2

Cryoablation: Mechanism of Actionand Devices

Christos S. Georgiades and Jonathan K. Marx

Though initially introduced in the 19th century, cryoablation entered mainstream clinical medicine in the mid-1960s.¹ Because of technical limitations, however (i.e., large-bore probes to accommodate the large gas flow required and to facilitate adequate heat exchange), its applications were limited in the operative setting. In the last few years, as technology spawned the introduction of thinner cryoprobes (down to 14-gauge), image-guided, percutaneous cryoablations have carved out an ever-expanding niche. Currently, cryoablation is extensively used in the treatment of prostate cancer and renal cancer (either during open surgery or laparoscopically). Image-guided percutaneous cryoablation is mostly used for the treatment of small renal cancers and for the palliation of painful bone lesions. To a lesser extent, it is also used for the treatment of primary and secondary liver and chest neoplasms.

The technical goal of cryoablation is to reduce the temperature of the target tissue as well as an appropriate surrounding margin (equivalent to the surgical margin) to below lethal levels, without damaging vital nearby organs. For mammalian tissue, lethal temperature is below approximately −20° to −25°C (or −4° to −13°F). Cytotoxicity is mediated via various mechanisms, including ischemia, denaturing of protein structures, and the breakdown of cellular and intracellular membranes due to the formation of ice crystals. Some authors believe that the extracellular scaffolding structure is relatively preserved during cryoablation. This theoretically allows for the cellular repopulation of the tissue, which may explain why cryoablation is more forgiving, compared with radiofrequency ablation (RFA), in cases of nontarget ablation.

♦ The Physics of Cryoablation

The objective of cryoablation (kryo = “cold” in Greek; and ablatus = “to carry away” in Latin) is to reduce the temperature of the target tissue to below the lethal −20°C. This requires a continuous removal of energy from the entire target tissue. There are three ways that thermal energy may be transferred from one volume to another: (1) by conduction, the transfer of energy (or heat) from an object to a another at a lower temperature by virtue of physical conduct (direct molecular kinetic energy exchanges); (2) by convection, the transfer of energy (or heat) from an object to another at a lower temperature, with one of the objects being a flowing fluid (thus providing a continuously refreshed heat sink or heat source); and (3) by radiation, which removes radiant energy from all bodies above absolute zero temperature. No significant radiation losses are observed in our case; therefore, we can concentrate on conduction and convection (Fig. 2.1).

Conduction is a steady-state transfer because (in our case) it depends only on the temperature difference, which is relatively stable (i.e., tissue temperature [37°C] versus probe temperature [approximately −140°C]). On the other hand, convection depends on the flow rate of the blood through or adjacent to the volume of interest. Therefore, large-diameter arteries or very vascular tissues, to a lesser degree, will limit the extent of cryoablation. The reverse of this phenomenon (cooling of tissue by nearby flowing blood) has been termed “heat sink” in the case of thermal ablations (such as RFA or microwave ablation).

One advantage of cryoablation over RFA and microwave ablation is that the tissue energy transfer characteristics do not appreciably change with temperature. In contrast, in the case of RFA, tissue desiccates at high temperature and no further transfer of energy is possible. Irrespective of the tissue temperature, further temperature reduction is always theoretically possible with cryoablation.

The removal of energy (cooling) from the tissue surrounding the cryoprobe is effected by the Joule-Thomson law.² During gas expansion two opposing mechanisms take effect. First, as intermolecular distances increase with cooling, the potential energy stored between molecules increases. Because total energy must be conserved, this translates into lower molecular kinetic energy and therefore lower temperature. Second, as intermolecular distance increases, however, fewer molecular collisions result in a relative decrease in potential energy stored between molecules. Again, because total energy must be conserved, this results in greater molecular kinetic energy and therefore greater temperature. Which of the two competing mechanisms dominates depends on the so-called inversion temperature of the gas, which is an intrinsic property. For argon, nitrogen, or oxygen, the inversion temperature factor is greater than room (or ambient) temperature and the first mechanism dominates. For helium and hydrogen, the inversion temperature is less than room temperature. Therefore, expansion of compressed argon, nitrogen, or oxygen causes cooling, whereas expansion of helium or hydrogen causes warming.

