Long-term efficacy and safety of microwave ablation for hepatocellular carcinoma adjacent to the gallbladder with a diameter ≤ 5 cm: a multicenter, propensity score matching study

Abstract

Objective

To compare the long-term efficacy and safety of microwave ablation (MWA) as first-line therapy for hepatocellular carcinoma (HCC) adjacent versus nonadjacent to the gallbladder.

Materials and Methods

From 2006 to 2018, 657 patients with ≤5 cm HCC who underwent percutaneous ultrasound-guided MWA as first-line therapy from 5 hospitals were enrolled in this retrospective study. Patients were grouped into the adjacent group (n = 49) and the nonadjacent group (n = 608) according to whether the tumor was adjacent to the gallbladder. Propensity score matching (PSM) was used to balance baseline variables between the two groups.

Results

Forty-eight patient pairs were matched after PSM. For the PSM cohort, during a median follow-up time of 60 months, there were no differences in PFS (hazard ratio [HR], 1.011; 95% confidence interval [CI], 0.647-1.578; p = 0.963) or OS (HR 0.925; 95% CI 0.522-1.639; p = 0.789) between the adjacent and nonadjacent groups. Univariate and multivariate analyses revealed that the tumor adjacent to the gallbladder was not an independent risk factor for PFS or OS (all p > 0.05). Subgroup analysis showed comparable PFS and OS between the two groups in the <3 cm subgroup and the 3-5 cm subgroups (all p > 0.05). In addition to more use of assistive technology (p < 0.05), the adjacent group shared comparable local tumor progression, complications, technical success rate, and hospital stay (all p > 0.05) to the nonadjacent group.

Conclusion

There were comparable long-term efficacy and complications between patients with HCC adjacent and nonadjacent to the gallbladder treated with MWA.

HIGHLIGHTS

• The application of MWA to HCC adjacent and nonadjacent to the gallbladder resulted in comparable PFS and OS and complications.

• For both cohorts, MWA shared comparable complications (immediate and delayed), LTP, hospitalization, and operative time.

• MWA might be a first-line alternative for ≤5 cm HCC adjacent to the gallbladder with the use of assistive technologies and advances in technology.

Keywords:

• Microwave ablation
• gallbladder
• hepatocellular carcinoma
• propensity score matching

Introduction

According to the latest Global Cancer Statistics 2020, primary liver cancer, a common malignant tumor of the digestive tract, has the sixth highest incidence of malignancies and the third highest fatality rate in the world and is a serious threat to human life and health [1]. As the most common subtype of liver cancer (85% ∼ 90%), hepatocellular carcinoma (HCC) is highly aggressive and heterogeneous [2]. Thus, one of the key issues in HCC management is effective therapy for tumors at an early stage, which is crucial for improving the patient’s prognosis[3]. As an alternative therapy, liver transplantation has been a phased success in the field of liver cancer treatment. However, most patients are unable to undergo liver transplants due to a shortage of donor organs, strong posttransplant rejection, severe complications, and elevated treatment costs [4]. In addition, patients with advanced HCC tend to have a poor general condition and impaired hepatic function due to cirrhosis or portal thrombosis, which makes curative resection impossible [5]. Since 80% of HCC patients are not candidates for liver resection and transplantation, a variety of nonsurgical locoregional interventional therapies have been developed over the past 20 years [6]. Among those, image-guided percutaneous ablation, such as radiofrequency ablation (RFA) and microwave ablation (MWA), has been considered a potentially curative treatment for patients with early-stage HCC because of its advantages of preserving liver function and achieving the effect of similar surgical resection [7,8]. The Barcelona Clinic Liver Cancer (BCLC) 2022 UPDATE suggests ablation as the first treatment approach for HCC patients with very early-stage BCLC 0 [9]. As a new direction in thermal ablation technology, MWA has shown numerous advantages over other technologies, including higher intratumoral temperatures, as well as reduced operation time and electrical conductivity dependence [10–12].

However, thermal ablation remains a controversial treatment option for HCC in high-risk locations due to the risks associated with direct puncture, such as hemorrhage, metastasis, pulmonary gas interference, and unsatisfactory localization. For instance, the use of thermal ablation for HCC adjacent to the gallbladder may result in reversible thermal damage to the gallbladder, perforation, and acute cholecystitis due to heat conduction [13]. Previous studies have demonstrated the efficacy of percutaneous thermal ablation for the treatment of HCC adjacent to the gastrointestinal tract [14], large blood vessels [15], bile ducts [16], gallbladder [17], diaphragm [18], etc. Recent studies have shown that MWA does not differ significantly in terms of technical efficacy, local tumor progression (LTP), or complication rates between high-risk and safe locations [19–21]. However, these studies included several different high-risk sites (gastrointestinal tract, bile duct, gallbladder, diaphragm, etc.). Therefore, confounding factors may have influenced the findings. In addition, only a few single-arm studies have shown that MWA is safe and effective in treating HCC adjacent to the gallbladder [13,22]. However, these studies were single-center, with short follow-up periods and unbalanced essential clinical characteristics of patients in the adjacent group and the nonadjacent group.

