Percutaneous Cryoablation Versus Microwave Ablation for Hepatocellular Carcinoma: Comparative Study of Safety and Oncologic Outcomes

Abstract

Purpose

To compare safety and oncologic outcomes of percutaneous image-guided hepatic cryoablation and microwave ablation (MWA) for treating hepatocellular carcinoma (HCC).

Materials and Methods

This single center, retrospective study evaluated percutaneous image-guided cryoablation and MWA procedures in patients with HCC from December 2015 to December 2023. Adverse events (AEs) were recorded per procedure. Technical success and local tumor progression-free survival were assessed per tumor and compared after propensity score matching. Intrahepatic progression-free survival, extrahepatic progression-free survival, and overall survival were measured per patient when cryoablation or MWA was used as the initial curative-intent treatment for early-stage HCC. Survival metrics were estimated using Kaplan-Meier curves and compared using log-rank testing.

Results

A total of 176 ablation procedures in 123 patients with 228 HCCs enabled three cohort analyses (mean age ± SD, 69 years ± 9; 85 male). Per-procedure AE analysis of 170 procedures (69 cryoablation and 101 MWA) showed no evidence of differences in overall or major AE rates (cryoablation: overall AEs, 19%, major AEs, 1%; MWA: overall AEs, 10%, major AEs, 3%; P = .60). After propensity score matching, 158 HCC tumors were studied (79 cryoablation and 79 MWA): technical success for both modalities was 97% (P = .96); 3-year per-tumor local tumor progression-free survival rates were 57% for cryoablation and 58% for MWA (P = .53). Per-patient analysis of 51 patients with early-stage HCC (17 cryoablation and 34 MWA) yielded 1-year intrahepatic progression-free survival, extrahepatic progression-free survival, and overall survival rates of 82%, 88%, and 100% for cryoablation and 73%, 94%, and 88% for MWA, respectively, without evidence of a difference in any survival outcome (P = .76).

Conclusion

Percutaneous cryoablation and MWA demonstrated similar favorable safety and oncologic outcomes for initial treatment of HCC.

Keywords: Comparative Studies, Ablation Techniques, CT, Interventional-Oncology, Abdomen/GI, Liver

© RSNA, 2026

Summary

Percutaneous image-guided cryoablation and microwave ablation demonstrated similar safety and oncologic outcomes for treatment of hepatocellular carcinoma.

Key Points

• ■ In this retrospective study of 170 procedures in 120 patients, percutaneous cryoablation and microwave ablation for treatment of hepatocellular carcinoma (HCC) demonstrated no evidence of differences in major adverse event rates (1% vs 3%; P = .60).

• ■ After propensity score matching of 158 HCC tumors in 84 patients, 3-year local tumor progression-free survival rates were similar between cryoablation and microwave ablation (57% vs 58%; P = .53).

• ■ Among 51 patients with early-stage HCC, curative-intent cryoablation and microwave ablation demonstrated no evidence of differences in intrahepatic progression-free survival, extrahepatic progression-free survival, or overall survival (P = .76).

Introduction

Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related mortality worldwide (1). Since 1980, the incidence of HCC has more than tripled, and mortality has doubled, with more than 700 000 deaths reported annually (1). At initial diagnosis, approximately 45% of patients present with intermediate- to advanced-stage disease (2).

For patients with solitary tumors without extrahepatic disease, current National Comprehensive Cancer Network and Barcelona Clinic Liver Cancer guidelines recommend partial hepatectomy as a potentially curative treatment, with perioperative mortality rates of less than 5% and 5-year survival rates of approximately 70% (3–5). Additionally, percutaneous image-guided thermal ablation may be considered a first-line treatment option for patients with very early– and early-stage HCC who are not candidates for surgery (3–6). Percutaneous image-guided ablative techniques, including radiofrequency ablation (RFA), microwave ablation (MWA), and cryoablation, are alternatives for patients with small or unresectable HCC (7–9). Cryoablation has historically been used less commonly than heat-based methods, partly due to concerns regarding bleeding, bile leak, and cryoshock observed during treatment of large tumors using large probes during open surgery (10,11). However, the development of thin, needle-like probes and argon and/or helium–based cryoablation systems has led to a more favorable safety profile (12–14).

