared. BioSpectrum India, a trade journal, reported on 9 June 2026 that Biorad Medisys of Pune had launched a 20 W thulium fiber laser for kidney stones, adding to its 40 W and 75 W units. As of October 2026, India's 2021 classification lists place general-purpose electrosurgical units in class C, and Chapter 27 sets out what that class requires.
Before a design review or a competitive analysis, list the devices of the same energy from the regulator's database rather than from press releases: product code, record type, applicant as filed, decision date and the indication in the summary. Merge the name forms of one company, and check whether a count covers products or only records. Label every figure that comes from a company as the company's own, and give every status with the date on which it was checked.
- FDA's open data hold 153 premarket notifications under the electrosurgical product code from January 2024 to October 2026, and the eight surgical makers this chapter describes hold about one in six of them.
- FDA cleared generators, sealers, ultrasonic instruments, laser systems or laser fibers from Medtronic, Ethicon, Olympus, Erbe, Intuitive and Boston Scientific in that period, while all eight records under the ultrasonic code, LFL, went to aspirators and similar devices, and the ultrasonic shears appear under the electrosurgical code.
- Four pulsed field systems were approved in the US between December 2023 and November 2024 and two more in 2025, while AngioDynamics left RF ablation in 2024, an investor consortium's majority stake in HistoSonics was announced in 2025 and Stryker announced an investment in CMR Surgical in 2026.
- India's makers appear in FDA's data with lasers and resection electrodes, and as of October 2026 no Indian electrosurgical generator or vessel sealer holds a US clearance there.
The frontier.
+ The questionWhat is changing in surgical energy, and what is likely to stay hard?
Non-thermal ablation goes mainstream
In January 2025, about two months after FDA approved it, Johnson & Johnson paused US cases with its Varipulse pulsed field catheter (Chapter 15). The correction that followed came after 4 of the 132 patients in the catheter's US evaluation had a stroke or a transient ischemic attack, about 3 % where 1 % or less was expected. FDA classed the correction as the most serious type of recall (Chapter 26 gives FDA's record). The case for pulsed fields rests on sparing the esophagus and the phrenic nerve, which heat and cold can injure, but the first months of one approved system showed that a new energy can bring a risk of its own.
The correction left the catheter on the market. Johnson & Johnson announced on 14 July 2025 that FDA had approved an updated irrigation flow rate for the Varipulse platform, and by November 2025 the company stated that more than 25,000 Varipulse procedures had been performed worldwide. The catheter's newer version reached Europe first. Johnson & Johnson states that Varipulse Pro received CE marking in March 2026, and it launched the catheter in Europe on 7 April 2026. Varipulse Pro had no US approval as of October 2026. On 24 April 2026 Johnson & Johnson announced PERSIGMA, a randomized trial of Varipulse Pro against Boston Scientific's FARAPULSE in persistent atrial fibrillation, with up to 466 participants at 50 sites.
In its release of 14 July 2025, Johnson & Johnson gives a neurovascular adverse event rate for Varipulse of less than 0.5 % and states that more than 10,000 procedures had been performed worldwide. The company also states that the rate was calculated from internal customer-reported data and complaints rather than from a prospective clinical investigation, and calls it preliminary and pending publication. A rate built from complaints counts only the events that someone reports, while the 4 of 132 came from a defined group of patients in an evaluation. The two figures measure different things, and neither shows by itself whether the risk has fallen.
One published safety record covers 17,642 patients, far more than the 132 of the Varipulse evaluation. Ekanem and colleagues published it in 2024 as MANIFEST-17K, a retrospective survey of the use of Boston Scientific's FARAPULSE at 106 centers in Europe and Israel after the system reached the European market. Its cases date mostly from early 2022 to June 2023, before FDA approved the system. It had no external funding, and many of its authors report ties to the company. Its findings bear on the case for pulsed fields made above: it found no esophageal complications and no persistent palsy of the phrenic nerve (Chapter 15).
MANIFEST-17K counted 173 major complications among 17,642 patients. Dividing 173 by 17,642 gives about 0.0098, and multiplying by 1,000 gives about 9.8 major complications per thousand patients, the 0.98 % the registry reports. Transient palsy of the phrenic nerve, in 11 patients, works out the same way at about 0.6 per thousand, and esophageal complications, of which there were none, at zero. The denominator decides how much one patient moves a rate. One patient more or less changes the registry's rate by 1,000 divided by 17,642, about 0.06 per thousand, while in the Varipulse evaluation of 132 patients one patient changes it by 1,000 divided by 132, about 7.6 per thousand. Its four strokes or transient ischemic attacks come to about 30 per thousand, against an expected 10 or fewer. The two rates cannot rank the catheters, because they count different events in different kinds of study; the arithmetic shows that a rate from 132 patients rests on far fewer cases than a rate from 17,642.
In tumor ablation, histotripsy received its first US authorization, a De Novo grant for liver tumors, on 6 October 2023. The evidence behind it came from HistoSonics' own single-arm trial in 44 patients (Chapter 16). HistoSonics has since sought an indication beyond the liver. It announced a De Novo request for kidney tumors on 11 May 2026, and no decision had been announced as of October 2026. In June 2026 HistoSonics itself still described the kidney request as pending, with use outside the liver "limited to investigational use". In Europe, HistoSonics announced CE marking of its Edison system for liver tumors on 13 July 2026, and Chapter 16 follows both steps. Investors had backed the company before both steps: an investor consortium's purchase of a majority stake had been announced in August 2025 (Chapter 23). Elsewhere, a clearance announced in December 2024 extended one IRE system's indication to name prostate tissue, and cryoablation, an older energy, gained a US indication in low-risk breast cancer through a De Novo granted on 3 October 2025 (Chapter 14).
What pulsed fields and histotripsy share is the idea of Chapter 6: for some tasks, the harm that heat does to the collagen scaffold of vessels and ducts matters more than how well it kills cells. Both energies now hold US authorizations. Histotripsy holds its grant for the liver, and four pulsed field systems were approved in the US between December 2023 and November 2024, with two more in 2025 (Chapter 15). The question for the next few years is how far the idea extends beyond the heart and the liver, and what evidence each new indication will need.
The rules and standards that these devices must meet changed in the same years. The plate marks those that changed between 2023 and 2026: the US quality regulation in force since 2 February 2026, a 2025 edition of the ultrasonic standard IEC 61847, an April 2026 amendment to the laser standard IEC 60601-2-22, and FDA's May 2026 database entry for edition 6.1 of the electrosurgical standard. Chapters 25 to 27 give each with its date and what it changes.
Robotic energy
Robotic surgery is changing how energy is controlled as much as where it is delivered. Intuitive's E-200 generator holds a US clearance dated August 2024, and Intuitive states that it works with its da Vinci systems with the energy settings controlled from the surgeon console, and that it is also available as a backup and standalone unit. The company also describes force feedback on some da Vinci 5 instruments, which applied up to 43 % less force to tissue in its preclinical trials.
Medtronic's Hugo is newer to the US: it was cleared there on 3 December 2025 for urologic procedures, and as of October 2026 those remained its only US indication. On 3 June 2026 Medtronic announced that it had submitted 510(k)s to extend Hugo to general surgery, including hernia repair, and to gynecology, and as of October 2026 both were pending. Medtronic states that the LigaSure sealer cleared for Hugo on 15 September 2026 is designed for use only with its Valleylab FT10 generator. Hugo's 510(k) summary shows that generator on the robot's tower, so the robot's sealer runs on a generator that the robot itself carries (Chapter 21).
As the robot takes over more of what the surgeon's hand once did, the questions of Chapter 7 grow: which controller decides when energy flows, how the robot and the generator are verified together, and how reused wristed instruments keep their insulation. Chapter 21 treats the first as an interface between products, and Chapter 7 gives the insulation failures of robotic instruments.
Instruments, lasers and power electronics
In urology, thulium fiber lasers now compete directly with holmium for stones and prostate work. The engineering difference is efficiency: a diode-pumped thulium fiber laser turns about 12 % of its input into light and can be air-cooled, against under 2 % for a flashlamp-pumped holmium laser that needs water cooling (Chapter 12). The difference shows in the mains supply each needs. Olympus specifies a supply of 100 to 240 V at 1,200 VA for its 60 W air-cooled SOLTIVE thulium systems, while Boston Scientific lists up to 46 A at 200 to 240 V for its 120 W Pulse 120H holmium system (Chapter 12). FDA cleared new thulium fiber systems from Olympus in 2024 and 2026 and from Allengers, an Indian maker, in December 2024 (Chapter 23). Holmium systems continue to develop as well, and FDA cleared Boston Scientific's MOSES Raydar holmium laser system in May 2026 (Chapter 23).
Instruments change in smaller steps. Olympus states that THUNDERBEAT II, cleared on 7 October 2025, adds an ultrasonic-only mode to its hybrid jaw and moves the transducer's cable to the base of the handle (Chapter 11). The generators that drive such instruments are controlled by software today, and they carry the software and cybersecurity obligations of Chapters 20 and 26. Faults in such generators reach the recall record as well. In July 2025 Olympus began a recall of 155 ESG-410 electrosurgical generators that could show a system error and then restart, sometimes repeatedly, and FDA classed it as Class II (Chapter 26).
What stays hard
Some problems in this guide are not going away, because they come from physics rather than from technology.
The effect is still inferred rather than measured. A sealer infers a seal from impedance, an RF ablation generator infers a zone from temperature or impedance, and the margin of an ablation is judged afterward on a scan. Direct, real-time measurement of the tissue effect remains rare. MRI thermometry during focused ultrasound, the bubble cloud of histotripsy and the visible ice ball come closest, and even the ice ball's visible edge marks 0 °C, with the lethal edge farther in (Chapter 14).
Energy still escapes out of view. Insulation still fails in reusable instruments, more often in robotic ones, and capacitive coupling is a property of any shaft in a cannula (Chapters 7 and 18). Monitoring helps, and the physics stays the same.
Tissue still varies. Perfusion, water content, collagen content and large vessels change what a given output does. Bench tests in unperfused tissue leave out the heat sink of Chapter 2, so they show the most favorable case.
Evidence is still hard to compare. Many performance studies are written or funded by manufacturers, conditions differ between studies, and each device generation changes the answer. The newest energies show the range. The trial behind histotripsy's liver authorization was its maker's own, the 17,642-patient registry had no external funding but many authors with ties to the maker, and the randomized comparison of two pulsed field catheters announced in April 2026 was launched by one of the two makers. Independent, standardized comparisons remain the scarcest evidence in the field. The words for regulatory status carry a hazard of their own, and Chapter 26 sets out how a request, a clearance, a grant, an approval and a certificate differ.
Treat every claim about a new energy, indication or rate as dated evidence before it enters a design review, a plan or a regulatory file. Record who made it and in which document, whether it is an announcement, a submission, a decision or a result, and its date; for a rate, record the count, the denominator and how the cases were found. Check the status against the regulator's own record before the claim is reused, and date it "as of" the day it was checked.
- The Varipulse correction of February 2025, after strokes or transient ischemic attacks in about 3 % of the patients in the catheter's US evaluation, showed that a new pulsed field system can bring a risk of its own.
- As of October 2026, histotripsy is authorized in the US for liver tumors only and no FDA decision on its kidney request has been announced, while one IRE system's indication names prostate tissue and cryoablation has gained an indication in low-risk breast cancer.
- Intuitive states that its generator's energy settings are controlled from the surgeon console, and as of October 2026 Hugo remains cleared in the US for urologic procedures only, while thulium fiber lasers turn more of their input into light and generators carry software and cybersecurity obligations.
- The effect is still inferred rather than measured, energy still escapes out of view, tissue still varies, and independent comparisons remain the scarcest evidence in the field.
+ Part IX · Quality and regulation, as of October 2026
What the regulators ask to see.
On 25 May 2026, FDA entered edition 6.1 of IEC 60601-2-2, the international standard for high-frequency surgical equipment, in its database of recognized standards, and it will accept declarations of conformity to the earlier edition 6.0 until 1 July 2029. Until that date, a maker may support a premarket submission with either edition. The earlier chapters showed why each requirement in such a standard exists. Chapter 10 explains why a generator's output is measured across a range of loads, Chapter 20 why stray current is measured through a resistor that stands in for a person, and Chapter 8 why the skin under a return pad is held to a small rise in temperature. This part maps three jurisdictions onto those reasons: the standards and the tests first, then the United States, then the European Union and India. In each chapter the mechanism comes before the clause, and every status is stated as of October 2026.
What the bench proves.
+ The questionWhat does compliance testing check on a surgical generator, which editions do regulators accept in October 2026, and what does a public test record show?
Five layers of standards
A resistor of 200 Ω, wired between an electrosurgical generator's output and earth, stands in for a person on the test bench. The high-frequency current that escapes to earth, instead of returning through the generator's own return lead, is measured through it and held to the limit of Chapter 20. That resistor is one detail of one standard, and the standard is one layer of a stack.
At the base are process standards, which govern how a device is designed and made rather than what it does: ISO 13485 for the quality management system, ISO 14971 for risk management, IEC 62304 for the software life cycle, IEC 62366-1 for usability engineering and IEC 81001-5-1 for security in the software life cycle. Above them sits the general standard for medical electrical equipment, IEC 60601-1, which sets the basic safety and essential performance of Chapter 20 and stands, as of October 2026, at edition 3.2 of 2020. Collateral standards add requirements that many kinds of equipment share: IEC 60601-1-2 for electromagnetic compatibility, IEC 60601-1-6 for usability and IEC 60601-1-8 for alarm systems.
On top sit the particular standards, one for each type of equipment, which modify the general standard and override it where the two differ: IEC 60601-2-2 for electrosurgery, IEC 60601-2-22 for surgical lasers and IEC 60601-2-62 for high-intensity therapeutic ultrasound. FDA's electrosurgical guidance also cites IEC 60601-2-18, for endoscopic equipment, when laparoscopic electrodes are tested for capacitive coupling (Chapter 18). Some energies have product standards for their output: IEC 61847 sets how an ultrasonic surgical system's output is measured and declared, and IEC 60825-1 classifies laser products by hazard (Chapter 19).
The stack has a gap. No particular standard for ultrasonic surgical systems was found for this guide, and ultrasonic generators are tested to the general standard, IEC 60601-1, and also to IEC 60601-2-2 where they deliver bipolar current, as hybrid instruments do (Chapter 11).
Which editions count, as of October 2026
Regulators accept specific editions of a standard. FDA lists the editions it accepts as recognized consensus standards, and a maker may support a premarket submission with a declaration of conformity to a recognized edition. For IEC 60601-2-2 the current entry is edition 6.1, a consolidated version of 2023 that was corrected in November 2025. FDA entered it in its database on 25 May 2026 and published the change in a Federal Register notice on 24 August 2026, the date from which the change applies. Edition 6.0, of 2017, is now marked as superseded, and FDA will accept declarations to it until 1 July 2029.
The laser standard shows the same pattern, and its transition ends sooner. FDA recognizes edition 4.0 of IEC 60601-2-22, of 2019, and accepts declarations to edition 3.1 until 4 July 2027. IEC published an amendment to edition 4.0 on 24 April 2026, and no FDA recognition of it was found as of October 2026. Two product standards are recognized in older editions than the latest: IEC 61847 in its edition of 1998 rather than that of 2025, and IEC 60825-1 in its edition of 2007 rather than that of 2014. Under Laser Notice 56, however, FDA does not intend to enforce most of its own laser rules for products that conform to the comparable clauses of the 2014 edition (Chapter 19). Recognition can also be partial. FDA recognizes IEC 60601-2-62, for high-intensity therapeutic ultrasound, in full in its edition of 2013, but IEC 60601-1-2, the EMC collateral, only in part in its edition 4.1 of 2020, leaving out one figure and a note to one table.
Two editions are still being drafted. Montenegro's national standards institute mirrors the IEC's project records, and its copy shows the second edition of IEC 62304 at the committee-draft stage since 7 August 2026, with no forecast date for publication. BSI, the British standards body, states that the merged committee draft for vote of the fourth edition of IEC 60601-1 is anticipated for 2029. A device that does not change can therefore fall behind the recognized editions, and a test report that supported one submission may need updating before the next.
Bench tests for an electrosurgical generator
IEC 60601-2-2 adds tests of its own to the electrical-safety, temperature and EMC tests of IEC 60601-1 and IEC 60601-1-2. Four of them follow from earlier chapters.
Output against load. The generator drives a bank of non-inductive resistors that covers the generator's load range, while a power meter records the output. The results are the power curves of Chapter 10, at maximum and at half setting, and the declared waveforms at rated load. The 2023 amendment to the standard revised the rule on output reduction, replacing its first sentence, and separated the scenarios for simultaneous activation.
High-frequency leakage. The generator stands on an insulating table with its leads laid out as the standard specifies, and the current that escapes to earth is measured through the 200 Ω resistor, against the limit of 150 mA for monopolar outputs given in Chapter 20. The 2023 amendment clarified the set-up: each electrode of an isolated monopolar output is tested first with the output unloaded and then loaded at its rated load.
Return-electrode heating. The standard calls the return pad the neutral electrode. In its heating test a pad carries a current set by the weight category of patient it is rated for, and the skin beneath it must not warm by more than 6 °C. Edition 6.0 required adult return electrodes to be of the contact-quality-monitoring kind, the split pad of Chapter 8. The 2023 amendment reaches below the high-current threshold of Chapter 10, a heating factor of 30 A²s in any 60 seconds. Below that threshold too, heating under the return electrode must now be considered in risk management for any mode whose intended use has an activation duty cycle above 45 %, the share of time the output is switched on (Chapter 10).
Accessories. Active accessories, such as pencils and laparoscopic electrodes, face a high-frequency dielectric test at 120 % of their rated voltage for 30 seconds, with a test waveform whose crest factor rises with the rating, for the reason Chapter 18 gave. They also face a leakage test through their insulation. The 2023 amendment requires the instructions for use to state each accessory's length, a figure the standard's clause on electromagnetic compatibility also uses, since accessories act as antennas (Chapter 21).
