Thermal Ablation in the Liver: Heat versus Cold—What Is the Role of Cryoablation?

Abstract

Cryoablation is commonly used in the kidney, lung, breast, and soft tissue, but is an uncommon choice in the liver where radiofrequency ablation (RFA) and microwave ablation (MWA) predominate. This is in part for historical reasons due to serious complications that occurred with open hepatic cryoablation using early technology. More current technology combined with image-guided percutaneous approaches has ameliorated these issues and allowed cryoablation to become a safe and effective thermal ablation modality for treating liver tumors. Cryoablation has several advantages over RFA and MWA including the ability to visualize the ice ball, minimal procedural pain, and strong immunomodulatory effects. This article will review the current literature on cryoablation of primary and secondary liver tumors, with a focus on efficacy, safety, and immunogenic potential. Clinical scenarios when it may be more beneficial to use cryoablation over heat-based ablation in the liver, as well as directions for future research, will also be discussed.

Radiofrequency ablation (RFA), microwave ablation (MWA), and cryoablation are the most widely used thermal ablation modalities for treating tumors in the liver, kidney, lung, breast, and soft tissue. Liver tumors are predominantly treated using heat-based ablation, and in the United States MWA has largely replaced RFA, with cold-based techniques via cryoablation used much less commonly. ^{1 2}

Early cryoablation techniques in the liver used an open surgical method in conjunction with large probes and nitrogen cryogen. ^{3 4 5} A retrospective analysis of these early procedures in liver tumors raised concerns about cryoshock and hemorrhage that influenced the subsequent era of thermal ablation procedures, and cryoablation was relegated as a riskier alternative to RFA. ^{6 7} Today, using smaller argon-helium probes, percutaneous technique, and cautery options, cryoablation appears to be a suitably safe and effective alternative to RFA and MWA. ⁸ In cirrhotic patients, the use of heat-based ablation has historically been preferred in the liver due to the perceived lower bleeding rates in this coagulopathic population. However, with the aforementioned advances in cryoablation technology, the hemorrhage gap is closing between the two mechanisms. ⁸

Cryoablation has several advantages over RFA and MWA. First, the ice ball is easily visualized intraprocedurally on ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) during the procedure, allowing for easy monitoring of the size and location of the ablation zone. ^{9 10 11} Second, the edge of the ice ball is at a non-lethal tissue temperature, and combined with easy visualization of the ice ball, it is postulated that cryoablation may be safer next to critical structures. ^{9 10 12} Third, cryoablation causes relatively little pain compared with RFA and MWA, making it easier to perform under local anesthesia (LA) or conscious sedation and to avoid the need for general anesthesia (GA). ^{9 10 12} Finally, cryoablation appears to induce more of an immunogenic response in tumors compared with heat-based ablation, ^{13 14 15 16} paving the way for potential combination treatment with immunotherapy to improve outcomes. The disadvantages of cryoablation include the need for multiple cryoprobes in each tumor. Furthermore, there remains a risk of cryoshock due to release of cytokines, and thrombocytopenia due to local platelet trapping, particularly when treating large tumors. ^{9 17}

This article will review the current literature on cryoablation of liver primary and secondary tumors regarding efficacy, safety, comparison to RFA and MWA, and immunomodulatory effects. Clinical scenarios when it may be more beneficial to use cryoablation over heat-based ablation in the liver, and future directions for research, will also be discussed.