The gas is supplied by a gas tank to the probe via a pressure regulator (Fig. 2.2). Gas expansion occurs within the dual chambered cryoprobe (Fig. 2.3). The gas, once it expands, cools, absorbs thermal energy from the surrounding tissues, and is siphoned back through the tubing and eventually released into the atmosphere. The diameter of the “ice ball” depends on the rate of energy transfer, which in turn depends on the flow rate of the cooling gas through the probe. The length of the ice ball mostly depends on the uninsulated length of the probe (Fig. 2.4).

It is worth repeating that the lethal temperature for mammalian tissue is −20°C or colder. This does not correspond to the visible margin of the ice ball as seen under computed tomography (CT), ultrasound (US), or magnetic resonance (MR) guidance, which is at 0°C. Studies have shown that the −20°C isotherm is at least 5 mm inside the 0°C isotherm.³

♦ Ablation Systems

There are currently two U.S. Food and Drug Administration– approved cryoablation systems available in the United States, which at the time of writing of this chapter were in advanced merger discussions. Both systems, Endocare and Galil, make use of the above basic physical properties of the gases, and both utilize argon for freezing and helium for thawing. This may change in the near future as technologic advancements may allow the reintroduction of a single gas cooling-thawing system using nitrogen.

Fig. 2.2 Diagram of the gas supply during cryoablation. The gas is stored in a high-pressure cylinder. For ultrahigh purity medical gases, the pressure is usually 6000 psi. At this pressure one argon tank provides enough gas for approximately 8 to 10 single-probe cryoablations. A full helium tank will suffice for 10 to 15 single-probe thawings. The latter is not absolutely necessary. Active thawing reduces the waiting time to 2 to 4 minutes from that of 6 to 12 minutes for passive thawing. The gas tanks are connected to the regulator via high-pressure cables. The operator uses the regulator to start or stop cryoablation once the probes are in the proper position. Rarely, when the operator needs to limit the size of the ice ball, the regulator can decrease the rate of gas flow to the probe, which decreases the rate of heat removal from the target tissue and consequently the final size of the ice ball. The pressurized gas is transferred to the probe, where it is allowed to decompress to 1 atmosphere within the probe. The Joule-Thomson effect dictates that, for certain gases, including nitrogen, argon, and oxygen, this rapid decompression results in cooling of the gas. The cooled gas removes energy (Q_out) from the tissue around the probe and circulates back toward the regulator. Finally the gas is exhausted into the room (inert gas) at 1 atmosphere and at room temperature.

Fig. 2.3 Schematic of the cryoprobe and its function. The cooling gas (argon in this case) is carried to the probe via a high-pressure cable at room temperature and under high pressure. Near the tip of the two-chambered probe, the gas is released into the probe, where its pressure equilibrates with ambient pressure (approximately 1 atmosphere). This rapid decompression results in the supercooling of the gas (Joule-Thomson effect) and gradually of the probe and the surrounding tissues, forming the ice ball. The margin of the ice ball as noted on CT or US represents the 0°C isotherm. This temperature is not lethal to mammalian tissue, and therefore the ice ball should not be thought of as the ablation region. Rather, at least a 5-mm ablation margin is required to reach the lethal −20°C and ensure tissue death.