MEDTEQ, a maker of test equipment, sets out the standard's rule. The test peak is 120 % of the accessory's rated peak voltage. Up to a rating of 1,600 V the test waveform's crest factor is at most 2; from 1,600 to 4,000 V it is the rating less 400 V, divided by 600 V; above 4,000 V it stays at 6. Three ratings show the rule: the two ends of the sloping range and its midpoint. For an accessory rated at 1,600 V, the test peak is 1.2 times 1,600, or 1,920 V, and at a crest factor of 2 the RMS test voltage is 960 V. For 2,800 V, the crest factor is 2,400 divided by 600, or 4, the test peak is 3,360 V and the RMS voltage 840 V. For 4,000 V, the crest factor is 6, the peak 4,800 V and the RMS voltage 800 V. From 1,600 V to 4,000 V, the test peak rises two and a half times while the RMS voltage falls by a sixth. A higher rating is therefore tested with a taller, more sharply peaked waveform, the shape of the coagulation waveforms of Chapter 8, at a slightly lower RMS voltage. Above 4,000 V the crest factor no longer rises, so the RMS voltage climbs again with the rating: WhaleTeq, another maker of test equipment, states in a 2017 application note that a rating of 6,000 V requires a test at 1,200 V RMS. The standard's rule gives the same figure: the test peak is 1.2 times 6,000, or 7,200 V, and at a crest factor of 6 the RMS voltage is 1,200 V.
Tissue and animal testing
Bench tests show that a generator is electrically safe and delivers what it declares. They do not show what that output does to tissue, and FDA asks for that evidence separately, in its guidance on 510(k) submissions for electrosurgical devices in general surgery, first issued on 15 August 2016 and current in its version of 9 March 2020. For a general soft-tissue claim, it expects thermal effects to be measured in at least three tissue types, such as liver, kidney and muscle, with the length, width and depth of the thermal damage zone measured under magnification on stained sections. The tests cover best-case, realistic and worst-case temperature–time histories, and each test runs at least in triplicate at the minimum, default and maximum power settings.
The guidance adds that an animal model may be necessary, with a chronic study for claims about specific sites such as lung or neural tissue. It asks for capacitive-coupling tests of monopolar instruments against a conductive cannula (Chapter 18), for biocompatibility, sterility and reprocessing data, and for labeling that warns about surgical smoke and hybrid trocars. Vessel sealers add the burst-pressure and animal requirements of FDA's sealer guidance (Chapter 9). IEC 60601-2-2's bench tests and FDA's tissue and animal tests therefore answer different questions, and a report of one gives no evidence of the other.
Reading one public record
Much of this evidence stays in the maker's submission, but FDA publishes the maker's 510(k) summary, and one summary read closely shows what an outsider can learn. FDA cleared Erbe's VIO 3n electrosurgical generator with accessories on 29 August 2025, under product code GEI in 21 CFR 878.4400, class II (Chapter 26). Its summary names two predicates: primarily Erbe's VIO 3 generator, cleared in 2024 as part of the company's HybridTherm System, and secondarily the older VIO 300 D. It describes monopolar and bipolar cutting and coagulation modes, argon plasma modes, a maximum output of 360 W at 350 kHz, and five named configurations.
The summary states that the differences from the predicates "include the device software, external interfaces, available modes and maximum power output". Its comparison table fills them in: a new software version, a service interface on a special USB slot, five configurations against one, and 360 W against 400 W for both predicates. It then lists the testing: functional testing and design controls, side-by-side tissue testing according to FDA's electrosurgical guidance, electromagnetic compatibility to IEC 60601-1-2, electrical safety to IEC 60601-1 and IEC 60601-2-2 "as applicable", software verification to IEC 62304 at the Enhanced documentation level of FDA's 2023 software guidance, and a cybersecurity assessment. Of that testing, the summary states only that it shows the different technological characteristics raise no new questions of safety and effectiveness. It cites non-clinical testing only and makes no statement about clinical data.
Set against the earlier chapters, the summary's two lists line up: the new software with the software verification, the service interface with the cybersecurity assessment, the configurations and modes with the functional and tissue tests, and the lower maximum power with the functional and electrical-safety tests. That pairing is this guide's reading; the summary itself pairs no test with any difference. It also gives no edition of any standard, no output curve and no tissue measurement. Those stay in the submission, which FDA reviewed and the public cannot see.
A second record shows how the framework has moved. Ethicon's Megadyne electrosurgical generator was cleared on 21 January 2022 with a lower maximum power than its predicate: 80 W in every mode except one, which delivers 150 W, against 300 W in the predicate's monopolar cut. Its summary cites IEC 60601-2-18 for capacitive coupling and reports thermal testing in triplicate at minimum, default and maximum settings. It documents the software for a "Major" level of concern, the framework that FDA's 2023 software guidance replaced, and states that the submission "does not include data from Clinical Studies".
Read a 510(k) summary as a map of the evidence. Take from it the device's classification, its predicates, the differences it lists and the families of tests it names. Where it names no edition of a standard, as the VIO 3n's summary names none, ask the maker for its declarations of conformity, and check them against FDA's recognition database as it stood on the submission's date. Never compare two devices' performance from their summaries: the numbers that would allow it stay in the submissions.
- Compliance rests on a stack of process standards, the general standard IEC 60601-1, collateral standards, a particular standard for each energy and product standards for output, and a particular standard overrides the general one where they differ.
- FDA accepts specific editions: as of October 2026 it recognizes edition 6.1 of IEC 60601-2-2, entered on 25 May 2026, and accepts declarations to edition 6.0 until 1 July 2029.
- IEC 60601-2-2 tests output into resistors across the load range, high-frequency leakage through 200 Ω, return-electrode heating against a 6 °C limit and accessory insulation at 120 % of rated voltage, and FDA's guidance asks separately for tissue and animal evidence.
- The VIO 3n's public summary lists four differences from its predicates and the families of tests run, but no editions, curves or tissue data, and no statement on clinical data.
Classification decides the route.
+ The questionWhat must a surgical energy device show to reach US patients, and what happens when it changes?
A regulation, a product code and a class
On 26 and 27 October 2022, an FDA advisory panel met to consider how ultrasonic surgical instruments should be classified, although the device type had been sold in the United States since before the 1976 amendments that created the country's classification system. FDA's briefing for the panel recommended class II with special controls, requirements written for that device type, such as particular tests and labeling, in addition to the rules for every device. As of October 2026, no proposed rule has been found, and FDA's product-code page still lists the instruments as unclassified, a pre-amendment device type that reaches the market by 510(k). Ultrasonic instruments are therefore an exception to the order in which the US system works: a device type is normally placed first in a regulation and a class, and the class decides what its maker must show.
FDA describes most device types in a classification regulation in Title 21 of the Code of Federal Regulations, which assigns each a class, and FDA files the records for each type under one or more three-letter product codes (Chapter 23). Most class II devices reach the market by a 510(k), a premarket notification showing that the device is substantially equivalent to a legally marketed predicate device, and FDA's decision on it is a clearance (Chapter 9). Devices in class III, the class of highest risk, need premarket approval, or PMA, and FDA's decision is an approval.
Electrosurgery sits in 21 CFR 878.4400, which describes a device "intended to remove tissue" and "control bleeding by use of high-frequency electrical current". The regulation places it in class II, FDA files these devices under product code GEI, and GEI devices are eligible for FDA's program in which accredited third parties review 510(k)s, while ultrasonic instruments are excluded from it. The table gives the codes that FDA's records show for devices in this book, as of October 2026. Where it shows a dash, the records cited give no regulation.
| Device type | Product code | Regulation, 21 CFR | Class | Route |
|---|---|---|---|---|
| Electrosurgical cutting and coagulation, with accessories | GEI | 878.4400 | II | 510(k) |
| Microwave ablation | NEY | 878.4400 | II | 510(k) |
| Powered laser surgical instrument | GEX | 878.4810 | II | 510(k) |
| Ultrasonic surgical instrument | LFL | — | unclassified | 510(k) |
| HIFU for prostate tissue ablation | PLP | 876.4340 | II | De Novo (2015), then 510(k) |
| Focused ultrasound for non-thermal, mechanical tissue ablation (histotripsy) | QGM | 878.4405 | II | De Novo (2023) |
| Cryoablation for low-risk breast cancer | QXW | 878.4355 | II | De Novo (2025) |
| MR-guided HIFU ablation | NRZ | — | III | PMA |
| Cardiac ablation catheter, pulsed field | QZI | — | III | PMA |
Hybrid instruments that deliver bipolar current and ultrasonic energy through one jaw have been cleared under the electrosurgical regulation with both product codes (Chapter 11). Other cryosurgical devices reach the market by 510(k) under product code GEH, and AngioDynamics states that its NanoKnife IRE system is cleared by 510(k) for the surgical ablation of soft tissue (Chapter 15).
When there is no predicate
The first device of a new kind has no predicate, and so no 510(k). If its risk is low or moderate, its maker can request a De Novo classification, and when FDA grants the request it creates a new regulation and product code that later devices can follow. The ablation chapters met three such grants: Sonablate's HIFU system for prostate tissue on 9 October 2015, HistoSonics' Edison system for liver tumors on 6 October 2023 (Chapter 16), and IceCure's ProSense cryoablation system for low-risk breast cancer on 3 October 2025 (Chapter 14). The route then works as designed. Chapter 16 tells how a competitor withdrew its own De Novo request after Sonablate's grant and was cleared by 510(k) on 6 November 2015, four weeks later, with Sonablate as its predicate.
An announced request is not yet a grant. As of 20 October 2026, no FDA decision had been found on the De Novo request for kidney tumors that HistoSonics announced on 11 May 2026 (Chapter 16). Devices of the highest risk go to premarket approval instead of a De Novo. FDA approved the cardiac pulsed field ablation systems of Chapter 15 by PMA, the first on 13 December 2023, and its first approval of a focused-ultrasound device was a PMA of 30 June 1988, for a glaucoma treatment (Chapter 16).
A 510(k) device is cleared, a De Novo request is granted and a PMA device is approved. A request that has been submitted or announced is an application, and only FDA's decision makes a device legally marketable for the use it names. That use is the indication on the label: an indication that names a tissue, as NanoKnife's names prostate tissue, makes no claim about a disease. Companies' own releases sometimes use the wrong verb. Medtronic's release of 30 September 2025 on launching its FT10 generators in India called them "USFDA-approved", although FDA clears such generators by 510(k), and HistoSonics' release of 22 June 2026 called the liver authorization a "De Novo clearance". In the EU, a notified body issues certificates and the manufacturer affixes the CE marking (Chapter 27). Voluntary ISO 13485 certificates come from certification bodies, and ISO itself issues none.
The quality system, software and cybersecurity
Whatever the route, the maker must run a quality system, and on 2 February 2026 the US rule for it changed its basis. The Quality Management System Regulation took effect that day. A final rule published on 2 February 2024 created it by amending 21 CFR Part 820. The new regulation incorporates ISO 13485:2016 by reference and replaces the older Quality System Regulation. Chapter 22 described design controls as the process that turns an intended use and its claims into requirements, risk controls, verification and validation, a trace from claim to test report, and a loop from the field back into the design. Under the new regulation, a US maker runs and records that process to ISO 13485's requirements for design and development. The VIO 3n's summary, prepared on 7 April 2025, shows the older form: it cites its design controls under 21 CFR 820.30, a section of the regulation then in force.
Chapter 20 placed software and connectivity in a generator's architecture, and in a US submission they carry obligations of their own. An opening notice in FDA's electrosurgical guidance says that general-surgery electrosurgical devices should generally address the Enhanced documentation level of FDA's 2023 software guidance (Chapter 20). The body of the electrosurgical guidance has not been revised since 2020 and still speaks of a "major" level of concern, a term from the framework that the 2023 guidance replaced (Chapter 25). The electrosurgical guidance therefore uses the terms of both frameworks, the newer in its notice and the older in its body.
Connectivity adds obligations that come from the statute itself. For premarket submissions made on or after 29 March 2023, section 524B of the Federal Food, Drug, and Cosmetic Act sets requirements for the sponsor of a cyber device. The section defines a cyber device by three criteria, which FDA's frequently asked questions restate: it includes software validated, installed or authorized by the sponsor, it can connect to the internet, and it has characteristics that could be vulnerable to cybersecurity threats. Among the requirements, the sponsor must provide a plan to monitor, identify and address vulnerabilities and exploits after sale, processes that give reasonable assurance that the device and related systems are cybersecure, including updates and patches, and a software bill of materials. FDA reissued its premarket cybersecurity guidance on 3 February 2026, superseding the edition of 27 June 2025, with "Quality Management System Considerations" now in its title.
When the device changes
A device is cleared as it was described, and every later change, to its software, components, labeling or materials, has to be assessed against that description. FDA's guidance on deciding when to submit a 510(k) for a change to an existing device, issued on 25 October 2017, walks through the assessment with flowcharts for labeling changes, for changes in technology, engineering or performance, and for changes in materials. Each flowchart serves the question that 21 CFR 807.81(a)(3) asks: whether the change "could significantly affect the safety or effectiveness of the device", or is a "major change or modification in the intended use of the device". If either holds, a new 510(k) is needed. If neither does, the guidance asks the maker to "document your analysis and file it for future reference". A companion guidance of the same month covers software changes.
Chapter 22 placed the change assessment inside the design loop, and these guidances set the US threshold for it. Both records of Chapter 25 are new models cleared against predicates: the VIO 3n differs from its predicates in software, interfaces, modes and maximum power, and the Megadyne generator in a lower maximum power among other changes. Neither summary says which difference made a new 510(k) necessary, so a reader learns where the threshold lies from the guidance, and the records show only that each maker submitted a new one.
After sale: reports, corrections and recalls
Clearance marks a date, and the record continues after it. FDA collects medical device reports of deaths, serious injuries and malfunctions and counts them by product code, and the counts can weigh in a classification decision: FDA's briefing for the 2022 panel on ultrasonic instruments counted the reports filed under LFL.
FDA's executive summary for the panel counted 46,673 medical device reports under product code LFL from 1 January 2002 to 31 December 2021: 44,354 malfunctions, 2,263 serious injuries and 56 deaths. The three add up to the total, since 44,354 and 2,263 make 46,617, and 56 more make 46,673. Malfunctions are 44,354 divided by 46,673, or 95 % of the reports. Serious injuries are 2,263 divided by 46,673, about 4.8 %, and deaths 56 divided by 46,673, about 0.12 %, or 1 report in about 830. Over the twenty years the reports average 46,673 divided by 20, about 2,300 a year. The summary gives no count of instruments in use or of procedures performed, so none of these figures is a rate of harm per operation. They describe what was reported, and 95 % of it was malfunctions.
When a marketed device puts patients at risk, its maker may correct it in place or remove it, and FDA classes the action on a scale of its own, separate from the device classes. FDA calls a Class I recall "the most serious type of recall", and such a recall need not remove anything. The Varipulse pulsed field ablation catheter shows such a recall. In February 2025 Biosense Webster, a Johnson & Johnson company, sent customers a correction updating the instructions for the catheter, after 4 of 132 patients in its US External Evaluation had a stroke or transient ischemic attack. That is about 3 %, against an expected rate of 1 % or less. FDA classed the correction as the most serious type although the correction left the catheter on the market, and its notice asked users to follow the updated instructions, including their indication for "symptomatic drug refractory paroxysmal atrial fibrillation" (Chapter 15 tells the pause that preceded it). Because the catheter is a PMA device, its later change of irrigation flow rate went back to FDA for approval, which FDA gave on 26 June 2025 (Chapter 22).
The return pad of Chapter 8 has its own entry. Megadyne began a correction of its MEGA 2000 and MEGA SOFT reusable return electrodes on 1 June 2023 after reports of burns, and FDA posted it as a Class I recall on 31 July 2024, covering 21,200 devices in the US, with 63 injuries reported and no deaths. Less serious actions fall in Class II. Olympus began one on 25 July 2025 for 155 units of its ESG-410 generator, which could show a system error and then restart, sometimes repeatedly, and FDA posted it on 19 August 2025 (Chapter 24). Each such action returns to the loop of Chapter 22, where the correction is itself a change to be assessed.
Write the intended use and the list of claims first, and only then classify the device: find the product code whose regulation describes it, its class and the route that class requires. Read the product code's page for third-party eligibility and the standards FDA lists against it. For each foreseeable change, record in advance the answer the 2017 flowcharts give, so that the change plan names the changes that need a new 510(k). Set up the procedures for reports and corrections before the first sale.
- In the US, a device type's regulation, product code and class set its route: electrosurgical generators and surgical lasers are class II by 510(k), ultrasonic instruments remain unclassified as of October 2026, and cardiac pulsed field ablation systems are class III by PMA.
- A new kind of device with no predicate can be classified by a De Novo grant, as Sonablate, Edison and ProSense were, and a later device can then follow by 510(k), as one did four weeks after Sonablate.
- Since 2 February 2026 the US quality system regulation incorporates ISO 13485:2016, under which a maker runs the design controls of Chapter 22, and section 524B adds cybersecurity obligations to submissions for cyber devices made on or after 29 March 2023.
- A change that could significantly affect safety or effectiveness needs a new 510(k), and after sale FDA counts medical device reports and classes corrections and recalls, as it classed the 2025 Varipulse correction among the most serious.
Same evidence, other assessors.
+ The questionHow do the European Union and India classify and assess the same devices, and what does each add?
Rule 9 in its own words
Sixteen words in the European Union's Medical Device Regulation decide whether an energy device is class IIa or IIb. Regulation (EU) 2017/745 classifies devices by the rules of its Annex VIII, and Rule 9 covers active therapeutic devices intended to administer or exchange energy. Such devices are class IIa unless they "may administer energy to or exchange energy with the human body in a potentially hazardous way", in which case they are class IIb. The rule names what decides the hazard: "the nature, the density and site of application of the energy". Those are the variables of this book. Chapter 6 showed that the nature of the carrier sets where its energy is deposited, Chapter 3 that current heats where its density is highest, and Chapters 2 and 13 that the site changes what the same dose does.
The Medical Device Coordination Group's guidance on classification, MDCG 2021-24, was revised as rev. 1, published by the Commission on 20 April 2026. The revision restructured the guidance's section on Rule 9 "to better match the rule" and kept the examples. As class IIb under Rule 9 it names "high-frequency electrosurgical generators, and electrocautery equipment, including their electrodes", "surgical lasers", "lithotriptors, surgical ultrasound devices" and "high-intensity focused ultrasound (HIFU)". By contrast, its class IIa examples include muscle stimulators and phototherapy. It places the cryo-ablation of the heart or spine under another rule, Rule 6, in class III, and it does not mention pulsed field ablation under Rule 9. The guidance does not bind; the rule in Annex VIII does.
A notified body, a certificate and a register
A class IIb device needs a notified body, an organization designated under the regulation to assess conformity. Under the usual route, in Annex IX, the notified body audits the manufacturer's whole quality management system and assesses the technical documentation of at least one representative device in each generic device group; a route by type examination is also available. The notified body then issues its certificates, and the manufacturer affixes the CE marking.