Importantly, cryoablation offers unique technical advantages, including direct visualization of the ice ball and real-time monitoring of ablation margins, which may help ensure complete tumor coverage while avoiding injury to adjacent critical structures. Recent randomized controlled trials have demonstrated comparable efficacy between MWA and RFA, with the additional advantages of shorter procedure times and reduced susceptibility to the heat-sink phenomenon for MWA (15,16). However, head-to-head comparisons between cryoablation and MWA for treatment of HCC remain limited, with available data being largely indirect and lacking evaluation of adverse event (AE) rates (17). The purpose of this study was to compare the safety and oncologic outcomes of cryoablation and MWA for treatment of HCC.

Materials and Methods

Study Design

This single-center, retrospective, Health Insurance Portability and Accountability Act–compliant study was approved by the institutional review board and performed at a quaternary care center (18). The requirement for written informed consent was waived. The institutional interventional radiology procedure database was searched for cryoablation and MWA procedures performed for treatment of HCC between December 2015 and December 2023; the end of the study period was selected to allow adequate clinical and imaging follow-up for time-to-event outcome analyses.

Three analyses were conducted to compare cryoablation versus MWA for treatment of adult patients (≥18 years) with HCC: (a) per-procedure analysis of AEs; (b) per-tumor analysis of technical success and local tumor progression-free survival (LTPFS); and (c) per-patient analysis of LTPFS, secondary efficacy, intrahepatic progression-free survival (IHPFS), extrahepatic progression-free survival (EHPFS), and overall survival (OS). Inclusion and exclusion criteria and end point definitions for each analysis are shown in Figure 1.

Figure 1:

Flowchart shows study sample derivation and categorization based on the respective inclusion and exclusion criteria for the per-tumor, per-procedure, and per-patient outcome analyses. BCLC = Barcelona Clinic Liver Cancer.

Per-Procedure Analysis of AEs

Per-procedure analysis of AEs included procedures in which only cryoablation or MWA was used to treat HCC. All patients had Barcelona Clinic Liver Cancer stage B or lower disease, except for two patients with stage C disease. All patients had preprocedural laboratory values with platelet counts greater than 50 000/µL and an international normalized ratio of 1.5 or less. Procedures combining treatment modalities, such as cryoablation and MWA during the same procedure, were excluded (Fig 1).

AEs were identified through medical record review and categorized according to Society of Interventional Radiology guidelines (18). Additional retrospectively collected variables included image guidance modality, number of probes used, need for adjunctive techniques, and type of anesthesia. Overall AE rates and major AE rates (grade 3 or higher) were calculated for cryoablation and MWA procedures.

Per-Tumor Analysis of Technical Success and LTPFS

Per-tumor analysis of technical success and LTPFS included HCC tumors treated with either cryoablation or MWA. All patients had Barcelona Clinic Liver Cancer stage B or lower disease, except for one patient with stage C disease. Exclusion criteria included tumors previously treated with surgery or arterially directed therapy, lack of imaging follow-up, noncurative intent, or combination treatment with radiation therapy (Fig 1).

Technical success was assessed by the operator on a per-tumor basis and defined as treatment according to protocol with complete tumor coverage by the ablation zone on the first postprocedural contrast-enhanced imaging examination (19). LTPFS was calculated per tumor and defined as the interval from the first ablation procedure to the date of local progression. Additional retrospectively collected variables included tumor type, size, segmental location, and history of prior cancer treatment. New nodular or thick crescentic tumoral enhancement at follow-up imaging was used to confirm local progression (19).

Per-Patient Analysis of Oncologic Outcomes

Patient-level analysis was conducted for all patients in whom cryoablation or MWA was used as the initial treatment for Barcelona Clinic Liver Cancer stage 0–A disease. Inclusion criteria were HCC Barcelona Clinic Liver Cancer stage 0–A disease, Child-Pugh class B or lower, and Eastern Cooperative Oncology Group performance status 0–1. Exclusion criteria included prior treatment for HCC or another malignancy, cryoablation combined with MWA or nonthermal ablation, and lack of imaging or clinical follow-up (Fig 1).

LTPFS was defined above and is reported here on a per-patient basis. Secondary efficacy per patient was defined as the LTPFS rate including repeat ablations of target tumors. IHPFS was defined as the interval from the first ablation procedure to first tumor progression either at the ablation site or elsewhere in the liver, as identified at imaging. EHPFS was defined as the interval from the first ablation procedure to first distant tumor progression confirmed at imaging. Patients without tumor recurrence were censored at the date of the last CT or MRI examination. OS was defined as the interval from the first ablation procedure to death from any cause, with censoring of patients alive as of April 30, 2024. Additional retrospectively collected variables included sex, age, maximum tumor size, number of treated lesions, procedure time, Child-Pugh class, Model for End-stage Liver Disease score, Eastern Cooperative Oncology Group performance status, etiology of liver cirrhosis, and preprocedural α-fetoprotein levels.