Herein, we conducted a multicenter, propensity score matching (PSM) study to compare the long-term efficacy and safety of MWA as first-line therapy for HCC adjacent versus nonadjacent to the gallbladder.

Materials and methods

Patient selection

All patients signed an informed consent form before the initiation of treatment. Owing to patient confidentiality being maintained, the Human Subjects Committee approved this study and waived the requirement for additional informed consent for this retrospective study (identifier number: S2020-464-01), which was in line with the 1975 Helsinki Declaration. We used data from a multicenter database of 2,395 HCC patients who received MWA from January 2006 to December 2018 at five hospitals across the country, from which 657 patients were screened and included in this study. The diagnosis of HCC was based on puncture biopsy (n = 465) or typical imaging findings (n = 192). The inclusion criteria were as follows: (a) number of nodules ≤3 and diameter ≤5 cm; (b) Child–Pugh class A or B; (c) BCLC stage 0 to B; (d) initial onset of HCC; and (e) perioperative laboratory results meeting the operation criteria. The exclusion criteria were as follows: (a) vascular and/or bile duct invasion on imaging, (b) extrahepatic metastasis (EM), (c) ascites, (d) history of malignant tumor of other organs, (e) history of liver transplantation, (f) preoperative chemotherapy and/or transcatheter arterial chemoembolization (TACE), (g) tumor near the gastrointestinal tract and/or diaphragm muscle and/or the liver capsule and/or pericardium, (h) tumors near large blood vessels and/or the hepatic hilum and/or primary to secondary branching bile ducts and (i) presence of both adjacent and nonadjacent gallbladder tumors. The location adjacent to the gallbladder was defined as the position where the tumor was located within 5 mm of the gallbladder. Patients were grouped into the adjacent group (n = 49) and the nonadjacent group (n = 608) according to whether the tumor was adjacent to the gallbladder (Figure 1).

Figure 1. Study flowchart. MWA: microwave ablation; BCLC: Barcelona Clinic Liver Cancer; PSM: propensity score matching.

Ablation equipment and procedures

MWA uses a commercial microwave ablation system (KY2000, Nanjing Kangyou Medical, China) under the guidance of real-time ultrasound, which consists of a 2.4-GHz microwave generator, a flexible coaxial cable, and a cold axis antenna with a diameter of 15 mm to establish an ablation treatment cold axis and generate sufficient ablation areas to surround the target lesion and an ablation edge of at least 5 mm. Before treatment, a detailed plan was developed for each patient, including antenna placement, power output settings, ablation time duration, and ablation zone prediction. When necessary, auxiliary ablation techniques were used, including ethanol injection [23], artificial pleural effusion [24], artificial ascites [14], temperature monitoring [25], and contrast-enhanced ultrasound (CEUS) guidance [26], to improve the visibility of the tumor and avoid damage to essential structures such as the intestinal tract, diaphragm muscle, blood vessels, bile duct, and gallbladder. Hydrodissection, also known as artificial pleural effusion, artificial ascites, and biliary water injection, was also utilized. Under the guidance of ultrasound, the microwave antenna was implanted and confirmed to reach the deep edge of the tumor after local anesthesia with 1% lidocaine. Depending on the thermal field effect of the antenna, the antenna tip was at least 3 mm away from the gallbladder (vertical approach, the antenna was aimed at the gallbladder), and the antenna body was at least 5 mm away (parallel approach, the antenna was placed parallel to the gallbladder). During the procedure, multiple overlapping ablations were performed on the tumor, and the high-echo region was monitored in real time by conventional gray ultrasound. The ablation was terminated when the hyperechoic region created by the overlapping ablation covered the entire tumor and the surrounding hepatic parenchyma along the antenna axis. While the antenna was being removed, the needle channel was routinely ablated to prevent the spread and bleeding of tumor cells. Conscious sedation was achieved with intravenous administration of midazolam (0.06 mg/kg), with continuous vital sign monitoring being performed. After ablation, flumazenil (0.5 mg/kg) was injected intravenously to aid in postoperative recovery from sedation. Antibiotic treatment was given for 2 days to prevent infection after the operation. For those patients with fever, their treatment was continued with extended antibiotic therapy. All treatments were performed by two experienced interventional ultrasound doctors according to the preoperative protocol.