What happens when a certified device changes depends on what the notified body certified. For a device whose technical documentation a notified body has certified, such as a class III device, Annex IX requires the notified body's approval for any change that "could affect the safety and performance of the device or the conditions prescribed for use of the device". A generator's documentation is assessed through a representative device of its group rather than certified device by device. The maker of a generator therefore informs its notified body of any planned substantial change to its quality management system or to the devices it covers, and any approval is issued as a supplement to the quality management system certificate. The EU's test for a certified device is the one Chapter 26 met in the US, asked without the word "significantly".
The EU's register for devices, EUDAMED, became partly mandatory on 28 May 2026. Four of its modules apply from that day: actor registration, device registration with unique device identifiers, notified bodies and certificates, and market surveillance. Ireland's Health Products Regulatory Authority sets out the transition: devices placed on the market before that date are to be registered by 28 November 2026, and certificates issued before mandatory use by 28 May 2027. Vigilance, the reporting of serious incidents after sale, was not among the four modules.
A transition and a proposal
Devices certified under the older directives are still moving across to the regulation. Regulation (EU) 2023/607, adopted on 15 March 2023, came into force on its publication on 20 March 2023 and extended the validity of certificates issued under the directives. For class III devices and implantable class IIb devices, with listed exceptions, the extension runs to 31 December 2027. For other class IIb devices, and for class IIa and sterile or measuring class I devices, it runs to 31 December 2028. It applies only where, among other conditions, the manufacturer had put a quality management system in place and lodged a formal application with a notified body by 26 May 2024, and signed a written agreement with it by 26 September 2024.
An electrosurgical generator is a non-implantable class IIb device under Rule 9, so the amended transition puts it in the later group, with a deadline of 31 December 2028 rather than 31 December 2027. Its directive certificate stays valid that long only if its maker met the conditions of the transition, among them an application by 26 May 2024 and an agreement by 26 September 2024. From the later of those dates to the deadline is four years, three months and five days, or about 51 months. From the end of October 2026, the month this book is current to, the deadline is two years and two months away, or 26 months. A generator still sold on a directive certificate in October 2026 therefore has a little over two years left on it, unless a later act moves the date.
On 16 December 2025 the Commission proposed a targeted revision of the regulation, to simplify it and reduce its burden. Its explanatory memorandum states that the proposal would remove the five-year maximum validity of certificates in favor of risk-proportionate periodic reviews, let notified bodies replace on-site audits with remote ones, and set surveillance audits every two years where there are no safety issues. The proposal would also adapt some classification rules to lower classes, naming reusable surgical instruments among others, and the memorandum does not name Rule 9. As of October 2026 the proposal has not been adopted. The Council's progress report of 5 June 2026 records that the member states had not reached a general approach, the Council's agreed position for negotiating the text with the Parliament. The report states that the European Parliament aims to vote on its position in early 2027, an aim rather than a scheduled date.
India: the Medical Devices Rules
India regulates devices under the Medical Devices Rules, 2017, which set four risk classes, A to D. Its regulator, the Central Drugs Standard Control Organisation, or CDSCO, publishes classification lists by specialty. Its 2021 lists for operation theatre devices and for dermatology and plastic surgery place general-purpose electrosurgical units and laser surgical instruments in class C, the class for moderate to high risk.
CDSCO began a new cycle of lists with a draft for comment in January 2025, and issued its final oncology list on 2 January 2026. The list has 77 entries, and the energy devices it names are all in class C: cryoablation systems, electroporation therapy systems, microwave ablation systems, a microwave or electrosurgical system generator, HIFU ablation systems and ultrasonic ablation systems. Not every device in this book could be placed in the new lists: no CDSCO class for vessel sealers, ultrasonic shears or cardiac pulsed field catheters was found in any list issued since late 2025.
The class sets who licenses the device. CDSCO's FAQ on the Rules states that the Central Licensing Authority licenses class C and D devices, while lower classes are licensed by the states or only registered, so a generator or a surgical laser is licensed centrally. An importer applies on Form MD-14 and receives its import license on Form MD-15. The FAQ gives timelines: an inspection of a class C or D manufacturing site within 60 days of the application, an import license within nine months, and approval of a major change within 45 days for a manufacturer or 60 days for an importer. A license stays valid until it is suspended or cancelled, against a retention fee every five years.
A device must meet the standards of the Bureau of Indian Standards where they exist, then ISO or IEC standards, then pharmacopoeial or validated manufacturer's standards, and after sale the license holder must submit post-marketing surveillance data. For a generator, the IEC tests of Chapter 25 therefore apply wherever no Indian standard covers the same ground.
One evidence core, filed three ways
The three systems ask in different words and through different assessors, but they draw on one body of evidence: a quality management system to ISO 13485, a risk management file to ISO 14971, IEC 60601 test reports, software documentation to IEC 62304, usability files, and the tissue and animal testing of Chapter 25. An electrosurgical generator shows how the forms differ. In the US the generator is class II under product code GEI, FDA clears it by 510(k), and its summary becomes public, as the VIO 3n's did. In the EU it is class IIb under Rule 9, a notified body audits its maker and assesses its technical documentation, and the maker affixes the CE marking. In India it is class C, licensed by the Central Licensing Authority against Indian standards first. A maker that builds the core once and keeps it current can present it in the form each assessor asks for, while one that builds it separately for each market repeats most of the work.
Classify the device in each target market before planning the evidence: the US product code and route, the EU rule and class, and the CDSCO list and class, noting where a list gives no class. Build the quality system, the risk file and the test reports once, to editions that each regulator accepts on the planned dates of submission. Keep one register of changes, and record for each change the US, EU and Indian answer to whether it needs a new submission or approval.
From Bovie's knobs to a public record
In the autumn of 1926, at the Peter Bent Brigham Hospital, William Bovie adjusted the knobs of his apparatus while Harvey Cushing operated. No standard governed the apparatus and no regulator had seen it. Its current went where the circuit allowed. By the account in Li and Isaacson's 2024 history of the collaboration, a shock once passed through a metal retractor to Cushing's arm, and the first unit lacked an effective return electrode. Each of those hazards now has a test. The current that escapes a generator by any path other than its return lead is measured through a 200 Ω resistor, and in the standard's test a return pad must keep the skin beneath it within 6 °C of its starting temperature (Chapter 25). A new generator may be sold in the United States only after FDA clears it, on evidence that includes test reports against IEC 60601-2-2 and the tissue tests of FDA's guidance. Whereas no regulator had seen Bovie's apparatus, the summary for Erbe's VIO 3n generator, cleared on 29 August 2025, is open to anyone.
The operating room of 1926 and the public record of 2025 together hold the answer this book has built. Where energy goes when it enters tissue is set by physics: current heats where it is densest, light where its wavelength is absorbed, focused sound at its focus, and the heat then spreads about 0.75 mm in one second and about 2.4 mm in ten (Chapters 2, 3, 5 and 16). Surgery needs eleven kinds of energy because its tasks need different geometries and times: a cut along a line, a seal across a vessel's wall, a zone of dead tissue around a needle, a lesion deep inside a closed body, and a kill that spares the vessels and ducts running through it. Keeping each energy where the surgeon intends takes three things more: a design that confines it, such as an isolated output, a monitored pad and a safety channel; tests that show the confinement holds; and an assessment by someone other than the maker, whether FDA, a notified body or CDSCO.
The physics of 1926 still holds. Bovie's patent, filed in 1928, separated two currents: a steady one that cuts, and a heavily damped one at about 4,000 V that dries tissue and stops bleeding. Nearly a century later, Medtronic declares the same distinction for each mode of its FT10 generator as a duty cycle and a crest factor (Chapter 8). The years since have added an account of where that current goes: the path by which it returns, the limits it must stay within, and a record of what was tested before it reached a patient.
- Under the EU's Rule 9, an energy device is class IIb when it may administer or exchange energy in a potentially hazardous way, judged by the nature, density and site of the energy, and the MDCG guidance of April 2026 keeps electrosurgical generators, surgical lasers, surgical ultrasound and HIFU as its examples.
- A class IIb device needs a notified body's assessment, directive certificates for non-implantable class IIb devices can run to 31 December 2028 under 2023/607, and the Commission's December 2025 proposal has not been adopted as of October 2026.
- Under its 2021 lists and its oncology list of 2 January 2026, India places electrosurgical units, surgical lasers and ablation systems in class C, licenses them centrally and applies Indian standards first.
- One core of quality, risk and test evidence serves all three jurisdictions; physics decides where energy goes, the tasks of surgery explain the eleven kinds, and design, test and assessment keep each where the surgeon intends.
Lessons.
The chapters reduce to a short set of working rules for anyone who designs, tests, buys, regulates or writes about surgical energy devices. Each one traces back to the chapters that explain it.
- State temperature and time together, or state the effect. A temperature alone says little; the injury depends on how long it is held, and the tissue decides how steeply (Chapters 1 and 2).
- Ask where the energy is deposited and how fast. Geometry and speed, set by the carrier and the control, decide cut, seal, coagulation or ablation more than the final temperature does (Chapters 2, 5 and 6).
- Follow the current density. Heat appears wherever current is concentrated, intended or not: at the tip, at a lifting pad edge, at a pinhole in insulation (Chapters 3, 8 and 18).
- Read the waveform, not the power setting. Crest factor and peak voltage decide whether a mode cuts, coagulates or fulgurates, and how strongly it stresses insulation and couples stray current (Chapters 8 and 18).
- Read power-versus-load curves before power ratings. A setting holds only at the rated load and across the constant-power band; outside it the generator delivers less, sometimes much less (Chapter 10).
- Treat every end point as an inference. Sealers and ablation generators infer the tissue's state from impedance or temperature; the strength of the result is established by testing, not by the tone (Chapters 9, 13 and 17).
- Stay within the indicated vessel size. In an independent bench comparison that Okhunov and colleagues published in 2018, four of five devices lost about half their seal strength or more on arteries larger than 7 mm. For the largest vessels, a mechanical closure is the alternative, because it does not rely on heating the tissue (Chapters 9 and 17).
- Remember that no current does not mean no hazard. Ultrasonic blades retain heat after activation; a laser's eye-hazard distance can exceed the room; each energy trades one hazard for another (Chapters 6, 11 and 19).
- Specify and read an ablation zone by its lethal edge, not its visible one. The ice ball's edge is 0 °C, heated zones shrink next to vessels, and manufacturer zone charts are measured without perfusion (Chapters 2, 13 and 14).
- Choose non-thermal energy when the scaffold must survive. Animal studies of pulsed fields and a 2026 review of preclinical histotripsy studies by Yu and colleagues report that these energies kill cells while largely sparing collagen-rich vessels and ducts. But pulsed fields still deposit some heat near their electrodes, and each of these energies needs its own verification (Chapters 15 and 16).
- Design for the fault as well as the function. An independent safety channel, isolation barriers, pad monitoring and usability engineering keep a single failure from delivering unintended energy (Chapter 20).
- Give every interface an owner, and trace every claim to its test. Each interface, where two parts meet or a part meets the people who use, clean and service it, needs a specification, a named party that writes it and a test that verifies it. Each claim needs requirements a test can fail, traced to their tests and to the label line, so that a change shows which tests to repeat (Chapters 21 and 22).
- Build one evidence core and present it three ways. A quality system, a risk file, standards test reports, and tissue and animal testing make up one body of evidence, which the US, the EU and India each assess in their own way (Chapters 25, 26 and 27).
- Keep filed, cleared, granted, approved and certified apart. Every regulatory claim names the status, the regulator, the indication and the date (Chapters 16, 23, 24 and 26).
Glossary.
Terms are defined as they are used in this guide.
- 510(k)
- A US premarket notification showing that a device is substantially equivalent to a legally marketed predicate device; FDA's decision on it is a clearance (Chapters 9 and 26).
- Ablation
- Destroying tissue in place without removing it, usually a tumor, or lines of heart muscle whose scar treats an irregular heart rhythm; it reaches its target through a needle, a catheter or the intact skin (Before we start).
- Absorption coefficient (µa)
- The probability per centimeter that light is absorbed; at a depth of 1/µa, the absorption depth, the intensity has fallen to about 37 % of its value at the surface (Chapter 5).
- Acoustic impedance
- A material's density multiplied by its speed of sound, measured in rayls; sound crosses a boundary well only when the impedances on its two sides match, and a mismatch reflects it (Chapter 4).
- Acoustic window
- A path to a target that is free of gas and bone, which focused ultrasound needs (Chapter 4).
- Active electrode
- The electrode through which monopolar current enters the patient, such as the pencil, hook or scissors in the surgeon's hand (Chapter 8).
- Active electrode monitoring
- A conductive shield around an instrument's insulated conductor, connected back to the generator's return and covered by a second insulation, which collects current that escapes through a defect or by capacitance and returns it; a monitor tells fault current, in phase with the voltage, from normal capacitive current, a quarter of a cycle apart, and stops the generator on a fault (Chapter 18).
- Advanced bipolar sealer
- A bipolar instrument that seals vessels by combining high, controlled jaw pressure, high current at low voltage and a closed loop that ends energy delivery by itself; many are indicated for vessels up to 7 mm (Chapter 9).
- Antinode
- A point of maximum motion in a standing wave; the blade tip of an ultrasonic instrument sits at one (Chapter 11).
- Applied part (B, BF, CF)
- A part of a device that contacts the patient, classified as type B, BF or CF from the least to the most isolated; type CF is required for anything applied directly to the heart (IEC 60601-1; Chapter 20).
- Arrhenius model
- A description of thermal damage as a chemical reaction whose rate rises exponentially with temperature, the damage accumulating for as long as the temperature is held (Chapter 1).
- Basic safety
- Freedom from unacceptable risk caused directly by physical hazards, such as electric shock, fire, mechanical injury and excessive temperature, when equipment is used in normal condition or with a single fault (IEC 60601-1; Chapter 20).
- Bioheat equation
- The heat balance for living tissue that Harry Pennes published in 1948: the temperature of a small volume rises by the heat deposited in it, minus the heat conducted to its neighbors, minus the heat carried off by blood; metabolic heat is also in it but is negligible during surgery (Chapter 2).
- Bipolar
- Electrosurgery with an instrument that carries both electrodes in its tips, usually the two blades of a forceps, so that current flows only through the tissue held between them and no return pad is needed (Chapter 9).
- Blocking capacitor
- A capacitor in series with each lead of a generator output that passes high-frequency current and blocks direct and low-frequency current, which could otherwise stimulate nerves and muscles (Chapter 8).
- Burst pressure
- The pressure at which a seal fails when the sealed vessel is filled with liquid and the pressure raised; the bench test of a seal's strength (Chapter 9).
- Butterworth–Van Dyke model
- The equivalent circuit of a piezoelectric transducer: the ceramic's own capacitance in parallel with a resonant branch whose inductance, capacitance and resistance stand for the moving mass, the stiffness and the mechanical losses, which include the tissue pressed against the blade (Chapter 11).
- Capacitive coupling
- Transfer of alternating current across an insulator without contact, for example from an instrument shaft to a metal cannula (Chapter 7).
- Carbonization
- Charring of tissue once it has dried out, with an onset put at about 150 to 200 °C, well above the 100 °C at which wet tissue stalls; char wastes energy, releases smoke and conducts current poorly (Chapter 1).
- Cavitation
- The growth and motion of gas bubbles in a sound field: in stable cavitation a bubble oscillates gently in step with the wave, and in inertial cavitation it grows to many times its size and collapses violently, and the collapse damages tissue (Chapter 4).
- CDSCO
- The Central Drugs Standard Control Organisation, India's device regulator under the Medical Devices Rules, 2017, which publishes classification lists by specialty and, through its Central Licensing Authority, licenses class C and D devices (Chapter 27).
- CE marking
- The mark a manufacturer affixes to a device for the EU market once its conformity assessment is complete; for a class IIb device, after a notified body has issued its certificates (Chapters 16 and 27).
- CEM43
- Cumulative equivalent minutes at 43 °C, the thermal dose used in hyperthermia; on this measure each degree above 43 °C halves the time needed for the same effect (Chapter 1).
- Chromophore
- A molecule that absorbs light and turns its energy into heat; in surgery, mainly water, hemoglobin and melanin (Chapter 5).
- Cleared
- Of a device in the United States: allowed onto the market by the Food and Drug Administration through a 510(k), as substantially equivalent to a device already legally marketed (Before we start; Chapter 26).
- Coagulation
- Irreversible unfolding and setting of proteins by heat, which kills cells and stops small vessels bleeding; published onsets range from about 50 to 60 °C, depending on the tissue, species and exposure time. To a surgeon the word means this change caused by heat, and the clotting by which blood closes a small wound on its own is a different process (Chapter 1).
- Collagen
- The protein that gives vessel walls, skin and tendons their strength, three chains wound around each other like the strands of a rope; in measurements on rat tissues, heat unwinds its helix between about 60 and 90 °C, depending on the tissue and the test, the fiber shrinks as it unwinds, and unwound layers pressed together as they cool fuse, which is how a vessel seal works (Chapter 1).
- Collateral standard
- A standard in the IEC 60601-1 family that adds requirements many kinds of equipment share, such as IEC 60601-1-2 for electromagnetic compatibility, IEC 60601-1-6 for usability and IEC 60601-1-8 for alarm systems (Chapters 20 and 25).
- Contact quality monitoring
- Monitoring of a split return pad's contact by the impedance between its two halves, which stops the output when the pad lifts; since edition 6.0 of 2017, IEC 60601-2-2 has required adult return electrodes of this kind (Chapters 8 and 25).
- Crest factor
- Peak voltage divided by RMS voltage: 1.41 for a continuous sine wave and higher the lower the duty cycle, so it is low for cutting waveforms and high for fulguration and spray (Chapter 8).
- Critical task
- In FDA's human factors guidance, a user task that, if performed incorrectly or not performed at all, would or could cause serious harm (Chapter 22).
- Cryoablation
- Ablation by freezing: a probe colder than tissue draws heat out, and ice grows outward from it (Chapters 6 and 14).
- Cutting
- Dividing tissue with current or a laser by delivering energy so quickly into a thin line that the water in its cells boils explosively and the tissue parts; no sharp edge is involved (Before we start; Chapter 1).
- Cyber device
- Under section 524B of the US Federal Food, Drug, and Cosmetic Act, a device that includes software validated, installed or authorized by its sponsor, can connect to the internet and has characteristics that could be vulnerable to cybersecurity threats; submissions for one made on or after 29 March 2023 must include a vulnerability plan, update processes and a software bill of materials (Chapters 20 and 26).
- Damping
- The decay of an oscillation. In Bovie's apparatus, as his patent filed in 1928 describes it, the less damped and steadier current severed tissue, and the heavily damped one, whose bursts of oscillation died away quickly, dried tissue and stopped bleeding (the opening story).