Overall Efficacy and Safety of Cryoablation in the Liver

A few recent large single-arm studies have evaluated the overall efficacy and safety of cryoablation for primary and secondary liver malignancies. In a large retrospective series by Littrup et al, ¹⁷ 443 liver tumors (370 metastases and 73 HCC) were treated with cryoablation in 212 patients, using either mild or deep sedation depending on anticipated procedural complexity. Average tumor size was 2.8 cm ± 1.4, with an average of 4.5 ± 2.2 cryoprobes used per tumor. At a mean follow-up time of 1.8 years, local recurrence (LR) rates were 5.5, 11.1, and 9.4% for the HCC, metastatic colorectal cancer (mCRC), and non-mCRC groups, respectively. The overall major complication rate was 5.8%, mostly hematologic derangements, with three deaths from cryoshock. The authors noted that the major complication rate decreased to 3.2% when only tumors < 4 cm and patients with platelet count > 100,000 were included. Another large retrospective study by Glazer et al ¹⁰ assessed 299 liver tumors (243 metastases and 56 HCC or cholangiocarcinoma) in 186 patients treated with cryoablation using conscious sedation or GA. Mean tumor diameter was 2.5 cm (range: 0.3–7.8 cm). Mean follow-up time was 2.5 years, and local tumor progression (LTP) was seen in 18.0% of tumors < 4.0 cm and 63.3% of tumors ≥ 4 cm. The grade 3–5 adverse event rates were 8.7% in tumors < 4 cm and 19.5% in those ≥ 4 cm ( p  = 0.04). Platelet transfusion due to post-ablation thrombocytopenia was administered in 9.2% of tumors < 4 cm versus 22% of those ≥ 4 cm ( p  = 0.02).

In Pusceddu et al's ¹⁸ retrospective series, where 64 tumors (50 liver metastases and 14 HCC) were treated with cryoablation under conscious sedation in 49 patients, mean tumor diameter was 2.15 cm (range: 0.5–5 cm) and mean follow-up was 19.8 months. They found LTP at the treatment site in 9% of lesions, and no major complications. Finally, in a study by Kim et al ¹¹ evaluating cryoablation in 45 patients for early or very early stage HCC per Barcelona Clinic Liver cancer (BCLC) guidelines, performed with intramuscular pethidine and LA only, mean tumor size was 1.8 cm ± 0.5 and patients were followed up for a mean of 28.1 ± 15.6 months. LTP at the treatment site occurred in 11.1%, and the LTP-free survival rates at 1 and 2 years were 93.3 and 88.9%, respectively. LTP rates were found to be significantly higher in patients with an ablation margin < 5 mm. No major complications occurred in the cohort.

Cryoablation Adjacent to Critical Structures

Several advantages of cryoablation make it particularly suitable for treating liver tumors adjacent to critical structures such as the gallbladder, central bile ducts, major blood vessels, bowel, diaphragm, or heart. Namely, the ability to visualize and monitor the location and size of the ice ball intraprocedurally, and the non-lethal ice temperature at the periphery of the ice ball. Several studies have evaluated the safety and efficacy of hepatic cryoablation adjacent to such structures without the use of hydrodissection or pneumodissection, and will be discussed here.

Gallbladder

In a retrospective review of cryoablation of tumors adjacent to the gallbladder by Fairchild et al, ¹⁹ 19 tumors (18 metastases and 1 HCC) were treated: 33% directly abutted the gallbladder, 33% were within 5 mm of the gallbladder, and 33% were between 6 and 10 mm from the gallbladder. Mean tumor size was 2.7 cm (range: 1.0–5.0 cm) with a mean of 3.5 cryoprobes (range: 1–6) used. In 95% of cases, the ice ball extended into the gallbladder lumen for a mean distance of 6 mm (range: 2–18 mm). No gallbladder-related complications occurred; asymptomatic mild focal gallbladder wall thickening was seen in 42% on imaging at 24 hours that subsequently resolved, and one patient developed pericholecystic fluid that resolved. The lack of complications was attributed to the protective effect of warm bile. In addition, several other retrospective studies ^{13 20 21 22} assessing the safety of cryoablation in various high-risk locations that included tumors next to the gallbladder also did not show any gallbladder-related complications; across these studies, a total of 36 liver tumors 5 to 10 mm from the gallbladder were treated. However, contrasting results were seen when Zeng et al ²³ compared the effects on the gallbladder wall in 12 rabbits treated with cryoablation to 12 rabbits treated with irreversible electroporation (IRE). In this study, the area of liver near the gallbladder was treated with each modality with probes placed 5 mm from the gallbladder. Ablation sizes were comparable in each group, but four rabbits in the cryoablation group showed histologic evidence of full-thickness gallbladder wall necrosis and perforation while no such changes were seen in the IRE group.