Cryoablation and MWA Procedures and Follow-up

Procedures were typically performed under general anesthesia (n = 160) and less frequently under monitored anesthesia care with intravenous sedation (n = 10). Patients were selected for treatment with cryoablation or MWA at the discretion of the interventional radiologist, generally based on tumor size and proximity to adjacent critical structures, both of which informed the need for enhanced visualization of the ablation zone. Ablation procedures and assessment of intraprocedural imaging outcomes (technical success and ablation zone coverage) were performed by one of four interventional radiologists with 1–28 years of tumor ablation experience (S.K.B., S.G.S., K.T., or P.B.S.). All ablations were performed during a single procedure, with detailed technical descriptions of CT- and MRI-guided cryoablation and MWA procedures reported previously (20,21)

Cryoablation was performed using 14–17-gauge applicators (Boston Scientific) and a freeze-thaw-freeze cycle, typically consisting of a 15-minute freeze, 10-minute thaw, and 15-minute freeze. The ice ball was monitored with serial CT imaging during and after ablation. MWA was performed using 14–16-gauge applicators (HS Medical, MedWaves, and NeuWave Medical) with power and time settings based on manufacturer recommendations for predicted ablation zone size. When feasible, ablations were performed with the intent of achieving a minimum 5–10-mm ablation margin. Selection of cryoablation versus MWA and image guidance modality (CT, MRI, or US) was at the discretion of the operator based on tumor location, size, and proximity to critical structures.

Following the procedure, patients were followed in the tumor ablation clinic and underwent MRI or CT (with and without intravenous contrast material unless contraindicated) within 30 days, every 3 months for 1 year, every 3–6 months for an additional 2 years, and annually thereafter. Imaging outcomes related to local tumor progression were determined from medical record reports generated by institutional radiologists with 1–28 years of abdominal radiology experience using established postablation response criteria (19).

Statistical Analysis

Continuous variables were reported as means ± SDs for normally distributed data and as medians with IQRs for nonnormally distributed data. Categorical variables were reported as counts and frequencies. Differences in categorical variables were assessed using the Fisher exact test or χ² test, as appropriate. For continuous variables, the Student t test or Mann-Whitney U test was used for normally and nonnormally distributed data, respectively. Statistical significance was defined as a two-sided P value less than .05. For the per-procedure analysis, overall and major AE rates for cryoablation and MWA were compared using the χ² test. For the per-tumor analysis, propensity score matching (PSM) was applied to balance clinical variables between treatment groups as described by Yao et al (22) and Zhao et al (23). Covariates used to estimate the propensity score were selected a priori on the basis of clinical relevance and included tumor size and history of prior cancer. Groups were matched in a 1:1 ratio without replacement. Propensity scores were estimated using logistic regression with nearest-neighbor matching. Covariate balance between matched cryoablation and MWA groups was assessed using standardized differences (24).

For the per-patient analysis, LTPFS, secondary efficacy, IHPFS, EHPFS, and OS were evaluated using Kaplan-Meier survival analysis and compared with the log-rank test. PSM was not performed in this subcohort because of the limited sample size and risk of small sample bias. All statistical analyses were performed using R software (version 4.4.0; R Foundation for Statistical Computing; https://www.r-project.org/).

Results

Patient, Tumor, and Procedural Characteristics

For this study, 176 liver cryoablation and/or MWA procedures performed in 123 adult patients for treatment of 228 primary HCCs were included in one or more of three cohort analyses (mean age ± SD, 69 years ± 9; 85 male patients) (Fig 1). The final per-procedure cohort included 170 procedures in 120 patients (Fig 1). Procedural characteristics for cryoablation (n = 69) and MWA (n = 101) are summarized in Table 1. Cryoablation required significantly more probes than MWA (mean ± SD, 2.87 ± 1.27 vs 1.18 ± 0.52; P = .001) and longer procedure times (139.1 min ± 38.3 vs 96.0 min ± 43.6; P = .001). Use of adjunctive techniques also differed between modalities (P = .034), with less frequent use of artificial ascites during cryoablation procedures (6% vs 15%) (Table 1 and Figs 2, 3). Before PSM, the per-tumor cohort consisted of 195 HCCs in 120 patients treated with cryoablation (n = 79) or MWA (n = 116). After PSM, the final per-tumor cohort included 158 HCCs in 84 patients treated with cryoablation (n = 79) or MWA (n = 79) (Fig 1). Mean tumor size was not significantly different between the cryoablation and MWA groups (1.98 cm ± 0.87 [range, 0.6–4.8 cm] vs 1.99 cm ± 0.97 [range, 0.3–5.2 cm]; P = .96; Table 2). Median imaging follow-up was 25 months (IQR, 10–32 months; range, 1–115 months) for the cryoablation cohort and 22 months (IQR, 10–32 months; range, 1–91 months) for the MWA cohort (P = .83). The final per-patient analysis included 51 patients treated with cryoablation (n = 17) or MWA (n = 34) as the initial curative-intent treatment for HCC (Fig 1). Mean patient age was 67 years ± 8, and 37 of 51 patients (73%) were male. Baseline characteristics were comparable between groups (Table 3).