- De Novo
- The US route for a new kind of device of low or moderate risk that has no predicate: when FDA grants the request it creates a new regulation and product code that later devices can follow by 510(k) (Chapters 14 and 26).
- Desiccation
- Slow drying of tissue: when energy arrives slowly, the water leaves gradually and the tissue dries and shrinks into a firm coagulum, which is how most contact coagulation works (Chapter 1).
- Design and development file
- The record that holds the trace from each claim to its requirements, tests and label line, which ISO 13485 requires for each device type or family (Chapter 22).
- Design controls
- The process that turns an intended use and its claims into design inputs, design outputs, verification and validation; FDA's guidance of 1997 draws it as a waterfall, and ISO 13485 sets out the same steps (Chapter 22).
- Design input
- A requirement the design must meet, written so that a test can fail it, such as seals of vessels up to 7 mm, measured rounded and perfused, holding at least 360 mmHg in each mode (Chapter 22).
- Design output
- What the design process produces to define the device, which must meet the input requirements; for a sealer, its jaw pressure, sealing mode, end-point algorithm and instructions for use (Chapter 22).
- Dielectric heating
- Heating by an alternating electromagnetic field that turns water molecules, which lag behind it, and drives ions back and forth; both motions dissipate energy as heat throughout a volume, with no conducting path to an electrode, as in microwave ablation (Chapters 6 and 13).
- Diffusion length
- An estimate of how far heat spreads in a given time: the square root of four times the thermal diffusivity times the time, about 0.75 mm in one second and 2.4 mm in ten for soft tissue (Chapter 2).
- Direct coupling
- Current passing from an active electrode into another conductor it touches, such as a metal grasper or the laparoscope, and from there into tissue at that conductor's point of contact (Chapter 18).
- Duty cycle
- The fraction of time an output is on. The waveform duty cycle of Chapter 8 is the share of each cycle in which the carrier is on, 100 % for pure cut and 6.25 % for the FT10's fulguration. The activation duty cycle of Chapter 10 is the share of time a generator's output is switched on, as in a rating of 10 seconds on and 30 off, and it is the sense of IEC 60601-2-2's 45 % rule (Chapters 8 and 10).
- Electromagnetic compatibility (EMC)
- The interface between a device and the other equipment in the room through emissions and immunity; for a surgical generator, set by IEC 60601-1-2 as modified by IEC 60601-2-2, with emissions tested in the idle state (Chapters 19 and 21).
- Electroporation
- The opening of pores in cell membranes by the voltage an external electric field induces across them, reported above about 0.5 to 1 V; reversible when the pores reseal and the cell survives, irreversible when the cell cannot repair them and dies (Chapter 15).
- Essential performance
- The performance a device must maintain for the patient to be safe, as identified by the manufacturer's risk analysis; for a generator, an output that matches its setting within its declared limits, appears only when commanded and stops when commanded (IEC 60601-1; Chapter 20).
- EUDAMED
- The EU's register for devices; four of its modules, for actor registration, the registration of devices and their unique device identifiers, notified bodies and certificates, and market surveillance, became mandatory on 28 May 2026, and vigilance was not among them (Chapter 27).
- Formative evaluation
- In FDA's human factors guidance, a test of a device's user interface while the interface is still changing, before human factors validation testing on the final design (Chapter 22).
- Fulguration
- Coagulation of a tissue surface by bursts of current at kilovolts that jump an air gap and dry and char the surface over a broad area, with no contact; also called spray coagulation (Chapter 8).
- General standard
- IEC 60601-1, the standard for the basic safety and essential performance of all medical electrical equipment, at edition 3.2 of 2020 as of October 2026 (Chapters 20 and 25).
- Harm
- In ISO 14971:2019, injury or damage to the health of people, or damage to property or the environment, such as a skin burn under a lifted pad (Chapter 22).
- Hazard
- In ISO 14971:2019, a potential source of harm, such as the monopolar current that leaves the patient through the return pad (Chapter 22).
- Hazardous situation
- In ISO 14971:2019, circumstances in which people, property or the environment are exposed to a hazard, such as a pad partly lifted during activation (Chapter 22).
- Heat sink
- Cooling of the tissue around a large vessel by the blood flowing through it; the "3 mm rule" of ablation practice treats a vessel 3 mm or more across beside a tumor as a heat sink (Chapter 2).
- Heating factor
- Current squared multiplied by time; under IEC 60601-2-2, a monopolar mode is a high current mode when its intended use, at maximum current and maximum activation duty cycle, gives more than 30 A²s in any 60 seconds (Chapter 10).
- HIFU
- High-intensity focused ultrasound, in which sound absorbed at a small focus deep inside the body heats it during exposures that last seconds (Chapter 16).
- High current mode
- Under IEC 60601-2-2, a monopolar mode whose intended use, at maximum current and maximum activation duty cycle, gives more than 30 A²s in any 60 seconds; such a mode brings stricter requirements for the return electrode (Chapter 10).
- High-frequency leakage current
- High-frequency current that escapes from an electrosurgical output by paths other than the intended return, through stray capacitance to earth; IEC 60601-2-2 limits it to 150 mA for monopolar outputs, measured through a 200 Ω resistor that stands in for a person (Chapter 20).
- Histotripsy
- Mechanical destruction of tissue by focused ultrasound pulses lasting microseconds, at a duty cycle below 1 %, whose inertial cavitation breaks cells apart while the tissue barely warms (Chapter 16).
- Human factors evaluation
- A study of how users work with a device, such as Intuitive's study of how users work with its E-200 generator in its integrated, backup and standalone set-ups (Chapter 21).
- Human factors validation testing
- FDA's term for the test at the end of development, on the final user interface, with in general at least 15 participants for each distinct user population; the Johner Institute notes that IEC 62366-1 calls it summative evaluation (Chapter 22).
- Hybrid cannula
- A metal cannula held in the abdominal wall by a plastic anchor that insulates it from the wall, so that current coupled into the metal can discharge into whatever touches it, which may be bowel (Chapter 18).
- Hybrid instrument
- An instrument that combines two energies in one jaw, such as bipolar current and ultrasonic vibration (Chapter 7).
- Impedance
- The opposition a circuit offers to an alternating current, measured in ohms like a resistance (Before we start).
- Impedance roll-off
- The point in RF ablation at which tissue at the needle dries and chars, the impedance rises sharply and the generator can no longer deliver power, so the zone stops growing (Chapter 13).
- Indication
- The use a device's maker claims for it, which a clearance or approval covers as the label states it, such as sealing vessels up to and including 7 mm in diameter; a general soft-tissue indication covers soft tissue in general rather than one procedure or organ (Chapters 2 and 15).
- Information for safety
- The third option for controlling a risk under ISO 14971, used where design and protective measures reduce the risk too little, such as the instruction not to cut or fold a pad; it ranks last because it works only if people act on it (Chapter 22).
- Inherently safe design
- The first and preferred option for controlling a risk under ISO 14971, in BSI's wording to eliminate the risk by the design itself, such as a return pad large enough to stay below the threshold of necrosis when two thirds of it lifts (Chapter 22).
- Interface
- A place where two parts of an energy system meet, or where a part meets a person: high-frequency power, signals, data or procedures cross it, and the party that writes its specification owns it (Chapter 21).
- Irreversible electroporation (IRE)
- Tumor ablation by electric pulses that open membrane pores the cells cannot repair (Chapter 15).
- Isolated output
- A generator output that floats electrically instead of being referenced to earth, so that current returns through the generator's own return lead rather than through earthed contacts on the patient; high-frequency current that still escapes to earth through stray capacitance is limited as leakage (Chapters 8 and 20).
- Joule heating
- Heating by current flowing through a resistance; in tissue, the heat released in each unit of volume is the square of the current density divided by the conductivity (Chapter 3).
- Joule–Thomson effect
- The temperature change of a real gas under high pressure expanding through a narrow orifice without doing work; argon and nitrogen cool, while helium, at room temperature, warms (Chapter 14).
- Langevin stack
- A few piezoelectric rings clamped between a heavy back mass and a front mass by a bolt through the middle, which keeps the rings in compression; the standard arrangement of an ultrasonic instrument's transducer (Chapter 11).
- Lateral thermal spread
- Thermal damage beyond the tissue a device was meant to affect, measured after the fact by cutting and staining the tissue and measuring the damaged zone under magnification (Chapter 2).
- Lethal isotherm
- The surface of equal temperature at the temperature that kills tissue, and so the edge of an ablation zone: about 60 °C in RF ablation, and below about −20 to −40 °C in cryoablation, several millimeters inside the visible edge of the ice (Chapters 13 and 14).
- Load
- The tissue as a generator sees it: an electrical impedance that changes as the tissue heats and dries, from about 20 Ω in a sealer's jaws to 2 kΩ or more across an arc (Before we start; Chapter 10).
- Maximum output voltage
- IEC 60601-2-2's figure for a generator: for each mode, the highest peak voltage that can appear between its patient connections; FDA asks for it in the generator's labeling (Chapters 10 and 21).
- Means of protection
- The barriers that keep a device's mains side apart from the patient and the operator, even after a fault: insulation, clearances through air and creepage distances along surfaces, each rated for a test voltage, with patient protection and operator protection rated separately (IEC 60601-1; Chapter 20).
- Medical device report
- A report to FDA of a death, serious injury or malfunction involving a device, which FDA counts by product code; FDA counted 46,673 under the ultrasonic-instrument code LFL from 2002 to 2021, 95 % of them malfunctions (Chapter 26).
- Monopolar
- Electrosurgery in which current flows from an active electrode through the patient to a return pad (Chapter 8).
- Motional current
- The current in the resonant branch of a piezoelectric transducer, proportional to the velocity of the blade; the generator regulates it to hold the blade's motion constant (Chapter 11).
- Neutral electrode
- IEC 60601-2-2's term for the return electrode, or pad; in the standard's heating test a pad carries a current set by the weight category of patient it is rated for, and the skin beneath it must not warm by more than 6 °C (Chapter 25).
- Node
- A point of no motion in a standing wave, half a wavelength from the next; the supports of an ultrasonic waveguide sit at nodes (Chapter 11).
- Nominal ocular hazard distance
- The distance along the direct beam over which a laser's exposure stays above the eye's maximum permissible exposure; reflections and focusing optics define a wider nominal hazard zone (Chapter 19).
- Notified body
- An organization designated under the EU's Medical Device Regulation to assess the conformity of devices; for a class IIb device, under the usual route, it audits the maker's quality management system, assesses the technical documentation of at least one representative device in each generic device group and issues certificates (Chapters 22 and 27).
- Particular standard
- A standard for one type of medical electrical equipment that modifies the general standard and overrides it where the two differ, such as IEC 60601-2-2 for electrosurgery (Chapter 25).
- Piezoelectric effect
- The change of shape of certain crystals and ceramics when a voltage is applied across them, and the voltage they produce when they are squeezed (Chapter 11).
- Population inversion
- More atoms in the upper laser level than in the level below it, the state a pump creates so that light passing through the gain medium is amplified by stimulated emission rather than absorbed (Chapter 12).
- Predicate device
- A legally marketed device to which a new device is shown to be substantially equivalent in a 510(k) (Chapter 9).
- Premarket approval (PMA)
- FDA's route for class III devices, the class of highest risk, by which a device is approved rather than cleared; the pulsed field ablation systems approved in the US, four between December 2023 and November 2024 and two more in 2025, reached patients this way (Chapters 16, 22 and 26).
- Process standard
- A standard that governs how a device is designed and made rather than what it does, such as ISO 13485 for the quality management system, ISO 14971 for risk management, IEC 62304 for the software life cycle and IEC 62366-1 for usability engineering (Chapter 25).
- Product code
- The three-letter code under which FDA files every record for one type of device, such as GEI for electrosurgical devices; a code can hold devices beyond the surgical ones, such as aesthetic RF devices under GEI (Chapter 23).
- Product standard
- A standard for how a kind of product's output is measured or how it is classed, such as IEC 61847 for ultrasonic surgical systems and IEC 60825-1 for laser hazard classes (Chapter 25).
- Protective measure
- The second option for controlling a risk under ISO 14971, after inherently safe design; on a lifting pad, the split pad and its monitor, which stops the generator when the impedance between the halves rises (Chapter 22).
- Pulsed field ablation (PFA)
- Irreversible electroporation used in the heart, to make the lines of scar that treat atrial fibrillation, usually with short biphasic pulses (Chapter 15).
- Quality Management System Regulation (QMSR)
- The US rule for device makers' quality systems, 21 CFR Part 820, which since 2 February 2026 incorporates ISO 13485:2016 by reference and replaces the older Quality System Regulation (Chapter 26).
- Rated accessory voltage
- The highest peak voltage an active accessory is designed to withstand, stated in its instructions for use; its insulation is tested at 120 % of it for 30 seconds, and the generator's peak in every mode used with it must stay at or below it (Chapters 18 and 21).
- Rated load
- The impedance at which a maker specifies a generator mode's power and waveform, such as 20 Ω for the FT10's sealing and 300 Ω for its monopolar cut (Chapter 10).
- Recall class
- FDA's grading of a correction or removal of a marketed device, on a scale of its own, separate from the device classes. FDA calls Class I "the most serious type of recall", and such a recall need not remove anything, as with the 2025 Varipulse correction of the catheter's instructions; less serious actions fall in Class II (Chapter 26).
- Recognized consensus standard
- A standard, in a stated edition, that FDA lists as one to which a maker may declare conformity in a premarket submission; FDA entered edition 6.1 of IEC 60601-2-2 on 25 May 2026 and accepts declarations to edition 6.0 until 1 July 2029 (Chapter 25).
- Recurrent test
- A safety test of medical electrical equipment in service, made at intervals and after repair under IEC 62353, whose abstract describes safety tests, such as leakage, and no test of output; a generator's output is checked by its maker's own procedure (Chapter 21).
- Return electrode
- The large pad on the patient's skin through which monopolar current leaves the body and returns to the generator; IEC 60601-2-2 calls it the neutral electrode (Before we start; Chapter 8).
- Rule 9
- The EU Medical Device Regulation's classification rule for active therapeutic devices that administer or exchange energy: class IIa unless they may do so in a potentially hazardous way, then class IIb, judged by the nature, density and site of application of the energy (Chapter 27).
- Sealing
- Closing a vessel held shut under pressure by heating it until the collagen of its walls fuses into a band strong enough to replace a suture, so that it stays shut when it is divided (Before we start; Chapter 1).
- Spark gap
- A small air gap across which each spark sets off a burst of high-frequency oscillation; Bovie's apparatus made both its currents this way (the opening story; Chapter 8).
- Special controls
- In the US, requirements written for one class II device type, such as particular tests and labeling, in addition to the rules for every device (Chapter 26).
- Split pad
- A return pad in two halves: the generator passes a small interrogating current from one half to the other through the skin, and stops when the impedance between them leaves its window, as it does when the pad lifts (Chapter 8).
- Summative evaluation
- IEC 62366-1's name, as the Johner Institute gives it, for human factors validation testing, the test of the final user interface at the end of development (Chapter 22).
- Thermal relaxation time
- The time a heated layer takes to lose its heat by conduction, roughly its thickness squared divided by four times the thermal diffusivity: about half a millisecond for the 16 µm layer that absorbs CO₂ light (Chapter 5).
- Traceability
- The links from each claim to the requirements that serve it, the tests that verify them and the label line that limits the claim, kept so that a change shows which tests must be repeated (Chapter 22).
- Transient ischemic attack
- A brief stroke-like episode; with stroke, the event that 4 of 132 patients in the Varipulse catheter's US evaluation had, about 3 % against an expected 1 % or less (Chapters 22 and 24).
- Use error
- In FDA's human factors guidance, a user action or lack of action that was different from what the manufacturer expected (Chapter 22).
- Use life
- The number of times a reusable instrument can be reused, which FDA's reprocessing guidance asks the maker to state from testing, with a way for users to tell when it is exceeded; Intuitive gives 10 uses for its da Vinci monopolar scissors and Covidien 50 for its reusable pencil (Chapter 21).
- Validation
- The assessment of whether devices made to the design meet user needs and intended uses, for a sealer by a chronic animal study, simulated use and human factors validation testing (Chapter 22).
- Vaporization
- Explosive boiling of the water in cells, which bursts them when energy arrives in a thin layer faster than steam can escape; in electrosurgery it begins once the water in the cells reaches about 100 °C, and a moving line of bursting cells is an incision (Chapter 1).
- Verification
- The check of each design output against its design input, such as a zero-failure burst test of seals (Chapter 22).
Sources.
Sources are listed by the chapter in which they are first used. Figures from manufacturers, and from studies written by manufacturers' staff, are identified as such in the text. Regulatory status and standards editions are as of October 2026.
- Opening — Li A.C., Isaacson G. Bovie, Cushing, and the era of electrosurgery. Ear Nose Throat J, published online 9 October 2024. doi:10.1177/01455613241289692 (secondary).
- Opening — DeLeon M.F., Yeo C.J., Maxwell P.J. IV. The evolution of cauterization: from the hot iron to the Bovie. Gibbon Society Historical Profiles, Thomas Jefferson University, 1574–1575 (December 2011). jdc.jefferson.edu/cgi/viewcontent.cgi?article=1039&context=gibbonsocietyprofiles (secondary; journal of first publication unconfirmed).
- Opening — Dagradi T. Electrosurgical in the operating room. Harvey Cushing/John Hay Whitney Medical Library, Yale University (12 September 2014). library.medicine.yale.edu/?p=1650 (secondary).
- Opening — Bovie W.T. Electrosurgical apparatus. US Patent 1,813,902, filed 18 January 1928, granted 14 July 1931; assigned to the Liebel-Flarsheim Company, Cincinnati. patents.google.com/patent/US1813902 (read as OCR text; filed in 1928, so it may describe the design of that year rather than the unit of 1926).
- Opening — Carter P.L. The life and legacy of William T. Bovie. Am J Surg 205(5), 488–491 (2013). doi:10.1016/j.amjsurg.2012.12.005 (not seen; the source Li and Isaacson give for the shock through a retractor).
- Opening — Cushing H., Bovie W.T. Electro-surgery as an aid to the removal of intracranial tumors. With a preliminary note on a new surgical-current generator by W.T. Bovie. Surg Gynecol Obstet 47, 751–784 (1928) (not seen; Cushing's words on first seeing the loop as quoted in Li and Isaacson 2024).
- Ch. 1 — Pearce J. Practical computer modeling for medical device development, Part 6: Including prediction of irreversible thermal alterations in numerical model work. Workshop notes, 37th IEEE EMBC, Milan (2015). embs.org/wp-content/uploads/2016/08/ws2015_5_Pearce_notes.pdf
- Ch. 1 — Sapareto S.A., Dewey W.C. Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys 10, 787–800 (1984).