Bile Ducts

Thermal ablation of tumors adjacent to central bile ducts is often avoided due to risk of biliary complications such as biliary stricture, bile leak, or biloma. ¹² Ko et al ¹² retrospectively compared 31 patients treated with RFA and 25 patients with cryoablation of solitary small HCC lesions abutting the bile duct on MRI. The number of tumors abutting the central versus peripheral ducts was not significantly different in each group. Median follow-up was longer than 3 years for both the RFA and cryoablation cohorts. The authors found a higher incidence of biliary complications in the RFA group ( p  = 0.007), predominantly asymptomatic biliary strictures, and multivariate analysis showed RFA was an independent risk factor for biliary complications. However, subanalysis of tumors abutting the central bile ducts revealed no difference in incidence and severity of biliary complications between the RFA and cryoablation. Additionally, the LTP rate was not different between the two groups ( p  = 0.806). Another study by Yan et al ²² evaluating the safety of hepatic cryoablation in high-risk locations included 16 patients with tumors adjacent to the hepatic hilum, and no biliary complications were reported, though follow-up time was not specified.

Blood Vessels

When ablating tumors adjacent to blood major hepatic blood vessels, there is a risk of incomplete ablation due to heat-sink effect (for RFA and MWA) and cold-sink effect (for cryoablation), as well as a risk of vessel injury and thrombosis. ²⁴ A study by Kim et al ²⁴ reviewed the safety and efficacy of cryoablation in 58 patients with small solitary HCC lesions directly abutting a major vessel, defined as a first- or second-order branch of a portal or hepatic vein at least 3 mm in diameter: 51 tumors abutted a portal vein and 7 abutted a hepatic vein. Thrombus developed after ablation in the peritumoral vessel in four patients (6.9%) but spontaneously resolved without anticoagulation on subsequent follow-up imaging in half of these cases, and there were no cases of hepatic infarction. LTP occurred in six patients (10.7%) at a median follow-up of 22 months, with 5/6 (83.3%) occurring along the peritumoral vessel. The same group subsequently published a comparison of cryoablation and RFA for perivascular HCC with propensity score matching (PSM). ²⁵ In their unmatched cohort of 111 patients, there was a significant difference in the peritumoral vessel diameter between the cryoablation and RFA groups ( p  = 0.003), as well as a significant difference in the proportion of peritumoral vessels that were portal versus hepatic vein ( p  = 0.001). Therefore, PSM was performed where 50 patients (25 in each group) were matched for these factors among others, and they found no difference in local tumor recurrence rates ( p  = 0.379) between cryoablation and RFA. The vascular complication rate was reported only for the unmatched cohort and they found a lower rate of peritumoral vessel thrombosis and hepatic infarction in the cryoablation group than the RFA group (9.8 vs. 16.0%, and 3.3 vs. 12.0%, respectively), but this did not reach statistical significance ( p  = 0.137). In Ma et al's ¹³ study assessing the safety of cryoablation in various high-risk locations that included eight tumors directly abutting a portal or hepatic vein of at least 3 mm in diameter, there was no incidence of vessel thrombosis or infarction; however, LTP rates were not specifically reported for their patient cohort with perivascular tumors.

Diaphragm and Heart

Thermal ablation of hepatic tumors adjacent to the diaphragm carries a risk of injury to the diaphragm, lung, and heart; with tumors adjacent to the heart, grave complications such as cardiac perforation, tamponade, and arrhythmias must be avoided. ^{26 27} Yang et al ²⁶ retrospectively evaluated cryoablation of 74 HCC tumors abutting the diaphragm; mean distance from the tumor margin to the diaphragm in their cohort was 2.9 mm (range: 0–4). Mean tumor diameter was 3.3 cm (range: 0.8–7.0) and mean ablation zone size was 4.5 cm (range: 2.3–7.8). Sixteen patients (26.2%) developed a pleural effusion, with 4 of those requiring a chest tube. Three patients developed a pneumothorax and 1 of those required a chest tube; in all 3 patients, the cryoprobes were inserted via a transpulmonary approach and this was deemed to be the cause of pneumothorax. Ten patients developed minor complications of cough or pneumonitis. LTP was seen in 40.9% of tumors; this high LTP rate was not addressed by the authors, but is presumably due to larger lesion sizes in this cohort and technical challenges with probe insertion under the diaphragm. Kwon et al ²⁷ focused on outcomes of cryoablation of left hepatic lobe HCC adjacent to the heart. They compared 22 patients with juxtacardiac tumors (< 10 mm from the heart border) to 48 patients with non-juxtacardiac tumors (> 10 mm from the heart border). Mean tumor size for the entire cohort was 17.6 mm with no difference between the two groups ( p  = 0.781) and mean ablation zone size for the entire cohort was 42.8 mm, with no difference between the two groups ( p  = 0.115). No major complications occurred in either group and no complications related to thermal injury to the heart or diaphragm were reported. Overall LTR rate was 16.9%, with no difference between either group ( p  = 0.725). Several other retrospective studies ^{13 20 21 22} have assessed the safety of cryoablation in various high-risk locations that included a total of 72 tumors adjacent to the diaphragm or heart. Three cases of pleural effusion were reported in one of study, ²⁰ but no other complications related to local thermal injury were noted.