Table 1:

Percutaneous Cryoablation Versus MWA for Hepatocellular Carcinoma in Per-Procedure Cohort

Characteristic	Overall (n = 170)	Cryoablation (n = 69)	MWA (n = 101)	P Value
Image guidance modality				.009
CT	118 (69)	57 (83)	61 (60)
CT/US	22 (13)	5 (7.2)	17 (17)
MRI	30 (18)	7 (10)	23 (23)
No. of probes used	1.86 ± 1.23	2.87 ± 1.27	1.18 ± 0.52	.001
Procedure time (min)	113.8 ± 46.5	139.1 ± 38.3	96 ± 43.6	.001
Special technique				.03
Artificial ascites	19 (11)	4 (6)	15 (15)
Hydrodissection	17 (10)	10 (14)	7 (7)
Balloon separation	2 (1)	2 (3)	0
None	132 (78)	53 (77)	79 (78)
Anesthesia type				.70
General anesthesia	160 (94)	66 (96)	94 (93)
Monitored anesthesia care (intravenous)	10 (6)	3 (4)	7 (7)
Overall AEs	23 (14)	13 (19)	10 (10)	.10
Major AEs (grade III, IV)	4 (2)	1 (1)	3 (3)	.60
Highest AE class				.11
0 (none)	147 (86)	56 (81)	91 (90)
Grade I	9 (5)	7 (10)	2 (2)	.06
Asymptomatic pneumothorax	5 (3)	4 (6)	1 (1)
Serum creatinine level >0.3 mg/dL	2 (1)	1 (1)	1 (1)
Segmental portal vein thrombosis	2 (1)	2 (3)	0
Grade II	10 (6)	5 (6)	5 (5)	.77
Thrombocytopenia (platelets 50 000–75 000/μL)	4 (3)	2 (3)	2 (2)
Pleural effusion requiring catheter drainage	2 (1)	1 (1)	1 (1)
Pneumothorax requiring chest tube	2 (1)	1 (1)	1 (1)
Hypertension requiring pharmacologic treatment	1 (0.5)	0	1 (1)
Hemorrhage not requiring transfusion	1 (0.5)	1 (1)	0
Grade III	3 (2)	1 (1)	2 (2)	.79
Symptomatic portal vein thrombosis requiring treatment	1 (0.5)	0	1 (1)
Symptomatic hemorrhage requiring nonemergent transfusion	1 (0.5)	0	1 (1)
Thrombocytopenia (platelet count 25 000–50 000/μL) requiring transfusion	1 (0.5)	1 (1)	0
Grade IV	1 (0.5)	0	1 (1)	.99
Bowel perforation requiring surgery	1 (0.5)	0	1 (1)
Grade V	0	0	0

Note.—Data are numbers of procedures with percentages in parentheses (rounded) or means ± SDs. AE = adverse event, MWA = microwave ablation.

Figure 2:

Images in a 68-year-old female patient with a history of metabolic syndrome and cirrhosis. (A) A contrast-enhanced axial T1-weighted MR subtraction image demonstrates an arterial-phase, hyperenhancing, 2.2-cm Liver Imaging Reporting and Data System (LI-RADS) 5 observation (arrow) in hepatic segment 8. (B) An intraprocedural axial CT image shows the microwave ablation probe (arrow) along the medial aspect of the hypodense subcapsular tumor. Artificial ascites (*) using normal saline was infused to separate the liver from the body wall and diaphragm. This maneuver reduces potential postprocedural pain from thermal injury to the peritoneum, ribs, intercostal nerves, and soft tissues. (C) A contrast-enhanced axial CT image immediately following microwave ablation demonstrates the size and shape of the hypoenhancing ablation zone (arrows), which would be difficult to accurately assess at unenhanced CT. This ablation procedure provided a bridge to liver transplantation 17 months later. The ablated tumor was completely nonviable on explant pathologic assessment, although the patient had two separate small foci of hepatocellular carcinoma elsewhere in the liver.