- Ch. 1 — Niemz M.H. Laser-Tissue Interactions, Chapter 3. Springer.
- Ch. 1 — Holmes K.R. Thermal properties of biological tissues (compilation). users.ece.utexas.edu/~valvano/research/Thermal.pdf
- Ch. 1 — US FDA. Premarket Notification (510(k)) Submissions for Electrosurgical Devices for General Surgery. Guidance (2016; current version 9 March 2020). fda.gov/media/87995/download
- Ch. 1 — Yarmolenko P.S., Moon E.J., Landon C., Manzoor A., Hochman D.W., Viglianti B.L., Dewhirst M.W. Thresholds for thermal damage to normal tissues: an update. Int J Hyperthermia 27(4), 320–343 (2011). pmc.ncbi.nlm.nih.gov/articles/PMC3609720
- Ch. 1 — Engineering ToolBox. Data pages: latent heat of evaporation; water, thermal conductivity and thermal diffusivity; thermal conductivity of metals and alloys; speed of sound in water and in air (undated; read 20 October 2026). engineeringtoolbox.com (secondary).
- Ch. 1 — Dargent D., Schettini S. Physical bases of electrosurgery. Contemporary OB/GYN (1 September 2005). contemporaryobgyn.net/view/physical-bases-electrosurgery (trade journal).
- Ch. 2 — Berjano E.J. Theoretical modeling for radiofrequency ablation: state-of-the-art and challenges for the future. BioMed Eng OnLine 5, 24 (2006).
- Ch. 2 — Hall J.E. Guyton and Hall Textbook of Medical Physiology, 12th ed., Chapter 17 (2011).
- Ch. 2 — Mohammadi A., Bianchi L., Asadi S., Saccomandi P. Measurement of ex vivo liver, brain and pancreas thermal properties as function of temperature. Sensors 21(12), 4236 (2021). doi:10.3390/s21124236
- Ch. 2 — The use of a physiologically based pharmacokinetic model to evaluate deconvolution measurements of systemic absorption. BMC Clin Pharmacol 3, 1 (2003), Table 1, organ weights and blood flows for a standard human (70 kg, 20 % fat). bmcclinpharma.biomedcentral.com/articles/10.1186/1472-6904-3-1/tables/1 (table read; authors not shown on the table page).
- Ch. 2 — Chambers R., Sarno D., Roweton S. New bipolar electrosurgical vessel sealing device provides improved performance and procedural efficiency. Med Devices (Auckl) 18, 75–86 (2025). doi:10.2147/MDER.S498873. Funded by Medtronic; all three authors are Medtronic employees.
- Ch. 2 — Lu D.S.K., Raman S.S., Vodopich D.J., Wang M., Sayre J., Lassman C. Effect of vessel size on creation of hepatic radiofrequency lesions in pigs: assessment of the 'heat sink' effect. AJR 178(1), 47–51 (2002). doi:10.2214/ajr.178.1.1780047
- Ch. 2 — Lu D.S.K., Raman S.S., Limanond P., Aziz D., Economou J., Busuttil R., Sayre J. Influence of large peritumoral vessels on outcome of radiofrequency ablation of liver tumors. J Vasc Interv Radiol 14(10), 1267–1274 (2003). doi:10.1097/01.RVI.0000092666.72261.6B
- Ch. 2 — Pennes H.H. Analysis of tissue and arterial blood temperatures in the resting human forearm. J Appl Physiol 1(2), 93–122 (August 1948). doi:10.1152/jappl.1948.1.2.93
- Ch. 3 — Guy A.W. Biological effects of electromagnetic radiation. Engineering and Technology History Wiki (on d'Arsonval, 1891 and 1893). ethw.org
- Ch. 3 — IEC 60601-2-2: Particular requirements for the basic safety and essential performance of high frequency surgical equipment and high frequency surgical accessories. Ed. 6.0 (2017) and AMD1 (2023); consolidated Ed. 6.1, corrected version 2025-11.
- Ch. 3 — Kotnik T., Pucihar G. Induced transmembrane voltage: theory, modeling, and experiments. In Advanced Electroporation Techniques in Biology and Medicine, 51–70 (2010).
- Ch. 3 — Woloszko J., Marion D.W. (assigned to ArthroCare). Method and system of reduction of low frequency muscle stimulation during electrosurgical procedures. US Patent application 2011/0245826 A1, filed 6 April 2010, published 6 October 2011.
- Ch. 3 — Gabriel S., Lau R.W., Gabriel C. The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues. Phys Med Biol 41, 2271–2293 (1996); values computed from the model's parameters as tabulated by IFAC-CNR, webnir.eu/04-Dosimetry/04-Gabriel.php.
- Ch. 3 — IEC TR 61289: High frequency surgical equipment and high frequency surgical accessories – Operation and maintenance, term 3.18 (citing IEC 60601-2-2:2017, 201.3.220), as reproduced at certifico.com (secondary host).
- Ch. 3 — Soma Tech Intl. Product pages reproducing the Valleylab Force FX and Force FX-C specifications (undated). somatechnology.com/pdfFiles/Medtronic-Covidien-Valleylab-Force-FX-ESU.pdf; somatechnology.com/pdfFiles/Covidien-Valleylab-Force-FX-C.pdf (reseller; secondary copy of manufacturer data).
- Ch. 3 — US FDA. 510(k) summaries K251108, Erbe ESU Model VIO 3n with Accessories (Erbe Elektromedizin; received 11 April 2025, cleared 29 August 2025), and K240932, HybridTherm System, of the VIO 3 generator, the ERBECRYO 2 cryosurgical unit and HybridTherm probes (cleared 5 August 2024). accessdata.fda.gov/cdrh_docs/pdf25/K251108.pdf; .../pdf24/K240932.pdf
- Ch. 3 — US FDA. 510(k) summaries K111202 (THUNDERBEAT, 2012), K211838 (USG-410, 2021) and K252150 (THUNDERBEAT II, 7 October 2025).
- Ch. 4 — Vieira-Coimbra M., Ferreira H. Ultrasonic surgical aspiration (CUSA) for laparoscopic excision of endometriosis. Front Surg 12, 1735940 (2025). Publication support from Integra LifeSciences, maker of CUSA.
- Ch. 4 — Culjat M.O., Goldenberg D., Tewari P., Singh R.S. A review of tissue substitutes for ultrasound imaging. Ultrasound Med Biol 36(6), 861–873 (2010). doi:10.1016/j.ultrasmedbio.2010.02.012
- Ch. 4 — Verdaasdonk R.M., van Swol C.F.P., Grimbergen M.C.M. High-speed and thermal imaging of the mechanism of action of the Cavitational Ultrasonic Aspirator (CUSA). In Ryan T.P. (ed.), Surgical Applications of Energy, Proc SPIE 3249, 72–84 (1998).
- Ch. 4 — Vlaisavljevich E., Lin K.W., Maxwell A., Warnez M.T., Mancia L., Singh R., Putnam A.J., Fowlkes B., Johnsen E., Cain C., Xu Z. Effects of ultrasound frequency and tissue stiffness on the histotripsy intrinsic threshold for cavitation. Ultrasound Med Biol 41(6), 1651–1667 (2015). doi:10.1016/j.ultrasmedbio.2015.01.028. Funded by NIH, NSF, the American Cancer Society, the Hartwell Foundation and the Focused Ultrasound Foundation; Cain, Fowlkes and Xu declare financial interests and/or other relationships with HistoSonics, maker of histotripsy systems.
- Ch. 4 — Homma T. Advances and safe use of energy devices in lung cancer surgery. Gen Thorac Cardiovasc Surg 70(3), 207–218 (2022).
- Ch. 4 — Yu et al. Prospects for clinical translation of histotripsy in liver applications. Quant Imaging Med Surg 16(1), 92 (2026).
- Ch. 4 — US FDA. 510(k) K172721, Focal One (EDAP; received 8 September 2017, cleared 7 June 2018; 21 CFR 876.4340, product code PLP).
- Ch. 4 — US FDA. De Novo decision summary DEN150011, Sonablate 450 (granted 9 October 2015).
- Ch. 4 — Ethicon. Surgical generator for ultrasonic and electrosurgical devices. US Patent 9,060,776 B2 (2015); Ultrasonic transducer for surgical instrument. US Patent 10,420,580 B2 (2019).
- Ch. 4 — Ethicon. US Patent 6,678,621, read at Google Patents (manufacturer; 'approximately 55.5 kHz' and a blade excursion of 'about 40 to 100 microns'; title, inventors and dates not recorded).
- Ch. 5 — Hale G.M., Querry M.R. Optical constants of water in the 200 nm to 200 µm wavelength region. Appl Opt 12, 555–563 (1973); with the data sets of Pope R.M. and Fry E.S. (1997) for 532 nm and Segelstein D.J. (1981) for 1,470 nm; data via omlc.org/spectra/water.
- Ch. 5 — Prahl S. Tabulated molar extinction coefficient for hemoglobin in water (1998). omlc.org/spectra/hemoglobin
- Ch. 5 — Pang P., Andreana S., Aoki A., et al. (Science and Research Committee, Academy of Laser Dentistry). Laser energy in oral soft tissue applications (position paper). J Laser Dent 20(2), 65–73 (2012). laserdentistry.org
- Ch. 6 — IEC 60601-2-22: Particular requirements for surgical, cosmetic, therapeutic and diagnostic laser equipment. Ed. 4.0 (2019) and AMD1 (2026).
- Ch. 6 — IEC 61847: Ultrasonics — Surgical systems — Measurement and declaration of the basic output characteristics. Ed. 2.0 (20 May 2025). webstore.iec.ch/en/publication/73373
- Ch. 6 — IEC 60601-2-62: Particular requirements for high intensity therapeutic ultrasound (HITU) equipment. Ed. 1.0 (9 July 2013). webstore.iec.ch/en/publication/2678
- Ch. 6 — IEC 60825-1: Safety of laser products — Part 1: Equipment classification and requirements. Ed. 3.0 (15 May 2014). webstore.iec.ch/en/publication/3587
- Ch. 6 — Lubner M.G., Brace C.L., Hinshaw J.L., Lee F.T. Microwave tumor ablation: mechanism of action, clinical results, and devices. J Vasc Interv Radiol 21(8 Suppl), S192–S203 (2010). Two authors, Lee and Brace, held stock in NeuWave Medical, maker of the NEUWAVE system.
- Ch. 7 — Espada M. et al. Insulation failure in robotic and laparoscopic instrumentation: a prospective evaluation. Am J Obstet Gynecol (2011).
- Ch. 7 — Intuitive. da Vinci energy page (MAT03921 V3 US, September 2026), generators page (MAT03429 V3 US, September 2026) and SynchroSeal page (MAT03430 V3 US, March 2026; its footnote lists the Vessel Sealer Extend with the E-100, E-200 and Erbe VIO dV). intuitive.com/en-us/products-and-services/da-vinci/energy (manufacturer).
- Ch. 7 — Olympus. THUNDERBEAT II launch, news release (Tokyo and Hamburg, 21 October 2025), read as distributed by JCN Newswire. acnnewswire.com/press-release/english/103180/ (company statement).
- Ch. 7 — Johnson & Johnson MedTech. ENDOPATH XCEL Trocars product page (undated; accessed 21 October 2026). jnjmedtech.com/en-US/products/surgery/surgical-instruments/endopath-xcel-trocars/ (manufacturer; sizes 5 to 15 mm).
- Ch. 8 — Anderson R.K. (Valleylab). Electrosurgical generator. US Patent 3,699,967, filed 30 April 1971, granted 24 October 1972 (an all-solid-state generator with isolated outputs).
- Ch. 8 — Koninckx P.R., Ussia A., Amro B., Prantner M., Keckstein J., Keckstein S., Adamyan L., Wattiez A., Romeo A. Electrosurgery: heating, sparking and electrical arcs. Facts Views Vis ObGyn 16(3), 281–290 (2024). doi:10.52054/FVVO.16.3.026. The article page lists one author as affiliated with BOWA-electronic GmbH & Co. KG, a generator maker.
- Ch. 8 — Ehninger M.D. et al. (MegaDyne). Universal self-limiting electrosurgical return electrode. US Patent 10,085,791 B2, filed 26 December 2014, granted 2 October 2018 (manufacturer; cites the Emergency Care Research Institute for the threshold of necrosis).
- Ch. 8 — Newton D.W., Hulett F.M. III, Ovens C. (Valleylab). Return electrode monitoring system for use during electrosurgical activation. US Patent 4,416,277, filed 3 November 1981, granted 22 November 1983.
- Ch. 8 — AORN. Sample policy and procedure: Electrosurgery (2016). aorn.org (professional association).
- Ch. 8 — Medtronic. Valleylab FT10 energy platform, specification guide, US-SE-2300184 (November 2023). medtronic.com/content/dam/medtronic-wide/public/united-states/products/surgical-energy/valleylab-ft10-generator-specifications-information-sheet.pdf (manufacturer; re-opened 21 October 2026).
- Ch. 8 — ASGE Technology Committee (Tokar J.L., Barth B.A., Banerjee S. et al.). Electrosurgical generators. Technology status evaluation report. Gastrointest Endosc 78(2) (2013). asge.org (professional society).
- Ch. 8 — Nessler N., Reischer W., Salchner M. (University of Innsbruck). Electronic skin: test device for electrosurgical neutral electrodes. 12th IMEKO TC4 Symposium, Zagreb, 25–27 September 2002, citing AAMI HF18. imeko.org/publications/tc4-symp-2002/IMEKO-TC4-2002-223.pdf; ANSI/AAMI HF18:2001 is listed as historical (ANSI Webstore, read 21 October 2026).
- Ch. 8 — Valleylab. Force EZ-C Users Guide, part 1015380 (effective November 2008), and Force EZ Electrosurgical Generator CS/8CS Service Manual (February 2011), read in previews on bioclinicalservices.com.au (front chapters only). Manufacturer documents.
- Ch. 8 — Fuchshuber P., Madani A., Jones D.B., Schwaitzberg S. The Fundamental Use of Surgical Energy (FUSE) program: promoting operating room safety. Metabolic & Bariatric Care Times (Bariatric Times) 11(10), 22–24 (2014). bariatrictimes.com/the-fundamental-use-of-surgical-energy-fuse-program-promoting-operating-room-safety/
- Ch. 9 — US FDA. 510(k) K981916, LigaSure Vessel Sealing System (cleared 28 August 1998); 510(k) summary K043273 (2005).
- Ch. 9 — US FDA. Premarket Notification (510(k)) Submissions for Bipolar Electrosurgical Vessel Sealers for General Surgery. Guidance (August 2016). fda.gov/media/87987/download
- Ch. 9 — Okhunov Z., Yoon R., Lusch A., Spradling K., Suarez M., Kaler K.S., Patel R., Hwang C., Osann K., Huang J., Lee T., Landman J. Evaluation and comparison of contemporary energy-based surgical vessel sealing devices. J Endourol 32(4), 329–337 (2018). doi:10.1089/end.2017.0596. escholarship.org/content/qt0x84c136/qt0x84c136.pdf. University of California, Irvine, and Duke University; no funding statement; "No competing financial interests exist"; how the devices were obtained is not stated.
- Ch. 9 — US FDA. 510(k) K120729, HARMONIC ACE Shears + Adaptive Tissue Technology (Ethicon Endo-Surgery; cleared 17 May 2012; the summary's conclusion gives sealing of vessels up to 5 mm). accessdata.fda.gov/cdrh_docs/pdf12/K120729.pdf
- Ch. 9 — Medtronic. LigaSure RAS Maryland jaw vessel sealer information sheet, US-SE-2600945 (September 2026). medtronic.com (manufacturer; burst pressure standard of 360 mmHg, three times systolic).
- Ch. 9 — Truckai C., Shadduck J.H. (SurgRx; later Ethicon Endo-Surgery). Surgical sealing surfaces and methods of use. US Patent 7,220,951 B2 (filed 2005, granted 2007) (manufacturer; positive-temperature-coefficient jaw material).
- Ch. 9 — Nezhat C., Lewis M., King L.P. Laparoscopic vessel sealing devices. In Prevention and Management of Laparoendoscopic Surgical Complications, 3rd ed., chapter 11, SLS (published online 3 February 2021). sls.org/the-3rd-edition-prevention-management/chapter-11/
- Ch. 10 — Gilbert J.A., Friedrichs D.A. (Covidien). Electrosurgical generator for minimizing neuromuscular stimulation. US Patent 9,839,470 B2 (filed 2015, granted 12 December 2017).
- Ch. 10 — Friedrichs D.A., Erickson R.W., Gilbert J. A new dual current-mode controller improves power regulation in electrosurgical generators. IEEE Trans Biomed Circuits Syst 6(1), 39–44 (2012). The same controller is patented by the three authors as US 9,379,643 B2 (filed 21 December 2011; assigned to Covidien LP and the University of Colorado); the paper's affiliations were not read.
- Ch. 10 — Friedrichs D., Erickson R., Gilbert J. Electrosurgical generator controller for regulation of electrosurgical generator output power. US Patent 9,379,643 B2 (filed 21 December 2011, granted 28 June 2016; assigned to Covidien LP and the University of Colorado).
- Ch. 10 — Rafiq M.M., Ibeas A., Ullah N. Expert knowledge-based peak current mode control of electrosurgical generators for improved output power regulation. J Electr Bioimpedance 14(1), 32–46 (2023). doi:10.2478/joeb-2023-0005 (CECOS University, Universitat Autònoma de Barcelona, Taif University; no conflict of interest stated).
- Ch. 10 — IEC 60601-2-2:2017 (Ed. 6.0) and IEC 60601-2-2:2017/AMD1:2023, sample pages (introduction, definitions, 201.4.2.3.101, 201.7.9.2.14, 201.8.7.3.101, 201.12.4.3.101). cdn.standards.iteh.ai (read 20 October 2026; the full clause texts on load range, test currents and non-activated outputs were not reached); and the Ed. 6.0 sample pages to 201.3.230, including 201.3.218 HEATING FACTOR and 201.3.219 HIGH CURRENT MODE ('MAXIMUM OUTPUT CURRENT and maximum DUTY CYCLE'), cdn.standards.iteh.ai/samples/22495/e21eb15864fc4ca4b4b8368c9e52eb65/IEC-60601-2-2-2017.pdf (read 21 October 2026).
- Ch. 11 — Govekar et al. Residual heat of laparoscopic energy devices: how long must the surgeon wait? SAGES abstract (c. 2011).
- Ch. 11 — Timm et al. Sealing vessels up to 7 mm in diameter solely with ultrasonic technology. Med Devices (Auckl) (2014). Authors affiliated with the manufacturer.