Gastrointestinal Tract

One of the most feared complications of ablation is perforation of the adjacent stomach or intestine from thermal injury. ²⁸ Li et al ²⁸ compared the results of cryoablation of peripheral liver tumors adjacent to the gastrointestinal (GI) tract, with and without artificial ascites. All tumors were located < 5 mm from the adjacent bowel. Forty-one tumors (mean size: 2.8 cm) were treated after instillation of artificial ascites to displace adjacent bowel, and 43 tumors (mean size: 2.6 cm) were treated without ascites ( p  > 0.05). Mean ablation zone size was not different between the two groups, 4.8 versus 4.4 cm, respectively ( p  > 0.05). However, there was a statistically lower rate of complete ablation in patients without artificial ascites than those with ascites, 81.3 and 95%, respectively ( p  < 0.05). Furthermore, there were six cases of GI tract injury in the group without ascites versus none in the group with ascites ( p  < 0.05), but no GI perforation occurred. Several other retrospective studies ^{13 20 21} have assessed the safety of cryoablation in various high-risk locations that included a total of 36 tumors adjacent to the GI tract and no cases of GI tract injury or perforation were seen.

Comparison of Cryoablation to Radiofrequency Ablation and Microwave Ablation

The safety and efficacy of cryoablation compared with RFA has been evaluated in multiple randomized control trials (RCTs). In a multicenter RCT by Wang et al ²⁹ involving 360 patients with cirrhosis and treatment-naive HCC ≤ 4 cm, 180 patients received cryoablation and 180 received RFA. The authors found no significant difference in LTP, tumor-free survival, overall survival (OS), or complication rate. However, they demonstrated lower cumulative LTP in the cryoablation group at 1, 2, and 3 years ( p  = 0.043) versus the RFA group. Additionally, in an ad hoc subgroup analysis, in patients with tumor > 3 cm, there was a significantly lower LTP rate in those treated with cryoablation compared with RFA ( p  = 0.041). In another multicenter RCT, Luo et al ³⁰ compared cryoablation to RFA in 223 elderly patients (>70 years old) with HCC ≤ 5 cm; 112 were treated with cryoablation and 111 with RFA. Like the previous RCT, ²⁹ they found no significant differences between the two groups in terms of local tumor recurrence, tumor-free survival, OS, or complication rate; and in an ad hoc subgroup analysis, the authors also reported in patients with tumor > 3 cm a significantly lower LTP rate in those treated with cryoablation compared with RFA ( p  = 0.039). Interestingly, the cryoablation group had a statistically lower distant intrahepatic tumor recurrence rate than RFA group at 1, 3, and 5 years ( p  = 0.007), suggesting a potential role of antitumor immunity after cryoablation (which will be discussed later). These two studies in HCC patients provided the strongest evidence to this day that cryoablation is a safe and effective alternative to RFA, and suggest superiority of cryoablation over RFA for lesions larger than 3 cm.