Figure 3:

Images in a 54-year-old male patient with chronic hepatitis B who underwent surgical wedge resection 5 years earlier for a 1.2-cm hepatocellular carcinoma in segment 8. He returned with a 1.5-cm Liver Imaging Reporting and Data System (LI-RADS) 5 observation in hepatic segment 4A/2. (A) The axial planning CT image shows the hypodense tumor (arrow). (B) The intraprocedural axial CT image shows excellent visualization of the ice ball (short arrows) surrounding one of the cryoprobes. Note that the ice ball was monitored as it approached the left hepatic bile duct (long arrow) helping to prevent injury. (C) An intraprocedural coronal CT image shows the ice ball (arrows) extending to the diaphragm just beneath the heart (*). Cryoablation into the diaphragm tends to cause less pain than microwave ablation. The cryoablation provided a bridge to liver transplantation 13 months later. Pathologic assessment of the explanted liver found the entire liver, including the ablation scar, without residual or viable tumor.

Table 2:

Baseline Hepatocellular Carcinoma Characteristics in Tumors Treated with Either Cryoablation or MWA Before and After PSM in Per-Tumor Cohort

Characteristic	Before PSM				After PSM
	Overall (n = 195)	Cryoablation (n = 79)	MWA (n = 116)	P Value	Overall (n =158)	Cryoablation (n = 79)	MWA (n = 79)	P Value
Type				.17				.72
Solitary	127 (65)	56 (71)	71 (61)		114 (73)	56 (71)	58 (73)
Multiple	68 (35)	23 (29)	45 (39)		44 (27)	23 (29)	21 (27)
Tumor size (cm)	1.91 ± 0.94	1.98 ± 0.87	1.87 ± 0.99	.15	1.98 ± 0.92	1.98 ± 0.87	1.99 ± 0.97	.96
Location (segment)				.34				.52
II	16 (9)	10 (13)	6 (5)		14 (9)	10 (13)	4 (5)
III	11 (6)	3 (4)	8 (7)		8 (5)	3 (4)	5 (6)
IV	30 (15)	15 (19)	15 (13)		25 (16)	15 (19)	10 (13)
V	30 (15)	12 (15)	18 (16)		25 (16)	12 (15)	13 (16)
VI	30 (15)	13 (16)	17 (15)		28 (18)	13 (16)	15 (19)
VII	31 (16)	11 (14)	20 (17)		22 (14)	11 (14)	11 (14)
VIII	47 (24)	15 (19)	32 (28)		36 (23)	15 (19)	21 (27)
Size category (cm)				.51				.97
<1	21 (11)	7 (9)	14 (12)		14 (9)	7 (9)	7 (9)
1 to <2	86 (44)	33 (42)	53 (46)		65 (42)	33 (42)	33 (42)
2 to <3	63 (32)	30 (38)	33 (28)		57 (37)	30 (38)	30 (38)
≥3	25 (13)	9 (11)	16 (14)		20 (13)	9 (11)	9 (11)
Prior treatment*	111 (57)	47 (59)	64 (55)	.56	91 (58)	47 (59)	45 (58)	.87

Note.—Data are numbers of tumors with percentages in parentheses or means ± SDs. MWA = microwave ablation, PSM = propensity score matching.

^*Prior treatment includes history of systemic chemotherapy, external beam radiation therapy, and/or local-regional therapy for a different tumor or type of cancer.

Table 3:

Baseline Characteristics of Patients Who Underwent Either Cryoablation or MWA as First-Line Curative-Intent Treatment for Early-Stage Hepatocellular Carcinoma in Per-Patient Cohort