- Ch. 11 — Engineering ToolBox. Sound speed in solids (extensional speed of titanium). engineeringtoolbox.com
- Ch. 11 — US FDA. 510(k) summary K132612, HARMONIC ACE+ Shears with Advanced Hemostasis (Ethicon Endo-Surgery; received 21 August 2013, cleared 17 October 2013; vessels up to and including 7 mm in the Advanced Hemostasis mode), marketed as HARMONIC ACE+7.
- Ch. 11 — Thomas A.E. (UltraCision). Method and apparatus for ultrasonic surgical cutting and hemostatis. US Patent 5,026,387 (filed 12 March 1990; later held by Ethicon Endo-Surgery) (manufacturer).
- Ch. 11 — Wiener E.T., Donofrio W.T. (Ethicon Endo-Surgery). Ultrasonic surgical system within digital control. US Patent 7,476,233 B1 (filed 20 October 2000) (manufacturer).
- Ch. 11 — SOMA Tech Intl. Ethicon GEN11 generator product sheet (undated). somatechnology.com/pdfFiles/Ethicon-Gen11-Generator-Harmonic-Scalpel.pdf (reseller; secondary copy of manufacturer data: ultrasonic 150 V RMS maximum, 35 W continuous, unless the instrument's IFU states otherwise).
- Ch. 11 — Houser K.L. (Ethicon Endo-Surgery). Temperature controlled ultrasonic surgical instruments. US Patent 9,044,261 B2 (filed 2008, granted 2 June 2015) (manufacturer).
- Ch. 11 — Ethicon. Ethicon receives 510(k) clearance for new HARMONIC ACE+7 Shears with Advanced Hemostasis, news release (Cincinnati, 28 October 2013). jnj.com/media-center/press-releases/ (company statement).
- Ch. 11 — Olympus America. Olympus introduces the world's only integrated bipolar and ultrasonic energy platform, news release (Center Valley, PA, 21 March 2012). medical.olympusamerica.com/node/1504 (company statement).
- Ch. 12 — Ling S.J., Sanny J., Moebs W. University Physics Vol. 3, section 8.6, Lasers. OpenStax (2016).
- Ch. 12 — Traxer O., Keller E.X. Thulium fiber laser: the new player for kidney stone treatment? World J Urol (2020).
- Ch. 12 — Sánchez-Puy A. et al. New generation pulse modulation in holmium:YAG lasers. J Clin Med 11, 3208 (2022).
- Ch. 12 — Aktas A.R. et al. Lasers Med Sci 30, 1583 (2015), comparing 980 nm and 1,470 nm endovenous ablation.
- Ch. 12 — US FDA. 510(k) summary K120870, GreenLight MoXy side-firing fiber, 750 µm core, for the GreenLight XPS (American Medical Systems; cleared 15 June 2012).
- Ch. 12 — Zorn K.C., Liberman D. GreenLight 180W XPS photovaporization of the prostate: how I do it. Can J Urol 18(5), 5918–5926 (2011). Zorn reports consulting and proctoring for American Medical Systems, maker of GreenLight (secondary; KTP for the 80 W PV, lithium triborate for the HPS and XPS).
- Ch. 12 — Olympus America. SOLTIVE SuperPulsed Laser System product page, specifications (accessed 21 October 2026). content.medical.olympusamerica.com (manufacturer; SOLTIVE Premium: 60 W, 0.025 to 6 J, 1 to 2,400 Hz, 100 to 240 V at 1,200 VA, 40 kg, air-cooled).
- Ch. 12 — Fried N.M. Recent advances in infrared laser lithotripsy. Biomed Opt Express 9(9), 4552 (2018) (not opened; the Optica and PMC pages were blocked).
- Ch. 12 — Boston Scientific. Lumenis Pulse 120H brochure, URO-1769904-AB (© 2024), bostonscientific.com; Lumenis. Pulse 120H operator's manual UM-10012510 Rev. C (copy hosted by Agiliti Health, agilitihealth.com; read 21 October 2026) (manufacturer).
- Ch. 12 — Aldoukhi A.H., Roberts W.W., Hall T.L., Ghani K.R. Holmium laser lithotripsy in the new stone age: dust or bust? Front Surg 4, 57 (2017). doi:10.3389/fsurg.2017.00057 (review; Ghani consults for Lumenis and Boston Scientific).
- Ch. 12 — Hardy L.A. Improving thulium fiber laser lithotripsy efficiency. PhD dissertation, University of North Carolina at Charlotte (2018), advisor N.M. Fried. ninercommons.charlotte.edu (secondary).
- Ch. 12 — NASA Earth Science Technology Office. Early laser research spins off to support medical applications. esto.nasa.gov/25years/medlaser (agency page; secondary).
- Ch. 12 — Gentec-EO. What is a CO2 laser and how does it work? (16 December 2021). gentec-eo.com (laser-measurement maker; secondary).
- Ch. 13 — US FDA. 510(k) summary K130399, NEUWAVE Certus 140 (2013).
- Ch. 13 — Ringe K.I. et al. Experimental evaluation of the heat sink effect in hepatic microwave ablation. PLOS One 10(7), e0134301 (2015).
- Ch. 13 — Huang et al. A review of antenna designs for percutaneous microwave ablation. Phys Med 84 (2021).
- Ch. 13 — De Vita E. et al. A review on radiofrequency, laser, and microwave ablations and their thermal monitoring through fiber Bragg gratings. iScience 26(11), 108260 (2023).
- Ch. 13 — Ahmed M. et al. Image-guided tumor ablation: standardization of terminology and reporting criteria, a 10-year update. Radiology (2014).
- Ch. 13 — Hong K., Georgiades C.S. Radiofrequency ablation: mechanism of action and devices. In Percutaneous Tumor Ablation, chapter 1, Thieme (2011). doi:10.1055/b-0034-81499 (read as reproduced at radiologykey.com).
- Ch. 13 — Hinshaw J.L. et al. Percutaneous tumor ablation tools: microwave, radiofrequency, or cryoablation—what should you use and why? RadioGraphics 34(5), 1344–1362 (2014). NIH-funded; three authors held shares in NeuWave Medical, maker of the NEUWAVE system, one of them also a patent holder and board member.
- Ch. 13 — Boston Scientific. Radiofrequency ablation brochure, RF 3000 generator and LeVeen electrodes, PI-298512-AA (February 2015) (manufacturer).
- Ch. 13 — AngioDynamics. Model 1500X RF generator page, angiodynamics.com/?p=466 (accessed 21 October 2026) (manufacturer).
- Ch. 13 — Multicenter and inter-software evaluation of ablative margins after thermal ablation of colorectal liver metastases. Eur Radiol (2024). doi:10.1007/s00330-024-10956-5 (author list not recorded when read).
- Ch. 13 — Simo K.A. et al. Microwave ablation using 915-MHz and 2.45-GHz systems: what are the differences? HPB 15(12), 991–996 (2013). pmc.ncbi.nlm.nih.gov/articles/PMC3843618
- Ch. 13 — US FDA. 510(k) summary K233838, Emprint ablation system (2024).
- Ch. 13 — Hui T.C.H., How G.Y., Chim M.S.M., Pua U. Comparative study of ablation zone of EMPRINT HP microwave device with contemporary 2.4 GHz microwave devices in an ex vivo porcine liver model. Diagnostics 13(16), 2702 (2023). doi:10.3390/diagnostics13162702 (funded by Medtronic).
- Ch. 13 — Cornelis A.A., Bodard S., Cornelis F.H. Lethal isotherm ratio to standardize the comparison of cryoablation needles and systems. CVIR Oncology 1, 3 (2025). doi:10.1007/s44343-025-00004-4 (not opened; the publisher's page returned an error on 21 October 2026).
- Ch. 13 — AngioDynamics Inc. Form 10-K for the fiscal year ended 31 May 2025. sec.gov/Archives/edgar/data/1275187/000127518725000014/ango-20250531.htm (manufacturer).
- Ch. 14 — Erinjeri J.P., Clark T.W.I. Cryoablation: mechanism of action and devices. J Vasc Interv Radiol 21(8 Suppl), S187–S191 (2010).
- Ch. 14 — US FDA. De Novo decision summary DEN220077, ProSense cryoablation system (granted 3 October 2025).
- Ch. 14 — IceCure Medical. The ICE3 clinical trial, company page (modified 13 October 2025). icecure-medical.com/?p=15444 (manufacturer).
- Ch. 14 — Boston Scientific. CX Needles Isotherm Guide, PI-754604-AC (© 2024): 0, −20 and −40 °C isotherms of single and clustered needles in 37 °C gel after two 10-minute freezes. bostonscientific.com (manufacturer; secondary for the isotherm ratio).
- Ch. 15 — US FDA. PMA P230017, PulseSelect pulsed field ablation system (Medtronic; approved 13 December 2023). openFDA, api.fda.gov/device/pma.json?search=pma_number:P230017
- Ch. 15 — TCTMD. Report of FDA's approval of Medtronic's PulseSelect, the first pulsed field ablation system approved in the US (14 December 2023). tctmd.com (secondary).
- Ch. 15 — Patil S., Kapa S., Deshmukh A.J. Pulsed field ablation: biophysical principles and practical considerations in atrial fibrillation treatment. Indian Pacing Electrophysiol J 26(2), 103–113 (2026). doi:10.1016/j.ipej.2026.03.006 (abstract only).
- Ch. 15 — US FDA. PMA P230030, FARAPULSE pulsed field ablation system (applicant Farapulse, Inc., Boston Scientific; approved 30 January 2024). openFDA, api.fda.gov/device/pma.json?search=pma_number:P230030
- Ch. 15 — US FDA. PMA P240013, Sphere-9 catheter and Affera ablation system, pulsed field and RF (Medtronic; approved 24 October 2024). openFDA, api.fda.gov/device/pma.json?search=pma_number:P240013
- Ch. 15 — US FDA. PMA P240006, VARIPULSE platform (Biosense Webster, Johnson & Johnson; approved 6 November 2024), with supplements S007 (labeling change to the recommended irrigation flow rate; received 1 April 2025, approved 26 June 2025) and S012 (TRUPULSE Generator software for automatic irrigation at 30 mL/min; received 30 September 2025, approved 23 December 2025). openFDA, api.fda.gov/device/pma.json?search=pma_number:P240006 (FDA states that openFDA data are unvalidated; 10 of 17 records read 20 October 2026).
- Ch. 15 — TCTMD. J&J halts Varipulse PFA cases in the US (8 January 2025); J&J restarts Varipulse PFA cases in the US (18 February 2025). tctmd.com (secondary).
- Ch. 15 — Ekanem E. et al. Safety of pulsed field ablation in more than 17,000 patients with atrial fibrillation in the MANIFEST-17K study. Nat Med (2024). doi:10.1038/s41591-024-03114-3. Retrospective registry of Boston Scientific's FARAPULSE at 106 centers in Europe and Israel; no external funding; many authors report ties to Boston Scientific.
- Ch. 15 — Guidepoint Qsight. Inside the pulsed field ablation market 2026 (25 June 2026; modified 30 June 2026). qsight.guidepoint.com (market-intelligence firm; market models built from the purchasing data of more than 4,500 US facilities).
- Ch. 15 — Kelly S. MedTech Dive report on a Citi physician survey (16 December 2024; 72 physicians, 36 electrophysiologists and 36 cardiovascular surgeons). medtechdive.com (secondary; bank research).
- Ch. 15 — Kos B. et al. Determination of lethal electric field threshold for pulsed field ablation in ex vivo perfused porcine and human hearts. Front Cardiovasc Med 10, 1160231 (2023). doi:10.3389/fcvm.2023.1160231. Table 3 compiles mean lethal thresholds reported in in vivo studies of other tissues. Funded by Medtronic; three authors were Medtronic employees.
- Ch. 15 — Davalos R.V., Mir L.M., Rubinsky B. Tissue ablation with irreversible electroporation. Ann Biomed Eng 33, 223–231 (2005) (full text not opened; values as in the authors' patent US 8,048,067, Example 1).
- Ch. 15 — AngioDynamics. NanoKnife System user manual, 160-104694 Rev 01, version 2.2.0, Table 10-1 (technical specifications). Read as hosted at manualslib.com (manufacturer document).
- Ch. 15 — US FDA. 510(k) records K060054, Oncobionic System (Oncobionics, Inc.; cleared 21 November 2006), and K102329, NanoKnife System (AngioDynamics; cleared 24 October 2011); product code OAB, 21 CFR 878.4400. openFDA, product_code:OAB (query run 21 October 2026).
- Ch. 15 — US FDA. 510(k) K242687, NanoKnife System (AngioDynamics; received 6 September 2024, cleared 6 December 2024; indication 'surgical ablation of soft tissue, including prostate tissue'; predicate K183385). accessdata.fda.gov/cdrh_docs/pdf24/K242687.pdf; AngioDynamics news release (9 December 2024), angiodynamics.gcs-web.com/node/22796 (company statement).
- Ch. 15 — Martin R.C.G. II, Durham A.N., Besselink M.G., et al. Irreversible electroporation in locally advanced pancreatic cancer: a call for standardization of energy delivery. J Surg Oncol (2016). doi:10.1002/jso.24404. The first author is a paid consultant for AngioDynamics.
- Ch. 15 — Martin R.C.G., Schwartz E., Adams J., Farah I., Derhake B.M. Intra-operative anesthesia management in patients undergoing surgical irreversible electroporation of the pancreas, liver, kidney, and retroperitoneal tumors. Anesth Pain Med 5(3), e22786 (2015).
- Ch. 15 — Campelo S.N., Huang P.-H., Buie C.R., Davalos R.V. Recent advancements in electroporation technologies: from bench to clinic. Annu Rev Biomed Eng 25, 77–100 (2023). doi:10.1146/annurev-bioeng-110220-023800. Authors hold patents in electroporation.
- Ch. 15 — Bi S. et al. Preclinical study of biphasic asymmetric pulsed field ablation. Front Cardiovasc Med 9, 859480 (2022). One author was employed by Tianjin Intelligent Health Medical Co., Ltd.
- Ch. 15 — US FDA. PMA P240044, Globe Pulsed Field System (Kardium Inc.; approved 27 August 2025), and P250022, Volt PFA System (Abbott Medical; approved 19 December 2025); product code QZI. openFDA, api.fda.gov/device/pma.json?search=pma_number:P240044 and P250022 (queries run 21 October 2026).
- Ch. 15 — Ha F.J., Han H.-C., Nerlekar N., Brown A.J., Kotschet E. Advent of pulsed field ablation for atrial fibrillation: state-of-the-art review. Rev Cardiovasc Med (2025). doi:10.31083/RCM47497 (secondary).
- Ch. 15 — Belfiore M.P., De Chiara M., Reginelli A. et al. An overview of the irreversible electroporation for the treatment of liver metastases: when to use it. Front Oncol 12, 943176 (2022). doi:10.3389/fonc.2022.943176 (review).
- Ch. 15 — Memorial Sloan Kettering Cancer Center. Irreversible electroporation ablation for colorectal metastases to the lung, protocol IRB 15-001A(5), amended 8 July 2016 (NCT02461550). cdn.clinicaltrials.gov/large-docs/50/NCT02461550/Prot_SAP_000.pdf (investigators' protocol, background section).
- Ch. 15 — FARAPULSE, Inc. ADVENT pivotal trial protocol CS0934, Rev. D (1 November 2021), NCT04612244. cdn.clinicaltrials.gov/large-docs/44/NCT04612244/Prot_002.pdf (manufacturer); Boston Scientific. What is PFA? FARAPULSE page, en-EU (read 21 October 2026). bostonscientific.com (manufacturer).
- Ch. 16 — US FDA. De Novo decision summary DEN220087, HistoSonics Edison system (decision 6 October 2023).
- Ch. 16 — HistoSonics. FDA submission for kidney tumors, news release via Business Wire (11 May 2026); company news page (accessed September 2026).
- Ch. 16 — US FDA. PMA P860046, Sonocare therapeutic ultrasound for the treatment of glaucoma (Sonocare, Inc.; received 1 October 1986, approved 30 June 1988; generic name ultrasound, cyclodestructive; product code LZR).
- Ch. 16 — US FDA. PMA P040003, ExAblate 2000 (InSightec; received 27 January 2004, approved 22 October 2004; ablation system, high intensity focused ultrasound (HIFU), MR-guided; product code NRZ).
- Ch. 16 — Tyshlek D. et al. Focused ultrasound development and clinical adoption: 2013 update. J Ther Ultrasound 2, 2 (2014) (secondary; dates FDA's first approval of a focused-ultrasound device to 1988).
- Ch. 16 — EDAP TMS. News release on withdrawing its De Novo application for Ablatherm HIFU and filing a 510(k), via BioSpace (15 October 2015) (company statement).
- Ch. 16 — US FDA. 510(k) K153023, Ablatherm Integrated Imaging (EDAP Technomed; received 15 October 2015, cleared 6 November 2015; predicate Sonablate 450; 21 CFR 876.4340, product code PLP). accessdata.fda.gov/cdrh_docs/pdf15/k153023.pdf
- Ch. 16 — Turner B., Cranston D. A review of high-intensity focused ultrasound. Int J Transl Med 4(1), 197–207 (2024). doi:10.3390/ijtm4010011. mdpi.com/2673-8937/4/1/11. No external funding; one author is an adjunct professor at Chongqing Medical University.
- Ch. 16 — HistoSonics. HistoSonics receives CE mark, unlocking access across Europe and accelerating global commercialization of the Edison histotripsy platform, news release via Business Wire (13 July 2026). businesswire.com/news/home/20260713505048 (company statement; names no notified body or class).
- Ch. 16 — US FDA. openFDA 510(k) endpoint, records for applicant HistoSonics (the De Novo of 6 October 2023, then K233466, K241902 and K252947), data last updated 12 October 2026; no kidney decision. api.fda.gov/device/510k.json?search=applicant:histosonics (query run 21 October 2026).
- Ch. 17 — Ibanez Jimenez et al. Novel multifunctional robotically assisted bipolar instrument for simultaneous radiofrequency sealing and transection: preclinical and single-center experience. BMC Surg (2022). doi.org/10.1186/s12893-022-01483-5 (comparator Harmonic ACE+7; abstract and text read 21 October 2026). Two authors employed by the manufacturer; funded by Intuitive (declarations not re-read).
- Ch. 17 — Rimonda R. et al. Electrothermal bipolar vessel sealing system vs. harmonic scalpel in colorectal laparoscopic surgery: a prospective, randomized study. Dis Colon Rectum (2009). PMID 19404071 (listing read; abstract not opened).