Real-world retrospective data also support equivalence between cryoablation and RFA. A retrospective review of Surveillance, Epidemiology and End Results (SEER) database using PSM by Xu et al ³¹ found no statically significant difference in liver cancer–specific survival and OS between cryoablation and RFA in HCC patients. In another retrospective analysis with PSM of the SEER database, Chen et al ³² concluded that cryoablation is non-inferior to RFA in terms of cancer-specific survival and OS in HCC patients. Dunne et al ³³ performed a single-center retrospective review of the safety profile of RFA and cryoablation for HCC. There were 39 tumors treated in each group and no significant difference in overall and severe complication rate was found between these two modalities, despite the tumor size and ablation zone size being significantly higher in the cryoablation group ( p  = 0.0022 and p  = 0.001, respectively).

Regarding comparison of cryoablation to MWA, there is very little literature comparing these two modalities, and most noticeably cryoablation has not been compared with MWA in an RCT to the authors' knowledge. Hu et al ³⁴ performed a retrospective comparison of outcomes of cryoablation (56 patients) to MWA (64 patients) for HCC and found that recurrence-free survival and OS were not significantly different between the two groups, but LTP and major complications were significantly lower in the cryoablation group ( p  = 0.003 and p  = 0.039, respectively). Finally, a recent meta-analysis by Gupta et al ⁸ compared outcomes of RFA, MWA, and cryoablation for HCC using comparative studies (randomized and non-randomized) with at least two arms comparing at least two of the three modalities. Only one study compared cryoablation to MWA, ³⁴ while the remainder compared RFA to MWA and RFA to cryoablation. The authors reported similar OS and local progression at 1 and 3 years between RFA, MWA, and cryoablation in very early and early-stage HCC.

Immunomodulatory Effects of Cryoablation

The increased immunomodulatory effects of cryoablation over heat-based ablation may be advantageous when considering locoregional treatment of hepatic tumors. The immunogenic potential of thermal ablation in general, as opposed to surgical resection, is thought to arise secondary to the in situ death of the tumor. ¹⁴ The resulting exposure of the immune system to the released tumor antigens can be akin to a vaccination-like mechanism, promoting anti-tumoral immunity against both local recurrence and distant disease, the latter a phenomenon known as the abscopal effect.

The proposed superiority of cryoablation compared with other heat-based ablation methods in eliciting an immune response is thought to arise from the propensity of RFA and MWA to denature the crucial tumor antigens, making them less available to the immune system. ³⁵ Cryoablation, on the other hand, releases intact tumor antigens into the local environment. ¹⁴ Indeed both animal and clinical studies have demonstrated increased serum key immunomodulating factors such as tumor necrosis factor-α, interleukin-6 (IL-6), and white blood cell counts after cryoablation compared with RFA or MWA. ^{36 37 38} Specifically in human tumors, a study by Erinjeri et al ³⁶ of 36 patients demonstrated cryoablation resulted in an increase of IL-6 by a factor of 54 compared with just 3.5 for hyperthermic ablation techniques.

Further investigation has set out to exploit and amplify the anti-tumoral immunogenic activation potential of cryoablation by combining the technique with immunotherapy as an adjunct. There are different points in the immune system which can be targeted in immunotherapy; toll like receptor agonist treatment invokes the activation of signaling pathways that ultimately result in the production of inflammatory mediators crucial to the immune response ³⁹ and has been combined with cryotherapy with improved outcomes in studies treating basal cell carcinomas and subcutaneous melanomas. ^{40 41} The instillation of immune mediating cells such as dendritic cells or natural killer (NK) cells is another strategy that may amplify the anti-tumoral response in conjunction with cryoablation. Focusing on hepatic tumors, a study of 61 patients by Lin et al ⁴² with HCC lesions demonstrated longer progression-free survival, and higher response and disease control rates in the group treated with cryoablation and allogeneic NK cells compared with cryoablation alone.

Checkpoint inhibitors are a third class of immunotherapy agents which appear to be the most extensively studied to date as combination treatment with cryoablation. Checkpoint inhibitors are antibodies against ligand-receptor mechanisms such as programmed death-ligand 1 (PD-L1) or programmed cell death protein 1 (PD-1), which otherwise may be utilized by tumor cells to suppress and evade T-cell-mediated assault. In a study of 63 mice by Mandt et al, ¹⁵ each with a single HCC lesion treated by cryoablation and an additional non-ablated distal lesion, the group receiving adjunct therapy with PD-1 inhibitor demonstrated 2.8-fold decrease in non-ablated tumor growth and decreased time to progression (hazard ratio = 0.47) compared with the cryoablation-only group. Furthermore, a single-arm trial by Shen et al ¹⁶ of 15 patients with multifocal hepatic melanoma metastasis demonstrated dual therapy with cryoablation of one to two lesions followed by infusion of an intra-arterial PD-1 inhibitor resulted in distal lesions' response rate of 26.7%, higher than previously reported response rates with PD-1 inhibitor alone. Additionally, one patient achieved complete response of distal tumors after failure of prior treatment with PD-1 inhibitor alone, which lends support to the ability of cryoablation to invoke an immune response via the release of preserved tumor antigens.