Follow-up care

CEUS or contrast-enhanced magnetic resonance imaging (MRI) was performed within 3 days after the operation to evaluate the integrity and immediate complications of ablation early. If the target tumor was completely necrotic with a sufficient ablative margin, the treatment was regarded as a technical success (TS). If a residual tumor was found, additional therapy was given to achieve complete necrosis of the target tumor. The follow-up period for all patients was calculated from the first MWA treatment. One month after the first MWA treatment, chest X-ray enhanced CT or enhanced MRI and laboratory examination were performed every 3 months for the first 2 years and every 4-6 months thereafter. If the patient was not suitable for undergoing enhanced CT or MRI, CEUS was performed. We assessed postoperative complications based on clinical symptoms, imaging tests, and laboratory records. All postoperative complications were recorded, whereas the occurrence of tumor metastasis was recorded as a delayed event. If a recurrence was detected during follow-up, the best second-line treatment, such as ablation, surgery, TACE, radiotherapy, chemotherapy, or targeted therapy, was used, depending on the characteristics of the recurrence, the patient’s liver function, and the general condition of the patient.

Definitions of ablation terminology and index of therapeutic efficacy

Definitions are based on the standardization by the International Working Group on Image-Guided Tumor Ablation[27]. LTP referred to the occurrence of new tumor lesions in or adjacent to the ablation zone after at least one contrast medium follow-up study has proven complete ablation, intrahepatic metastasis (IDM) was defined as the appearance of new tumors in the liver parenchyma far from the ablation area, and extrahepatic metastasis (EM) was defined as the emergence of new tumors in other organs. One of the primary outcomes was progression-free survival (PFS), which was defined as the interval from the first MWA to the occurrence of any event, such as LTP, IDM, EM, or death. Another primary endpoint of this study was overall survival (OS), which was calculated as the interval from the initial MWA treatment to the death of the patient or the last follow-up before 5 September 2022. The secondary endpoints included TS rate, LTP, complication rates, hospitalization, and operative time. Complications referred to clinical events that resulted in additional treatment interventions, prolonged hospital stays, or unexpected increases in the level of care. We carried out an analysis of immediate and delayed complications using the standard classification for complications published by the Society of Interventional Radiology (SIR) [28]. Patients who were lost to follow-up were censored at the last contact.

Statistical analysis

All statistical analyses were carried out using SPSS software (version 26; IBM Corp., Armonk, NY). The missing values were imputed using various methods depending on the type of data. Logistic regression was used to estimate the propensity score. Known or suspected confounding factors (sex, age-adjusted Charlson codisease index (aCCI), Child–Pugh class, etiology, liver cirrhosis, tumor number, tumor size, aspartate aminotransferase (AST), alanine aminotransferase (ALT), serum albumin (ALB), serum alpha-fetoprotein (AFP), total bilirubin (TBIL), blood glucose (GLU), hemoglobin (HB), platelet count (PLT) and prothrombin time (PT)) were used as independent variables for the logistic regression model fitting, and then the prediction probability was calculated as the propensity score. The grouping variable (if HCC was adjacent to the gallbladder) was used as a dependent variable to fit the model. To reduce the selection bias and confounding factors between the two groups, 1:1 nearest-neighbor matching with an optimal caliper of 0.05 without replacement was used to generate 48 pairs of cases (Figure 1).

For the baseline comparison, continuous variables with normal distribution were statistically described as the mean ± standard deviation and compared by Student’s t test. Nonnormal data were described by median and quartile spacing, and compared with the Mann–Whitney U test. Categorical variables were represented by numbers and/or percentages and analyzed using Pearson’s chi-square test or Fisher’s exact test. The survival curves were plotted using the Kaplan–Meier method and compared using the log-rank test. To assess the independent risk factors for PFS and OS, hazard ratios (HRs) with 95% CIs were estimated using the Cox proportional hazards model. Statistical significance was defined as p < 0.05 (2-sided).

Result

Baseline features

Before PSM, 657 patients were included, of whom 528 were males and 129 were females, and the patients were divided into two groups: adjacent (n = 49) and nonadjacent (n = 608) to the gallbladder. The average age was 57.25 ± 11.33 (range 15-91) years. The median follow-up time for all patients was 61.4 months (range: 0.6-184.5 months). As shown in Table 1, significant differences were observed in four factors (all p < 0.05), including AFP (p = 0.031), HB (p = 0.006), PLT (p = 0.025), and red blood cell (RBC) count (p = 0.011), between the two groups.