- Ch. 18 — Wu M.P., Ou C.S., Chen S.L., Yen E.Y.T., Rowbotham R. Complications and recommended practices for electrosurgery in laparoscopy. Am J Surg 179(1), 67–73 (2000).
- Ch. 18 — El-Sayed M., Saridogan E. Principles and safe use of electrosurgery in minimally invasive surgery. Gynecol Pelvic Med (2021).
- Ch. 18 — Homma T., Uehara H., Saji H. Factors affecting insulation failure in reusable surgical devices. Sci Rep 13 (2023).
- Ch. 18 — WhaleTeq. High Frequency Dielectric Strength Test, ESU application note (2 October 2017). whaleteq.com/en/app/view15-high-frequency-dielectric-strength-test (test-equipment maker).
- Ch. 18 — MEDTEQ. IEC 60601-2-2 calculator for HF insulation testing, clauses 201.8.8.3.102 and 201.8.8.3.103 (19 April 2023). medteq.net/article/hf-calculator (test-equipment maker, which sells HF insulation testers).
- Ch. 18 — Newton D.W., Odell R.C., Boyle D.R. (Electroscope). Electrosurgical apparatus for laparoscopic and like procedures. US Patent 5,312,401, filed 1991, granted 1994.
- Ch. 18 — Feldman L.S., Fuchshuber P.R., Jones D.B., Robinson T.N., Schwaitzberg S. The SAGES FUSE program: bridging a patient safety gap. Bull Am Coll Surg (September 2014). bulletin.facs.org/2014/09/the-sages-fuse-program-bridging-a-patient-safety-gap/ (by members of the SAGES FUSE task force).
- Ch. 18 — Montero P.N., Robinson T.N., Weaver J.S. et al. Insulation failure in laparoscopic instruments. Surg Endosc 24(2), 462–465 (2010). doi:10.1007/s00464-009-0601-5 (abstract read).
- Ch. 18 — Feldman L.S., Fuchshuber P., Jones D.B., Mischna J., Schwaitzberg S.D., FUSE Task Force. Surgeons don't know what they don't know about the safe use of energy in surgery. Surg Endosc 26(10), 2735–2739 (2012). doi:10.1007/s00464-012-2263-y (abstract read).
- Ch. 19 — Liu Y. et al. Awareness of surgical smoke hazards and enhancement of surgical smoke prevention among the gynecologists. J Cancer 10(12), 2788–2799 (2019).
- Ch. 19 — OSHA. Hospitals eTool: surgical suite, smoke plume. osha.gov
- Ch. 19 — NIOSH. Control of smoke from laser/electric surgical procedures. Hazard Controls HC11, DHHS (NIOSH) Pub. 96-128 (1996).
- Ch. 19 — Bruley M.E. et al. Surgical fires: decreasing incidence relies on continued prevention efforts. Pa Patient Saf Advis 15(2) (2018).
- Ch. 19 — OSHA Technical Manual, Section III, Chapter 6: Laser hazards. osha.gov
- Ch. 19 — Crossley G.H. et al. HRS/ASA expert consensus statement on the perioperative management of patients with implantable defibrillators, pacemakers and arrhythmia monitors. Heart Rhythm 8(7), 1114–1154 (2011).
- Ch. 19 — AORN. Surgical Smoke-Free OR, policy-agenda page listing the states with smoke-evacuation laws (no visible date; read 20 October 2026). aorn.org/get-involved/government-affairs/policy-agenda/surgical-smoke-free-or (professional association).
- Ch. 19 — Tomita Y. et al. Mutagenicity of smoke condensates induced by CO2-laser irradiation and electrocauterization. Mutat Res 89, 145–149 (1981) (not seen; read through Liu et al. 2019).
- Ch. 19 — IEC. IEC 60601-2-22:2019+AMD1:2026 consolidated version (Ed. 4.1), IEC Webstore (publication date 24 April 2026). webstore.iec.ch/en/publication/113384
- Ch. 19 — US FDA. Laser Products – Conformance with IEC 60825-1 Ed. 3 and IEC 60601-2-22 Ed. 3.1 (Laser Notice No. 56). Guidance (issued 8 May 2019; current version 21 February 2023). fda.gov/media/110120/download; 21 CFR 1040.10 and 1040.11.
- Ch. 19 — IEC 60601-1: General requirements for basic safety and essential performance. Ed. 3.2 consolidated (2020); IEC 60601-1-2 Ed. 4.1 (2020); IEC 60601-1-8.
- Ch. 19 — McCormick P.W. Bovie smoke: a perilous plume. AANS Neurosurgeon (2008). aansneurosurgeon.org (secondary; gives the split of six cigarettes for electrosurgery and three for the laser, after Dikes 1999).
- Ch. 19 — Clarke J.R., Bruley M.E. Surgical fires: trends associated with prevention efforts. Pa Patient Saf Advis 9(4), 130–135 (December 2012). patientsafety.pa.gov/ADVISORIES/Pages/201212_130.aspx
- Ch. 19 — ECRI Institute. New clinical guide to surgical fire prevention. Health Devices 38(10), 314–332 (October 2009) (not seen; cited by Bruley et al. 2018 for the 70 %).
- Ch. 19 — APSF. Upper airway management guide for laser airway surgery (1993) (not seen).
- Ch. 19 — Rey J.F., Beilenhoff U., Neumann C.S., Dumonceau J.M. European Society of Gastrointestinal Endoscopy (ESGE) guideline: the use of electrosurgical units. Endoscopy 42(9), 764–772 (2010) (not seen; the publisher's page could not be opened).
- Ch. 19 — Boston Scientific. GreenLight XPS specifications page, en-US (read 21 October 2026): 180 W, 532 nm, NOHD 33.9 m, OD ≥ 6. bostonscientific.com (manufacturer).
- Ch. 19 — Casey V.J. et al. Comparison of surgical smoke generated during electrosurgery with aerosolized particulates from ultrasonic and high-speed cutting. Ann Biomed Eng 49, 560–572 (2021). doi:10.1007/s10439-020-02587-w (read 21 October 2026 at link.springer.com; four conditions, sizes by condition in Table 1; affiliations and funding not in the text retrieved).
- Ch. 19 — Lumenis. Pulse 120H operator's manual UM-10012510 Rev. C, laser safety section: worst-case NOHD 1.6 m for the laser and compatible delivery systems; eyewear OD 3.0. Copy hosted by Agiliti Health, agilitihealth.com (manufacturer).
- Ch. 19 — Thakkar Y., Fernandes A., Mohite S., Phalgune D. Spontaneous combustile pneumoperitoneum. Int Surg J 6(6), 2216–2218 (2019) (secondary: explosive ranges of hydrogen and methane, oxygen below 5 %).
- Ch. 19 — Guglietta E. et al. Heart Rhythm New Zealand consensus statement on the practical management of cardiac implanted electronic devices in the perioperative environment. N Z Med J 137(1590) (2024) (professional society; secondary for the ESGE precautions).
- Ch. 20 — US FDA. Content of Premarket Submissions for Device Software Functions. Guidance (14 June 2023).
- Ch. 20 — ISO 14971:2019, risk management; IEC 62304:2006+A1:2015, software life cycle; IEC 62366-1:2015+A1:2020, usability; IEC 81001-5-1:2021, security; ISO 13485:2016, quality management.
- Ch. 20 — Astute Labs. High-frequency surgical equipment testing (IEC 60601-2-2) (8 May 2026). astute-labs.com (test house; HF leakage limit of 150 mA through 200 Ω).
- Ch. 20 — US FDA. Cybersecurity in Medical Devices: Quality Management System Considerations and Content of Premarket Submissions. Guidance (3 February 2026; supersedes the edition of 27 June 2025). fda.gov/media/119933/download (guidance page read; PDF not opened); cybersecurity FAQs on section 524B.
- Ch. 20 — US FDA. 510(k) summary K213696, Ethicon Megadyne Electrosurgical Generator (applicant Megadyne Medical Products; cleared 21 January 2022). accessdata.fda.gov/cdrh_docs/pdf21/K213696.pdf
- Ch. 20 — Johner Institute (Klessascheck M.). Essential performance of medical devices (11 May 2021). blog.johner-institute.com (consultancy; quotes IEC 60601-1 clauses 3.10 and 3.27; secondary for the standard, whose own text was not reached).
- Ch. 20 — Cheong L. (MEAN WELL USA). The difference between MOPP and MOOP in IEC 60601-1 3rd (19 October 2018). medical.meanwell.com/newsInfo.aspx?c=5&i=635 (power-supply maker); Rigel Medical. A practical guide to IEC 60601, Rev 2 (2021), Appendix A (test-equipment maker); Biomedical Views, IEC 60601-1 leakage currents.
- Ch. 21 — Erbe USA. VIO 3 product page (undated; accessed 20 October 2026). us.erbegroup.com/us-en/products/electrosurgery/vior-3/ (manufacturer); Erbe Elektromedizin. VIO 3 brochure 85160-100 (December 2019; copy hosted by a third party) (manufacturer).
- Ch. 21 — Olympus. ESG-410 Electrosurgical Generator Sockets and Screens Quick Reference Guide, W9227760-01 (January 2021); read on a service distributor's listing page, bioclinicalservices.com.au (manufacturer).
- Ch. 21 — IEC 60601-2-2:2017/AMD1:2023/ISH1:2025, Interpretation Sheet 1 (November 2025), on accessory compatibility, the information each manufacturer provides and CISPR 11 groups; read at webdesk.jsa.or.jp (JSA's host of the IEC sheet).
- Ch. 21 — Cook Medical. ACU-1 and ACU-1-VL Active Cord, instructions for use 14535/0924. ifu.cookmedical.com/data/IFU_PDF/14535_0924.PDF (manufacturer; maximum rated input voltage 5 kVp-p, 2.5 kVp; names no generator).
- Ch. 21 — US FDA. 510(k) summaries K241635, da Vinci E-200 Electrosurgical Generator (cleared 5 August 2024), and K231212, E-200 with Erbe NESSY split pad compatibility (cleared 21 June 2023). accessdata.fda.gov/cdrh_docs/pdf24/K241635.pdf; .../pdf23/K231212.pdf. Applicant Intuitive Surgical.
- Ch. 21 — US FDA. Medical device recall Z-2597-2016, Class 2: Pediatric ThermoGard Dual Dispersive Electrodes, 5–15 kg, item 51-7710, Conmed Corporation (6,640 units; initiated 18 May 2016; FDA-determined cause: device design; terminated 24 May 2018). accessdata.fda.gov/scripts/cdrh/cfdocs/cfres/res.cfm?id=146781
- Ch. 21 — US FDA. 510(k) summary K250725, Hugo RAS System (Covidien, a Medtronic company; product code SCV; substantial-equivalence letter 3 December 2025, corrected letter 4 December 2025). accessdata.fda.gov/cdrh_docs/pdf25/K250725.pdf
- Ch. 21 — US FDA. Design Considerations and Pre-market Submission Recommendations for Interoperable Medical Devices. Guidance (6 September 2017; draft 26 January 2016); read from HHS's guidance repository.
- Ch. 21 — Medtronic. RapidVac smoke evacuator information sheet, US-SE-2000119 (September 2020) and RapidVac smoke evacuator system support page (undated). medtronic.com (manufacturer).
- Ch. 21 — US FDA. Medical device recall Z-0367-2025, Class 2: Stryker Neptune SafeAir Rocker-Switch Smoke Evacuation Pencil (2,620 units; initiated 4 October 2024; expanded by 47 lots, 2 September 2025; cause under investigation by the firm). accessdata.fda.gov/scripts/cdrh/cfdocs/cfRES/res.cfm?id=210435
- Ch. 21 — Schulman P.M., Treggiari M.M., Yanez N.D., Henrikson C.A., Jessel P.M., Dewland T.A., Merkel M.J., Sera V., Harukuni I. et al. Electromagnetic interference with protocolized electrosurgery dispersive electrode positioning in patients with implantable cardioverter defibrillators. Anesthesiology 130(4), 530–540 (2019). doi:10.1097/ALN.0000000000002571 (abstract only; supported by an unrestricted grant from Boston Scientific).
- Ch. 21 — US FDA. Reprocessing Medical Devices in Health Care Settings: Validation Methods and Labeling. Guidance (17 March 2015; Appendix E updated 9 June 2017). fda.gov/media/80265/download
- Ch. 21 — Intuitive. Da Vinci X/Xi System Instrument and Accessory Catalog, BUS00001 v1 US (November 2023; no-pricing edition). intuitive.com (manufacturer; use counts per instrument).
- Ch. 21 — IEC 62353:2014, Medical electrical equipment – Recurrent test and test after repair of medical electrical equipment, Ed. 2.0 (4 September 2014). IEC webstore, webstore.iec.ch/en/publication/6913 (abstract; read 20 October 2026).
- Ch. 21 — Medtronic Canada. FT10 technical training: output checks (undated). medtronic.com/en-ca/healthcare-professionals/services/ft10-technical-training-output-checks.html (manufacturer).
- Ch. 21 — Fluke Biomedical. Electrosurgical Analyzer Primer, 6013240a-en (February 2020). flukebiomedical.com (test-equipment maker, which sells electrosurgical analyzers).
- Ch. 21 — Covidien (Valleylab). Reusable Pencils E2100 and E2100E (Rocker Switch Pencil, Reusable), instructions for use, document PT00071867 (no revision date shown; copy hosted by a distributor at nda.ru; Medtronic's US product page (read 21 October 2026) still lists E2100 with '50 uses'; the voltage rating is not stated on Medtronic's pages, and no current Medtronic IFU was found). Manufacturer document: maximum peak voltage 5,600 V; 50-use life validated with vacuum-assisted steam; compatible generators listed on the cover.
- Ch. 21 — Elite Electronic Engineering. Introduction to CISPR 11 EMC testing (24 May 2022; modified 17 January 2025). elitetest.com/blog/emc-emi-testing/introduction-to-cispr-11-emc-testing (EMC test laboratory; paraphrases CISPR 11's group definitions).
- Ch. 21 — Kirwan Surgical Products. AURA 70 Watt Bipolar Electrosurgical Generators, models 28-2000 and 28-2600, Operating Instructions and Specifications, IFU-70WGEN-5 Rev. M (December 2019). ksp.com/assets/IFU-70WGEN-5-Rev-M.pdf (manufacturer).
- Ch. 21 — AORN. What's New in AORN's Guideline for the Safe Use of Surgical Energy Devices – 2026 Update (4 December 2025). aorn.org/article/whats-new-in-aorns-guideline-for-the-safe-use-of-surgical-energy-devices-2026-update (professional association's summary; the guideline itself not opened).
- Ch. 22 — US FDA. Design Control Guidance for Medical Device Manufacturers (11 March 1997; listed as final, 20 October 2026). fda.gov/media/116573/download
- Ch. 22 — US FDA, CDRH (J. Tartal). Quality Management System Regulation (QMSR): Design and Development, presentation slides quoting ISO 13485:2016 clause 7.3 (undated). fda.gov/media/189041/download
- Ch. 22 — US FDA. Quality Management System Regulation (final rule 2 February 2024; effective 2 February 2026).
- Ch. 22 — ISO. ISO 14971:2019, Medical devices — Application of risk management to medical devices, Ed. 3 (December 2019; confirmed 7 March 2025). iso.org/standard/72704.html (read 20 October 2026).
- Ch. 22 — US FDA, CDRH (J. Tartal). Risk Basics for Medical Devices, presentation slides quoting the ISO 14971:2019 definitions (undated). fda.gov/media/163912/download
- Ch. 22 — BSI. Risk control in the third edition of ISO 14971, excerpt of the white paper "Risk management for medical devices and the new ISO 14971". Compliance Navigator blog (undated; the authors' own views). compliancenavigator.bsigroup.com
- Ch. 22 — US FDA. Applying Human Factors and Usability Engineering to Medical Devices. Guidance (3 February 2016). fda.gov/media/80481/download
- Ch. 22 — Schleer P. Summative evaluation: the following points are essential. Johner Institute blog (12 April 2024). blog.johner-institute.com (consultancy).
- Ch. 22 — US FDA. Ablation catheter correction: Biosense Webster updates use instructions for Varipulse due to high rate of stroke or transient ischemic attack, medical device recalls and early alerts (content current as of 28 February 2025). fda.gov/medical-devices/medical-device-recalls-and-early-alerts/
- Ch. 22 — Johnson & Johnson MedTech. Johnson & Johnson receives FDA approval for updated VARIPULSE platform irrigation rate, news release (14 July 2025). jnjmedtech.com (company statement; the adverse-event rate is from internal customer-reported data and complaints, preliminary).
- Ch. 22 — US FDA. Megadyne recalls MEGA 2000 and MEGA SOFT reusable patient return electrodes due to risk of serious burn, medical device recalls (Class I; correction initiated 1 June 2023, posted 31 July 2024). fda.gov/medical-devices/medical-device-recalls-and-early-alerts/
- Ch. 22 — UL Solutions. Things to know about ISO 14971 3rd edition (28 September 2020). ul.com/news/things-know-about-iso-14971-3rd-edition (certification body).
- Ch. 22 — Medtronic. Medtronic announces FDA clearance for LigaSure RAS Maryland instrument on the Hugo RAS system, news release (16 September 2026). news.medtronic.com (company statement).
- Ch. 23 — US FDA. 510(k) records K261070, LigaSure RAS Maryland jaw sealer, nano-coated (Covidien; cleared 15 September 2026), and K240572, Valleylab SM Smoke Management Pencil with Edge Blade Electrode (cleared 1 July 2024). openFDA, api.fda.gov/device/510k.json?search=product_code:GEI+AND+decision_date:[20240101+TO+20261020]+AND+applicant:covidien (query run 20 October 2026).
- Ch. 23 — US FDA. 510(k) records K242191 (cleared 23 August 2024), K261221 (24 July 2026) and K262877 (16 September 2026), SOLTIVE Pro and Premium SuperPulsed lasers, fibers and accessories (Olympus). openFDA, api.fda.gov/device/510k.json?search=product_code:GEX+AND+decision_date:[20240101+TO+20261020]+AND+applicant:olympus (query run 20 October 2026).
- Ch. 23 — US FDA. 510(k) records K243067, The ETHICON Total Energy System (cleared 17 January 2025), and K254196, HARMONIC 1100 Shears (cleared 17 September 2026). openFDA, api.fda.gov/device/510k.json?search=product_code:GEI+AND+decision_date:[20240101+TO+20261020]+AND+applicant:ethicon (query run 20 October 2026).
- Ch. 23 — US FDA. openFDA 510(k) endpoint, count of records under product code GEI by applicant, decision dates 1 January 2024 to 20 October 2026 (153 records, 98 applicant name forms). api.fda.gov/device/510k.json?search=product_code:GEI+AND+decision_date:[20240101+TO+20261020]&count=applicant.exact (query run 20 October 2026; openFDA states its data are unvalidated).