Characteristic	Overall (n = 51)	Cryoablation (n = 17)	MWA (n = 34)	P Value
Sex				.10
Male	37 (73)	15 (88)	22 (65)
Female	14 (27)	2 (12)	12 (35)
Age (y)	67 (8)	65 (8)	68 (7)	.31
Maximum tumor size (cm)	2.34 ± 0.77	2.36 ± 0.85	2.31 ± 0.62	.88
No. of lesions treated	1.08 ± 0.27	1.12 ± 0.33	1.06 ± 0.24	.48
Child-Pugh class				.16
A	46 (90)	17 (100)	29 (85)
B	5 (10)	0	5 (15)
MELD score	8.00 (7.00–11.00)	8.00 (7.00–10.00)	8.00 (7.25–11.75)	.19
ECOG performance				.54
0	48 (94)	17 (100)	31 (91)
1	3 (6)	0	3 (9)
Origin of liver cirrhosis				.29
Hepatitis C virus	20 (41)	5 (29)	15 (47)
Steatohepatitis	14 (29)	5 (29)	9 (28)
Alcoholism	12 (24)	4 (24)	8 (25)
Alcoholism and hepatitis C virus	1 (2)	1 (6)	0
Hepatitis B virus	1 (2)	1 (6)	0
Hemochromatosis	1 (2)	1 (6)	0
Other	2	0	2
Preprocedural α-fetoprotein level (ng/mL)	28 ± 52	44 ± 74	19 ± 32	.18
Unknown	6	1	5

Note.—Data are numbers of tumors with percentages in parentheses, means ± SDs, or medians with IQRs in parentheses. ECOG = Eastern Cooperative Oncology Group, MELD = Model for End-Stage Liver Disease, MWA = microwave ablation.

Per-Procedure Analysis of AEs

Twenty-three AEs occurred in 170 procedures (14%) (Table 1). Nine were grade 1 events, including five asymptomatic pneumothoraces, two segmental portal vein thromboses, and two transient acute kidney injuries with serum creatinine elevation levels of 0.3 mg/dL or greater. Ten were grade 2 events, including four thrombocytopenias (platelet count, 50 000–75 000/μL), two pleural effusions requiring catheter drainage, two pneumothoraces requiring chest tube placement, one hemorrhage not requiring transfusion, and one episode of hypertension requiring pharmacologic treatment. Three grade 3 events occurred, including one symptomatic portal vein thrombosis requiring treatment, one symptomatic hemorrhage requiring nonemergent transfusion, and one thrombocytopenia (platelet count, 25 000–50 000/μL) requiring platelet transfusion. One grade 4 event consisted of bowel perforation requiring surgery. No cryoshock events or grade 5 events were observed. There was no evidence of differences in overall AE rates between cryoablation and MWA (19% vs 10%; P = .10) or in major AE rates (1.4% vs 3%, respectively; P = .60).

Per-Tumor Analysis of Technical Success and LTPFS

Technical success was achieved in 96% (76 of 79) of tumors treated with cryoablation and 98% (114 of 116) of tumors treated with MWA (P = .96). Pre- and post-PSM Kaplan-Meier curves for LTPFS are shown in Figure 4A and 4B. Estimated 1-, 2-, and 3-year per-tumor LTPFS rates were 82%, 68%, and 57%, respectively, for the cryoablation group and 77%, 64%, and 58%, respectively, for the MWA group, with no significant difference before or after PSM (P = .53) (Table 4).

Figure 4:

Kaplan-Meier curves show local tumor progression-free survival (per tumor) (A) before propensity matching and (B) with propensity matching for the full tumor cohort, as well as (C) poststratification for tumor sizes less than 3 cm and 3 cm or greater for hepatocellular carcinoma treated with either cryoablation or microwave ablation.

Table 4:

Kaplan-Meier–estimated LTPFS (Per Patient and Per Tumor), Secondary Efficacy, IHPFS, EHPFS, and Overall Survival for Cryoablation versus MWA

Oncologic Outcome at Follow-up	Thermal Ablation Modality		P Value
	Cryoablation	MWA
LTPFS (per tumor; before PSM)	79	116	.79
1 y	82 (74, 92)	79 (71, 87)
2 y	68 (57, 81)	67 (57, 78)
3 y	57 (44, 75)	61 (51, 74)
LTPFS (per tumor; after PSM)	79	79	.53
1 y	82 (74, 92)	77 (67, 87)
2 y	68 (57, 81)	64 (57, 81)
3 y	57 (44, 75)	58 (46, 73)
LTPFS (tumors <3 cm)	70	70	.47
1 y	85 (77, 94)	80 (70, 90)
2 y	73 (62, 86)	67 (56, 81)
3 y	61 (47, 80)	61 (48, 77)
LTPFS (tumors ≥3 cm)	9	9	.99
1 y	56 (31, 100)	58 (31, 100)
2 y	29 (9.1, 93)	42 (18, 94)
3 y	NA	42 (18, 94)
LTPFS (per patient)	17	34	.26
1 y	82 (66, 100)	81 (68, 96)
2 y	46 (26, 81)	67 (52, 86)
3 y	17 (3.5, 86)	49 (29, 81)
Secondary efficacy (per patient)	17	34	.59
1 y	100 (100, 100)	93 (85, 100)
2 y	92 (79, 100)	85 (72, 100)
3 y	79 (56, 100)	78 (62, 98)
IHPFS	17	34	.98
1 y	82 (65, 100)	73 (59, 90)
2 y	41 (22, 75)	41 (27, 62)
3 y	11 (1.9, 64)	27 (14, 55)
EHPFS	17	34	.89
1 y	88 (73, 100)	94 (86, 100)
2 y	88 (73, 100)	87 (76, 100)
3 y	80 (61, 100)	80 (65, 99)
Overall survival	17	34	.76
1 y	100 (100, 100)	88 (78, 100)
2 y	94 (84, 100)	82 (70, 96)
3 y	68 (46, 100)	79 (66, 94)