Table 1. Baseline characteristics of patients before and after PSM.

	Before PSM			After PSM
Characteristics	Adjacent (n = 49)	Nonadjacent (n = 608)	p value	Adjacent (n = 48)	Nonadjacent (n = 48)	p value
Sex(male/female)	38/11	490/118	0.606	37/11	39/9	0.615
Age(year), Mean ± SD	59.06 ± 11.97	57.10 ± 11.28	0.244	58.65 ± 11.74	62.21 ± 12.63	0.156
Child-pugh (A/B)	49/0	598/10	1.000	47/0	46/1	1.000
Median aCCI [P25,P75]	2(1,3)	2(1,3)	0.232	2(1,3)	2(1,4)	0.151
Etiology(HBV/HCV/NBNC)	35/4/10	424/52/132	1.000	34/4/10	35/5/8	0.893
Cirrhosis, n(%)	41(83.7)	491(80.8)	0.617	40(83.3)	40(83.3)	1.000
Tumor number (1/2/3)	39/9/1	443/120/45	0.335	38/9/1	35/9/4	0.482
Tumor size(cm), Mean ± SD	2.68 ± 0.85	2.53 ± 1.00	0.311	2.64 ± 0.80	2.88 ± 0.17	0.052
Median AFP(P25,P75), ug/L	6.4(3.0,23.5)	10.8(3.9,64.5)	0.031	6.4(3.1,25.3)	8.1(3.4,49.0)	0.577
Median ALT(P25,P75), U/L	25.3(18.7,33.1)	28.8(21.4,41.0)	0.093	25.3(19.8,33.2)	28.8(21.5,40.6)	0.479
Median AST(P25,P75), U/L	28.0(21.3,39.0)	26.5(21.7,38.2)	0.712	28.0(22.7,39.5)	26.5(21.5,43.2)	0.775
Median GGT(P25,P75), U/L	42.4(27.5,75.4)	41.7(29.0,69.3)	0.666	42.4(27.3,72.7)	41.6(29.7,55.4)	0.339
Median TBIL (P25,P75), μmol/L	12.6(9.3,19.8)	14.4(10.9,19.3)	0.334	13.0(9.4,19.9)	15.1(11.3,20.1)	0.305
Median DBIL (P25,P75), μmol/L	4.5(3.4,6.9)	4.9(3.6,7.0)	0.544	4.5(3.3,6.9)	5.1(4.2,6.8)	0.292
Median ALB (P25,P75), g/L	40.1(38.9,41.6)	39.8(37.8,42.5)	0.876	40.1(38.9,41.2)	39.8(36.8,42.0)	0.615
Median GLU (P25,P75), μmol/L	4.8(4.6,5.6)	5.0(4.7,5.8)	0.121	4.8(4.6,5.5)	5.2(4.6,6.1)	0.074
Median Hb (P25,P75), g/L	136(125,141)	139(129,149)	0.006	135(126,141)	139(130,146)	0.143
Median RBC (P25,P75), 10^9/L	4.3(4.1,4.5)	4.4(4.0,4.8)	0.011	4.3(4.1,4.5)	4.4(4.0,4.8)	0.276
Median PLT (P25,P75), 10^9/L	116(81,128)	119(83,156)	0.025	103(80,129)	117(89,143)	0.215
Median PT (P25,P75), s	14.1(13.2,15.1)	13.9(13.2,14.8)	0.566	14.1(13.2,15.1)	13.8(12.8,14.4)	0.089

PSM: propensity score matching; aCCI: age-adjusted Charlson comorbidity index; HBV: hepatitis B virus; HCV: hepatitis C virus; AFP: alpha-fetoprotein; ALT: alanine aminotransferase; AST: aspartate aminotransferase; GGT: gamma-glutamyl transferase; TBIL: total bilirubin; DBIL: direct bilirubin; ALB: albumin; GLU: blood glucose; Hb: hemoglobin; RBC: red blood cell; PLT: platelet; PT: prothrombin time; P₂₅: 25th percentile; P₇₅: 75th percentile.

After PSM, we included 96 patients (76 males and 20 females) who were divided into the adjacent group (n = 48) and nonadjacent group (n = 48), with an average age of 60.43 ± 12.26 years (range 34-91 years). Additionally, the median follow-up time for all patients was 60 (range 6-171) months. After performing PSM, the baseline characteristics were similar between the two groups (Table 1), including age, sex, etiology, number and size of tumors, liver function, and additional laboratory test results (all p > 0.05).