- Ch. 23 — Medtronic plc. Form 10-K for the fiscal year ended 24 April 2026 (business section). sec.gov/Archives/edgar/data/0001613103/000162828026044354/mdt-20260424.htm (manufacturer).
- Ch. 23 — Johnson & Johnson MedTech. DUALTO Energy System product page (undated; accessed 20 October 2026), with the GEN11 and MEGADYNE generator pages. jnjmedtech.com/en-US/products/surgery/surgical-energy/dualto-energy-system/ (manufacturer).
- Ch. 23 — US FDA. 510(k) record K241138, ArtiSeal Vessel Sealing System, instruments and generator (LivSmed; cleared 15 November 2024). openFDA, api.fda.gov/device/510k.json?search=product_code:GEI+AND+decision_date:[20240101+TO+20261020]+AND+device_name:generator (query run 20 October 2026).
- Ch. 23 — US FDA. 510(k) records K242266, PhotonBlade 3 (Stryker Instruments; cleared 21 October 2024), and K260636, Remanufactured LigaSure XP Maryland Jaw Sealer/Divider (Stryker Sustainability Solutions; cleared 18 June 2026). openFDA, applicant:stryker (query run 20 October 2026).
- Ch. 23 — US FDA. Latest 510(k) records under GEI for BOWA (K193591, ARC 400, cleared 2 April 2020), KLS Martin (K183526, Codman Electrosurgical Generator, cleared 11 April 2019), CONMED (K231732, CleanSeal Advanced Bipolar Vessel Sealer Maryland, 23 October 2023; K230547, PlumePen, 27 April 2023) and Apyx Medical (K230586 and K230272, Renuvion handpieces, 2023). openFDA, sorted by decision date (queries run 20 October 2026).
- Ch. 23 — US FDA. 510(k) record K254056, THUNDERBEAT shears, Grip Type S, front-actuated (Olympus; cleared 1 July 2026). openFDA, api.fda.gov/device/510k.json?search=product_code:GEI+AND+decision_date:[20240101+TO+20261020]+AND+applicant:olympus (query run 20 October 2026).
- Ch. 23 — US FDA. Product classification, product code LFL, instrument, ultrasonic surgical: Unclassified, reason Pre-Amendment (page read 20 October 2026). accessdata.fda.gov/scripts/cdrh/cfdocs/cfpcd/classification.cfm?id=LFL
- Ch. 23 — US FDA. openFDA 510(k) endpoint, count of records under product code GEX by applicant, 1 January 2024 to 20 October 2026. api.fda.gov/device/510k.json?search=product_code:GEX+AND+decision_date:[20240101+TO+20261020]&count=applicant.exact (query run 20 October 2026).
- Ch. 23 — Med-Tech News. Boston Scientific closes acquisition of Lumenis surgical business (2021). med-technews.com (secondary).
- Ch. 23 — US FDA. openFDA 510(k) endpoint, count of records under product code NEY (microwave ablation) by applicant, 2024 to 20 October 2026; K240480, IntelliBlate microwave ablation system (Varian; cleared 19 July 2024). api.fda.gov/device/510k.json?search=product_code:NEY+AND+decision_date:[20240101+TO+20261020] (queries run 20 October 2026).
- Ch. 23 — US FDA. openFDA 510(k) endpoint, count of records under product code GEH (cryosurgical) by applicant, 2024 to 20 October 2026. api.fda.gov/device/510k.json?search=product_code:GEH+AND+decision_date:[20240101+TO+20261020]&count=applicant.exact (query run 20 October 2026).
- Ch. 23 — Johnson & Johnson. Johnson & Johnson announces FDA approval for the Dual Energy THERMOCOOL SMARTTOUCH SF platform, press release (8 July 2026). jnj.com (company statement).
- Ch. 23 — Boston Scientific. 2025 Annual Report. investors.bostonscientific.com (company statement).
- Ch. 23 — HistoSonics. HistoSonics announces $2.25B acquisition by consortium of top-tier investors, news release via Business Wire (7 August 2025). businesswire.com/news/home/20250807749442 (company statement).
- Ch. 23 — HistoSonics. Financing announcement at a US$3.75 billion valuation, news release via Business Wire (22 June 2026). businesswire.com/news/home/20260622435314 (company statement).
- Ch. 23 — Insightec. Insightec announces FDA approval of staged bilateral focused ultrasound treatment for Parkinson's disease, news release (8 July 2025). insightec.com (company statement).
- Ch. 23 — Profound Medical Corp. Form 8-K, exhibit 99.1 (12 January 2026). ir.profoundmedical.com (company statement).
- Ch. 23 — EDAP TMS S.A. Form 10-K for the fiscal year ended 31 December 2025. sec.gov/Archives/edgar/data/1041934/000110465926034528/edap-20251231x10k.htm (manufacturer).
- Ch. 23 — Intuitive Surgical. Form 8-K, exhibit 99.1, on FDA clearance of da Vinci 5 (14 March 2024). sec.gov/Archives/edgar/data/1035267/000103526724000125/q124ex-991dv5fdaclearancex.htm (company statement).
- Ch. 23 — Cambridge Innovation Capital. CMR Surgical secures 510(k) clearance of its Versius Plus robotic surgical system (16 December 2025). cic.vc (investor page reproducing company news; secondary).
- Ch. 23 — Stryker. Stryker and CMR Surgical announce strategic alliance to advance robotic-assisted surgery, news release (12 October 2026). stryker.com (company statement).
- Ch. 23 — SS Innovations International. SS Innovations submits response to FDA request for additional information, news release via Stockhouse (13 October 2026). stockhouse.com (company statement).
- Ch. 23 — US FDA. 510(k) records with an Indian applicant address: K231718, Holmium Medical Laser (cleared 18 August 2023), and K242987, Thulium Fiber Laser FiberLAZE+ (cleared 17 December 2024), Allengers Global Healthcare; K153055, ACE electrosurgical resection and vaporization electrodes (Ace Medical Devices; cleared 11 July 2016). openFDA, country_code:IN with product codes GEX and GEI (queries run 20 October 2026).
- Ch. 23 — Department of Pharmaceuticals, Government of India. List of applicants approved under the Production Linked Incentive scheme for promoting domestic manufacturing of medical devices, as on 06.11.2023. pharma-dept.gov.in
- Ch. 23 — Medtronic. Medtronic launches advance Valleylab FT10 energy platform in India, news release (Delhi, 30 September 2025). news.medtronic.com (company statement).
- Ch. 23 — BioSpectrum India. Biorad Medisys unveils Quantino 20W laser system for kidney stones management (9 June 2026). biospectrumindia.com (secondary; company claims).
- Ch. 23 — US FDA. 510(k) record K241621, monopolar and bipolar adapters (Intuitive Surgical; cleared 2 August 2024). openFDA, product_code:GEI, applicant:intuitive (query run 20 October 2026).
- Ch. 23 — US FDA. 510(k) record K252487, POWERSEAL Open Extended Jaw Sealer and Divider, Double-Action (Gyrus ACMI, Inc., Westborough MA, an Olympus company since 2008; cleared 2 March 2026). accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMN/pmn.cfm?ID=K252487; Olympus America, Surgical Technologies page (accessed 21 October 2026) (manufacturer).
- Ch. 23 — US FDA. De Novo DEN230078, Versius Surgical System (CMR Surgical; received 21 November 2023, granted 11 October 2024; 21 CFR 878.4964, product code SCV), and 510(k) record K252111, Versius Surgical System (Versius Plus) (cleared 16 December 2025). accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/denovo.cfm?id=DEN230078; openFDA, product_code:SCV (query run 21 October 2026).
- Ch. 23 — US FDA. 510(k) records under product code LFL decided 1 January 2024 to 21 October 2026: K233036 (Hunan Handlike), K240493 and K251162 (Integra CUSA Clarity), K241700 (Trice Medical Tenex), K242894 (Shanghai Saints Sages SanAgile), K243930 and K262244 (Stryker Sonopet iQ; K262244, Sonopet iQ Series III, cleared 8 October 2026), K261448 (Misonix neXus). openFDA, api.fda.gov/device/510k.json?search=product_code:LFL+AND+decision_date:[20240101+TO+20261021] (query run 21 October 2026).
- Ch. 23 — US FDA. 510(k) records K251759, Flexiva Pulse laser fibers (cleared 9 July 2025), and K260100, MOSES Raydar holmium laser system (cleared 18 May 2026), Boston Scientific; product code GEX, 21 CFR 878.4810. openFDA, product_code:GEX, applicant:boston (query run 20 October 2026); Boston Scientific, MOSES Raydar Holmium Laser System product page (Ho:YAG, 2.1 µm) (manufacturer).
- Ch. 23 — US FDA. 510(k) record K250809, Stellar M22 (Lumenis Be, Yokneam; product code GEX, with the light-based aesthetic codes ONF and ONG); 510kdatabase.net, list of Lumenis Be clearances 2024–2026 (Stellar M22, FoLix, MILAN System) (secondary); Lumenis and BPEA, release of 2021 via BioSpace (company statement; the remaining company keeps the aesthetics and vision businesses).
- Ch. 23 — Medtronic. Affera momentum continues as Medtronic announces CE Mark in Europe and U.S. IDE first cases for Sphere-360 PFA catheter to treat paroxysmal atrial fibrillation, news release (23 January 2026); and Medtronic announces key Affera clinical study milestones for Sphere-9 and Sphere-360 catheters, news release (25 April 2026). news.medtronic.com (company statements).
- Ch. 23 — CDSCO. Medical Devices Rules, 2017; risk classification lists for operation theatre (13 September 2021) and dermatology and plastic surgery devices (26 July 2021) (lists not seen; their entries read in consultancies' accounts).
- Ch. 23 — Morulaa. India risk classification for operation theatre as medical device (undated; read 20 October 2026). morulaa.com/knowledge-hub/regulations/india/cdsco-medical-device-classification-operation-theater (consultancy; secondary).
- Ch. 24 — Johnson & Johnson MedTech. VARIPULSE release, Yokohama, via Kyodo PR Wire (3 November 2025). kyodonewsprwire.jp/release/202511038312 (company statement).
- Ch. 24 — Johnson & Johnson. Johnson & Johnson advances pulsed field ablation portfolio with the launch of VARIPULSE Pro in Europe (7 April 2026); Johnson & Johnson launches landmark head-to-head pulsed field ablation trial in persistent atrial fibrillation (PERSIGMA, 24 April 2026). jnj.com (company statements).
- Ch. 24 — Medtronic. Medtronic submits 510(k) filings to expand Hugo robotic-assisted surgery system into general and gynecologic specialties in the United States, news release (3 June 2026). news.medtronic.com (company statement).
- Ch. 24 — US FDA. Medical device recall, Class 2: Olympus ESG-410 electrosurgical generator, E0662 system error with reboots (155 units; initiated 25 July 2025; posted 19 August 2025). accessdata.fda.gov/scripts/cdrh/cfdocs/cfres/res.cfm?id=214934
- Ch. 24 — Intuitive. da Vinci 5 page, MAT01676 V4 US (April 2025). intuitive.com/en-us/products-and-services/da-vinci/5 (manufacturer; the force-feedback figure is from Intuitive's preclinical trials).
- Ch. 24 — Medtronic. Hugo RAS system page, US-RP-2500120 (April 2026). medtronic.com/covidien/en-us/robotic-assisted-surgery/hugo-ras-system.html (manufacturer).
- Ch. 25 — US FDA. Recognized Consensus Standards, detail pages: 6-521, IEC 60601-2-2 Ed. 6.1 2023-02, corrected version 2025-11 (date of entry 25 May 2026; declarations to 6-389, Ed. 6.0, accepted until 1 July 2029); 12-268, IEC 60601-2-22 Ed. 3.1 (superseded by 12-356, Ed. 4.0; accepted until 4 July 2027); 12-121, IEC 61847:1998; 12-273, IEC 60825-1 Ed. 2.0 2007; 12-281, IEC 60601-2-62 Ed. 1.0 2013; 19-36, IEC 60601-1-2 Ed. 4.1 2020 (partial). accessdata.fda.gov/scripts/cdrh/cfdocs/cfstandards/ (read 20 October 2026).
- Ch. 25 — US FDA. Food and Drug Administration Modernization Act of 1997: Modifications to the List of Recognized Standards, Recognition List Number 066. Federal Register, document 2026-17229 (24 August 2026). federalregister.gov
- Ch. 25 — Ehninger M.D. et al. (MegaDyne Medical Products). Universal self-limiting electrosurgical return electrode. US Patent 12,357,368 B2 (filed 10 October 2022, granted 15 July 2025); cited for the 6 °C limit on skin temperature rise under a return electrode, which it attributes to the IEC (manufacturer; a secondary source for the standard).
- Ch. 25 — Institute for Standardization of Montenegro (ISME). Mirror of the IEC project record for IEC 62304 ED2, Health software – Software life cycle processes, project 122433 (stage 30.20 from 7 August 2026). isme.me/en/project/show/iec:proj:122433 (secondary: a national body's copy of the IEC record; the IEC dashboard could not be opened).
- Ch. 25 — BSI Standards Development. Project TS 62A-1727 ED1, Requirements for assessment of electromagnetic exposure hazards, stating that the merged CDV of IEC 60601-1 Ed. 4 is anticipated for 2029. standardsdevelopment.bsigroup.com/projects/9026-13201 (secondary for the IEC plan).
- Ch. 25 — IEC Webstore. IEC 60601-1:2005+AMD1:2012+AMD2:2020 CSV, Ed. 3.2 (20 August 2020), webstore.iec.ch/en/publication/67497; IEC 60601-1-6:2010+AMD1:2013+AMD2:2020 CSV, Ed. 3.2 (22 July 2020), webstore.iec.ch/en/publication/67381 (read 21 October 2026).
- Ch. 25 — IEC 60601-2-2 Ed. 6.1 (2023-02, corrected version 2025-11), IEC preview pages (contents; Table 201.103, 'Test currents by weight range'; introduction listing the changes of Amendment 1). Preview PDF as hosted by technickenormy.cz; catalogue page webstore.iec.ch/en/publication/83195 (read 21 October 2026).
- Ch. 26 — US FDA. General and Plastic Surgery Devices Panel, executive summary on ultrasonic surgical instruments (26–27 October 2022).
- Ch. 26 — US FDA. Deciding When to Submit a 510(k) for a Change to an Existing Device. Guidance (October 2017).
- Ch. 26 — US FDA. Deciding When to Submit a 510(k) for a Software Change to an Existing Device. Guidance (October 2017).
- Ch. 26 — ISO. Certification (read 21 October 2026). iso.org/certification.html
- Ch. 26 — US FDA. A history of medical device regulation and oversight in the United States (content current as of 21 August 2023); Tarver M. Fifty years of the Medical Device Amendments. FDA Voices (28 May 2026). fda.gov
- Ch. 26 — US Code of Federal Regulations, 21 CFR 878.4400 and 878.4810 (accessed 2026), and 21 CFR 820.7 and 820.10 (QMSR; eCFR, up to date as of 16 October 2026); FDA product classification database, codes GEI, GEX and LFL, and NEY, NRZ, PLP and QZI (pages read 21 October 2026).
- Ch. 27 — Medical Device Coordination Group. MDCG 2021-24, Guidance on classification of medical devices (2021; Rev. 1, April 2026).
- Ch. 27 — Ireland, Health Products Regulatory Authority (HPRA). Mandatory use of EUDAMED modules for medical devices (23 March 2026). hpra.ie (national competent authority).
- Ch. 27 — European Commission. Proposal for a regulation to simplify the rules on medical and in vitro diagnostic devices, COM(2025) 1023 (16 December 2025).
- Ch. 27 — Council of the European Union, General Secretariat. Progress report on the proposal amending Regulations (EU) 2017/745 and 2017/746, interinstitutional file 2025/0404(COD), document 9801/26 (5 June 2026). data.consilium.europa.eu
- Ch. 27 — CDSCO. Categories of medical devices for risk-based classification, draft for stakeholder comments, File No. MED-16014(12)/1/2024-eoffice (6 January 2025). cdsco.gov.in
- Ch. 27 — Regulation (EU) 2017/745 on medical devices, Article 52(4), Annex VIII Rule 9 and Annex IX sections 2.4, 4 and 4.10, read as adopted in the National Archives' copy, legislation.gov.uk/eur/2017/745 (EUR-Lex's pages could not be read past Article 16); Regulation (EU) 2023/607, eur-lex.europa.eu/eli/reg/2023/607/oj/eng.
- Ch. 27 — Czech Metrology Institute (notified body). Conformity assessment procedure under the MDR, version 24-001. cmi.gov.cz (notified body's flowchart).
- Ch. 27 — IEC 60601-2-2:1982, Medical electrical equipment – Part 2: Particular requirements for the safety of high-frequency surgical equipment, Edition 1.0 (published 1 January 1982; withdrawn 31 October 1991). webstore.iec.ch/en/publication/16836
- Ch. 27 — National Academies (Institute of Medicine). Medical Devices and the Public's Health: The FDA 510(k) Clearance Process at 35 Years (2011), chapter 2. nationalacademies.org/read/13150/chapter/4
- Ch. 27 — CDSCO. Frequently Asked Questions on Medical Devices Rules, 2017, Doc. No. CDSCO/FAQ/MD/01/2024. cdsco.gov.in (FAQs 25–26, 27, 30, 55, 69, 80, 89, 97, 136, 148).
- Ch. 27 — CDSCO. Risk classification list of medical devices pertaining to Oncology, File No. MED-16014(12)/1/2024-eoffice (notice of 2 January 2026). cdsco.gov.in (read 21 October 2026).
- Ch. 27 — CDSCO. Medical Device and Diagnostics page, risk classification notices including the Oncology list (2 January 2026). cdsco.gov.in/opencms/opencms/en/Medical-Device-Diagnostics/ (read 20 October 2026).
- Ch. 27 — CliniExperts. DCGI classifies 55 medical devices pertaining to dermatology and plastic surgery (11 October 2021). cliniexperts.com (consultancy; secondary).
- Ch. 27 — European Parliament, Legislative Observatory. Procedure file 2025/0404(COD) (read 21 October 2026). oeil.europarl.europa.eu
Take the PDF with you.
Reading online needs no sign-up. For the PDF edition, laid out for print and offline reading, tell us where to send it and we will email you.
Check your inbox.
The download link is on its way to your email. If it has not arrived in a few minutes, check your spam folder or write to info@unplex.tech.
Preview only — nothing was sent
Let's Talk