Note.—Data are total numbers for each category or percentages with 95% CIs in parentheses. EHPFS = extrahepatic progression-free survival, IHPFS = intrahepatic progression-free survival, LTPFS = local tumor progression-free survival, MWA = microwave ablation, PSM = propensity score matching.

After PSM and stratification by tumor size, estimated 1-, 2-, and 3-year LTPFS rates for tumors smaller than 3 cm were 85%, 73%, and 61%, respectively, for the cryoablation group and 80%, 67%, and 61%, respectively, for the MWA group (Fig 4B). For tumors 3 cm or larger, estimated 1- and 2-year LTPFS rates were 56% and 29%, respectively, for the cryoablation group and 58% and 42%, respectively, for the MWA group (Fig 4C). No significant differences were observed between cryoablation and MWA for either subgroup (P = .47 and P = .99, respectively) (Table 4).

Per-Patient Analysis of Oncologic Outcomes

There was no evidence of differences between cryoablation and MWA for any survival outcome end point. For per-patient LTPFS, estimated 1-, 2-, and 3-year per-patient LTPFS rates were 82%, 47%, and 17% for the cryoablation group and 81%, 67%, and 49% for the MWA group, respectively (P = .26) (Fig 5A and Table 4). For secondary efficacy, estimated 1-, 2-, and 3-year secondary efficacy rates were 100%, 92%, and 79% for the cryoablation group and 93%, 85%, and 78% for the MWA group, respectively (P = .59) (Fig 5B and Table 4). For IHPFS, estimated 1-, 2-, and 3-year IHPFS rates were 82%, 41%, and 11% for the cryoablation group and 73%, 41%, and 27% for the MWA group, respectively (P = .98) (Fig 5C and Table 4). For EHPFS, estimated 1-, 2-, and 3-year EHPFS rates were 88%, 88%, and 80% for the cryoablation group and 94%, 87%, and 80% for the MWA group, respectively (P = .89) (Fig 5D and Table 4). For OS, median follow-up for patients without an event was 44.5 months (IQR, 28.5–63.5 months). Estimated 1-, 2-, and 3-year OS rates were 100%, 94%, and 68% for the cryoablation group and 88%, 82%, and 79% for the MWA group, respectively (P = .76) (Fig 5E and Table 4).

Figure 5:

Kaplan-Meier curves for (A) local tumor progression-free survival (per patient), (B) secondary efficacy (per patient), (C) intrahepatic progression-free survival, (D) extrahepatic progression-free survival, and (E) overall survival between patients who underwent microwave ablation or cryoablation as the first initial curative-intent treatment for hepatocellular carcinoma.

Discussion

The standard of care for initial HCC management is usually surgery; however, a substantial proportion of patients are not optimal surgical candidates or have tumors that are not amenable to resection (5). In this context, percutaneous thermal ablation has an established role in very early– and early-stage HCC, due to similar oncologic results and lower cost compared with surgical resection (6). Heat-based ablation modalities, RFA and MWA, are the most widely adopted (6,25); MWA enables shorter procedure times and less heat-sink effects than RFA (26). However, cryoablation, a cold-based ablation modality has the advantage of precise monitoring because the ice ball can be directly visualized (especially using CT or MRI) during the procedure (27,28). In contrast, the MWA zone is poorly visualized until intravenous contrast material is administered at the conclusion of the application. The ability to monitor the cryoablation zone precisely may make it more likely that tumors are completely treated. This is particularly important because complete ablation with adequate margins is one of the most significant factors impacting local tumor progression (19). A potential safety advantage of cryoablation is that the edge of the ice ball demarcates the 0 °C isotherm, enabling precise monitoring when ablating tumors near or adjacent to critical anatomic structures (21,31). Despite these advantages, the adoption of cryoablation has lagged behind RFA and MWA due to lingering concerns regarding the safety of liver cryoablation; this may explain the relatively limited volume of contemporary percutaneous cryoablation literature compared with MWA (10,32). We therefore sought to compare cryoablation with MWA. Our study showed that there was no significant difference in overall or major AEs between hepatic cryoablation and MWA for the treatment of patients with HCC: 19% versus 10% for overall AEs, respectively (P = .10), and 1% versus 3% for major AEs, respectively (P = .60). In addition, cryoablation and MWA demonstrated comparable 1-, 2-, and 3-year LTPFS, IHPFS, EHPFS, and OS rates (P > .05).