Discussion

Cryoablation has historically been unpopular in the liver, with RFA and MWA much more widely used to this day, due to concerns about complications that were seen in early series using open techniques and older cryotherapy technology. However, the advent of newer percutaneous technology combined with image-guided approaches has substantially improved the safety profile of hepatic cryoablation to the point of equivalency with heat-based ablation, as highlighted in this review. Hesitation to use cryoablation may also arise from the need to use multiple probes and longer ablation times.

However, as illustrated in this review, studies show that cryoablation is equally as safe and effective as RFA for HCC and liver metastases, and suggest that it may be more effective than RFA for lesions larger than 3 cm. The limited data comparing cryoablation to MWA suggest equivalency as well. Furthermore, studies evaluating cryoablation of liver tumors adjacent to critical structures such as the gallbladder, central bile ducts, bowel, diaphragm, and heart show a high safety profile.

As such, cryoablation deserves a place in the interventional toolbox for treating liver tumors. Outside of operator's preference, the authors believe that cryoablation should be the preferred method for treating liver tumors adjacent to critical structures when achieving adequate hydrodissection or pneumodissection for thermal protection seems unlikely. However, the authors still advocate for thermal protection of critical structures using instillation of fluid, air, or CO ₂ as much as possible, to safely allow for larger ablation zones with corresponding lower risk of local tumor recurrence. This is of particular importance when the tumor is adjacent to the GI tract, due to the significantly higher rate of tumor recurrence and GI tract injury seen in patients when artificial ascites is not used. ²⁸

Due to the lower degree of pain experienced, cryoablation should be considered when anesthesia or sedation resources are limited. In fact, cryoablation was performed under LA with or without IV analgesia, or moderate sedation, in many studies included in this review.

Care should still be taken when considering cryoablation of larger liver tumors. Several large studies demonstrated significantly increased risk of complications, namely, cryoshock and thrombocytopenia, when treating tumors over 4 cm. ^{10 17} One of these studies ¹⁷ also found the major complication rate was also significantly higher in patients with platelet counts < 100,000. This highlights the importance of careful patient selection. In tumors > 4 cm, the authors prefer to use chemoembolization combined with ablation, or radiation segmentectomy.

MWA generates hotter and larger ablation zones than RFA and has become the dominant thermal ablation modality in the United States. There is a notable paucity of comparative studies with cryoablation, and this is an area that should be considered for future research, particularly in the form of RCTs.

Perhaps a more important factor, however, in determining the role of cryoablation in the liver is its prominent immunomodulatory effects. In particular, the potential to harness these effects to improve local and systemic tumor response in combination with immunotherapy may elevate its status as a hepatic ablation modality. As discussed earlier, this approach is promising, and is in keeping with the shift in oncologic treatment toward synergistic therapies. ⁴³ While data have been emerging on the safety and efficacy of chemoembolization and radioembolization plus systemic therapy, ^{43 44} data on liver ablation plus systemic therapy are lacking. However, several clinical trials are ongoing to assess the outcomes of cryoablation combined with immunotherapy for HCC and liver metastases. More investigation into the sequencing of treatment and patient selection is needed to optimize safety, tumor response, and outcomes.

Conclusion

Cryoablation belongs in the interventional toolbox for locoregional treatment of primary and secondary liver tumors. It appears to be equally safe and effective as heat-based ablation, and may be particularly beneficial when ablating next to critical structures and when anesthesia or sedation resources are limited. The immunogenic potential of cryoablation is promising and further investigation into combining cryoablation with immunotherapy for hepatic malignancies is needed.