Fig. 2.4 Even though the temperature reading of the probe can reach −150° to −160°C, the distance between the tumor margin (dashed blue circle) and the edge of the ice ball (red arrows) is the crucial factor that determines the efficacy of ablation. The temperature rises quickly as a function of the distance from the probe, and reaches 0°C at the edge of the ice ball. The lethal temperature for mammalian tissue is −20°C or colder. Therefore, an ablation margin is required between the target tumor and the edge of the visible ice ball to ensure complete ablation. A 5-mm margin is thought to be adequate to ensure that the target tumor reaches the lethal temperature of −20°C. The solid black line piercing the target lesion (dashed blue circle) is the “ghost” of the removed probe, which remains filled with air as the surrounding ice ball prevents it from collapsing.

With the Endocare (Irvine, CA) system, there are four cryoprobe types available for use, differing in the shape and size of the generated ice ball (Fig. 2.5). Probe selection depends on the size and shape of the target lesion. For example, a 1-cm lesion in a difficult-to-access location is best approached with the Perc-15 probe. In general the Perc-24 is the workhorse probe for this system, as it yields the largest ablation zone. The respective isotherms for each probe are graphically shown in Fig. 2.6. Endocare also provides a percutaneously inserted temperature sensor (Fig. 2.5). This may be useful in rare instances when the temperature near a nontarget organ needs to be monitored.

The Endocare regulator is shown in Fig. 2.7. The regulator is composed of gas inlet hookups (argon and helium in); gas outlet hookups (argon, helium out); the internal regulator design; the control panel; and the screen, which can display the readings from up to eight cryoprobes and eight temperature sensors.

Galil

The Galil (Yokneam, Israel) Seednet and Presice regulators (which have the capacity to utilize and monitor up to five cryoprobes simultaneously) are shown in Fig. 2.8. The design of the cryoprobes differs slightly from that of Endocare’s in that they are a thinner (17-gauge) (Fig. 2.9). Four cryoprobes are available—the IceBulb, the IceRod, the IceSphere, and the IceSeed—which differ in the shape and size of the ice ball generated. The Galil cryoprobe isotherms are shown in Fig. 2.10 and sample ice balls are depicted in Fig. 2.11. As with Endocare, the choice of the number and type of cryoprobes depends on the size and shape of the target lesion. The operator must “sculpt” an ice ball by placing the appropriate number and type of probes at such angles with respect to each other so that the resulting ice ball covers the entire target lesion as well as a 5-mm margin, at least. For detailed temperature monitoring, Galil provides specific temperature sensors (Fig. 2.12), as the cryoprobes themselves do not have incorporated thermocouples. These sensors can be placed between the target lesion and nontarget organ to monitor the temperature near the nontarget organ, or in or near the target lesion to monitor the effectiveness of cryoablation. They are capable of providing real-time temperature measurements at multiple points, that is, at 5, 15, 25, and 35 mm from the tip.

Fig. 2.5 (A) The Endocare cryoprobes. The right-angle design facilitates probe insertion within the confines of a CT gantry and is especially useful for obese patients. Because of design limitations, the internal thermocouple is contained within the distal shaft of the cryo-probe (arrow). As a result, the actual temperature at the tip of the probe is approximately 10° to 12°C colder than the display reading. It is necessary to allow the probe to thaw to above +10°C; otherwise, the still frozen tip sticks to the tissue and makes insertion or removal difficult. (B) Temperature sensors can be used if necessary to monitor the temperature near a critical nontarget structure. They measure a single point temperature at the tip of the sensor.

Ablation Protocols

The ablation protocols have been developed mostly based on empirical observations. Deviations from the protocols should be discouraged unless patient safety is compromised in not doing so. For example, an expanding ice ball that threatens to invade the nearby bowel may precipitate either early termination or reduction in the power of the closest probe, if more than one is used. Cell death is precipitated not only by the low temperatures but also by the effects of rapid cooling, thawing, and recooling, which result in ice crystal formation and cell membrane disruptions. Cooling for longer than the protocol recommends does not appreciably increase the size of the ice ball as the system approaches a plateau along the heat exchange curve as explained in the physics section above. Similarly, applying more than the recommended ablation cycles does not appreciably increase the cytotoxic effect of the ice ball.