Regarding MWA safety, our results are in line with those reported in a recent systematic review and meta-analysis of randomized controlled trials comparing MWA to RFA for the treatment of HCC, where the aggregated pooled rate of AEs after MWA was 3.3% (95% CI: 1.7, 4.9). Regarding cryoablation safety, our results are similar or better than those of prior single-arm studies that reported overall AE rates ranging from 5.8% to 31.3% for tumors smaller than 4 cm (13,21,33). However, our study included no patients with bleeding diatheses (platelet count < 50 000/µL) and only eight patients with tumors 4 cm or larger, factors that may increase the risk of AEs following cryoablation (8,30). No patients in our study developed cryoshock after cryoablation. This result is consistent with a recent systematic review and meta-analysis that aimed to evaluate the overall incidence of cryoshock after cryoablation for the treatment of any liver lesion (34). After pooling the results of 26 studies including 4029 patients, cryoshock occurred in less than 0.3% (10 of 4029) of the procedures and was not reported for liver lesions 3 cm or smaller (34).

Our study showed that cryoablation and MWA yielded similar oncologic outcomes in terms of 1-, 2-, and 3-year LTPFS rates (P > .05). In addition, we found no evidence of differences in OS, IHPFS, and EHPFS when cryoablation and MWA were used as first-line curative-intent treatment for HCC. In this domain, there is a paucity of data directly comparing cryoablation to MWA in the literature. A recent systematic review and Bayesian network meta-analysis compared the efficacy of RFA, MWA, and cryoablation for very early– and early-stage HCC and found no significant difference in the local tumor progression rates and OS at 1 and 3 years (16). Two randomized studies compared cryoablation to RFA and found no evidence of differences in the LTPFS; OS at 1, 2, and 3 years; or AE rate (35,36). Both studies reported that in patients with tumors larger than 3 cm, cryoablation resulted in lower local tumor progression rates compared with RFA (P = .04), perhaps secondary to the intraprocedural benefit of being able to monitor ablation effects (35,36). Our results are also consistent with a retrospective study of 120 patients with HCC in high-risk liver locations, defined as lesions with a proximity of 1 cm or less to the abdominal wall and surrounding organs (37). Local tumor progression rates after 1 year were slightly lower for MWA (15.9%) compared with 19.0% for cryoablation (P = .04) but at 1, 3, and 5 years, the OS rate, distant metastasis incidence, and progression-free survival during the 5-year follow-up were not different.

Our study had limitations. This was a single institution, retrospective study. The retrospective nature of the study precluded analysis of operator rationale for choice of cryoablation and MWA and was not designed to assess operator-dependent differences. Accordingly, a selection bias is not excluded; however, whether before or after PSM, no significant differences in patient or tumor characteristics were identified between the two cohorts. PSM was used to balance selected tumor characteristics, but hidden allocation biases may not have been eliminated. In addition, even though no evidence of differences was found between the two treatment modalities when used as first curative-intent treatment, the sample size was relatively small (MWA, n = 34 and cryoablation, n = 17) for the per-patient oncologic outcomes and underscores the need for larger, prospective studies. Although not used in this study, contemporary ablation margin confirmation software may improve and standardize the assessment of thermal ablation margins (29,30). Additionally, data on tumor proximity to critical structures were not collected, limiting correlation with modality selection factors and procedural efficiency. Lastly, interobserver agreement for image interpretations was not formally assessed, which may limit the reproducibility of the imaging-based outcome measures of technical success, ablation zones, and local tumor progression. In conclusion, the comparable oncologic outcomes and safety of cryoablation relative to MWA in our study suggest that historical concerns regarding liver cryoablation safety should not dissuade the interventional radiologist from using image-guided cryoablation to treat small HCC.

Disclosures of conflicts of interest

Please see ICMJE form(s) for author conflicts of interest. These have been provided as supplemental materials.