Fig. 2.6 Isotherms for the Endocare probes. The margin of the ice ball as seen on CT or US represents the 0°C isotherm. This can be used as a safety factor to minimize the risk of nontarget ablation; however, it is not lethal. The lethal isotherm is the one at −20°C, and it varies from probe to probe. For example, the lethal isotherm for the Perc-24 is an oval/teardrop-shaped volume with a 44-mm long diameter and a 24-mm short diameter. The thicker the probe, the faster the cooling gas can circulate, and thus the larger the effective isotherm. A lesion with a diameter of 1.5 cm theoretically can be ablated using only one Perc-24 cryoprobe. In reality, however, this is unlikely, because such an outcome presupposes the insertion of the single probe perfectly within the center of perfectly spherical mass. It is prudent to be liberal with the use of cryoprobes and ensure adequate target mass and margin ablation, rather than risk having to re-treat the patient in the near future.

The protocol is set irrespective of the location, type, and number of probes or the location, size or kind, of tumor. It consists of a 10-minute freeze, followed by an 8-minute thaw and another 10-minute freeze, of all probes simultaneously. Some authors support the notion that the thaw cycle should be active (helium circulates) and not passive (no gas circulation) because faster thawing may improve the cytotoxic effect of cryoablation. In theory this is correct; however, prior to the second freeze cycle the location of the probes must be verified as they may have moved during active thawing.

Cryoablation theoretically can be used in any part of the body; however, its specific risks and benefits must be weighed against those of other ablation modalities. Lung cryoablation is possible and occasionally necessary (Fig. 2.13); however, one has to contend with a slightly higher risk of pulmonary hemorrhage compared with doing RFA. This can be clinically significant in patients with severe chronic obstructive pulmonary disease, who represent a good portion of the patients with primary lung cancer referred for ablation. In general, cryoablation for peripheral, pleura-based lesions is safer than for central ones. In addition, cryoablation for pleura-based lesions is essentially painless, in contrast to RFA. Primary and secondary liver cancer (Fig. 2.14) can be treated with cryoablation with probably similar efficacy and safety to that of RFA, though there is vastly more published data on the latter. The efficacy of cryoablation is affected by nearby large vessels, a phenomenon whose RFA counterpart is termed “heat-sink” effect. Blood vessels larger than 3 mm in diameter are known to affect the geometry of the ice ball. In such cases, additional probes, or even balloon occlusion of a hepatic vein (if this is the culprit), may be necessary. The most commonly treated tumor with cryoablation is renal cell carcinoma (Fig. 2.15), and it is also the one with the most published efficacy and safety data. The use of cryoablation for metastatic disease is not common, and its use should be limited to three indications: (1) in cases where tumor cytore-duction has been shown to improve survival (i.e., oligometastatic colon, breast, or renal cancer); (2) for the palliation of painful metastatic foci (Figs. 2.16, 2.17, 2.18, 2.19, and 2.20); or (3) to treat a metastasis whose further growth is expected to result in significant symptoms.

Fig. 2.14 An axial CT image during cryoablation of a large liver lesion. This cryoablation was performed as part of an experimental protocol during which only part of the mass is targeted, which explains why the cryoprobes were placed far apart. The result is the formation of two distinct ice balls (arrows) that do not coalesce. When more than one cryoprobe is used, the distance (D) between them should be less than 2 cm to ensure their ablation zones coalesce and that no viable tumor is left between them. Additionally, to maximize the size of the resulting ice ball, the cryoprobes should be placed as parallel to each other as possible.

Fig. 2.15 (A) An axial CT image shows an exophytic left renal mass (arrowhead). (B) During cryoablation, the probe(s) are placed so that the ice ball (arrowheads) covers the mass and a >5-mm margin. (C) On the 3-month contrast-enhanced CT, the mass appears completely devascularized and the ablation zone in the perinephric fat is demarcated by a hypervascular rim (arrowhead).