Comparison of treatment results before PSM

Before PSM, the recurrence rates of diseases (including LTP, IDM, EM, and death) in the adjacent and nonadjacent gallbladder groups were 83.7% (41/49) and 68.9% (419/608), respectively. The median PFS of the adjacent group was 27.0 months (95% CI 23.4-30.7), and for the nonadjacent group, it was 33.7 months (95% CI 28.7-38.7). PFS approached statistical significance at 1 year (78% vs. 78%), 3 years (40% vs. 48%), and 5 years (21% vs. 38%) between the two groups (hazard ratio [HR], 1.341; 95% confidence interval [CI], 0.972-1.849; p = 0.073) (Figure 2A). The overall mortality of the adjacent group and the nonadjacent group was 49.0% (24/49) and 29.9% (182/608), respectively. The median OS was 86.1 months (95% CI 47.7-124.5) in the adjacent group and 133.1 months (95% CI 120.6-145.5) in the nonadjacent group. In summary, OS between the two groups approached statistical significance at 1 year (98% vs. 98%), 3 years (80% vs. 85%), and 5 years (66% vs. 76%) (HR 1.517; 95% CI 0.991-2.324; p = 0.054) (Figure 2C). However, further validation is required due to the imbalanced distribution of the subgroup patients in the study.

Figure 2. Kaplan–meier curves of the adjacent gallbladder cohort and nonadjacent gallbladder cohort before and after PSM. (A) PFS before PSM; (B) PFS after PSM; (C) OS before PSM; (D) OS after PSM. PSM: propensity score matching; PFS: progression-free survival; OS: overall survival.

Comparison of treatment results after PSM

After PSM, the recurrence rates of diseases in the adjacent and nonadjacent gallbladder groups were 83.3% (40/48) and 83.3% (40/48), respectively. The median PFS of the adjacent group was 27.4 months (95% CI 22.7-32.1) and that of the nonadjacent group was 25.7 months (95% CI 14.3-37.1). There was comparable PFS at 1 year (77% vs. 77%), 3 years (41% vs. 37%), and 5 years (22% vs. 30%) between the two groups (HR 1.011; 95% CI 0.647-1.578; p = 0.963) (Figure 2B). Both the adjacent group and the nonadjacent group had an overall mortality rate of 50.0% (24/48). The median OS was 86.1 months (95% CI 48.7-123.5) in the adjacent group and 75.6 months (95% CI 54.8-96.3) in the nonadjacent group. There was also comparable OS between the two groups at 1 year (98% vs. 96%), 3 years (80% vs. 87%), and 5 years (65% vs. 67%) (HR 0.925; 95% CI 0.522-1.639; p = 0.789) (Figure 2D).

Analysis of risk factors affecting treatment outcome after PSM

In both univariate and multivariate analyses, whether the tumor was adjacent to the gallbladder was not found to be an independent risk factor for PFS (p = 0.963) or OS (p = 0.789). Specifically, in the univariate analysis, the results indicated that both PLT and AST were significant risk factors for PFS (both p < 0.05) and that age, aCCI, HBV, AST, GGT, DBIL, ALB, GLU, RBC, and PLT were significant risk factors for OS (all p < 0.05). Additionally, the multivariate analysis identified AST as a significant risk factor for DFS (p < 0.05), as well as GGT and ALB for OS (both p < 0.05) (Tables 2 and 3).

Table 2. Univariate and multivariate analyses of PFS after PSM.

	Univariate analysis		Multivariate analysis
Characteristics	HR (95% CI)	p value	HR (95% CI)	p value
Sex (male)	1.320(0.759,2.295)	0.326
Age (≥65 year)	0.810(0.508,1.291)	0.375
aCCI (>3)	1.169(0.688, 1.987)	0.563
Etiology (HBV)	1.108(0.663,1.851)	0.695
Etiology (HCV)	0.794(0.345,1.832)	0.589
Cirrhosis (yes)	1.158(0.623,2.151)	0.643
Tumor number (>1)	1.112(0.687, 1.798)	0.666
Tumor size (>3.0 cm)	0.983(0.550,1.756)	0.953
Adjacent to gallbladder(yes)	1.011(0.647,1.578)	0.963
AFP (>200μmg/L)	1.827(0.960,3.476)	0.066
ALT (>40U/L)	1.608(0.961,2.690)	0.070
AST (>40U/L)	1.727(1.054,2.831)	0.030	1.727(1.054,2.831)	0.030
GGT (>50 U/L)	1.407(0.881,2.247)	0.153
TBIL (>21μmol/L)	1.026(0.598,1.759)	0.926
DBIL (>8.6 μmol/L)	1.488(0.763,2.751)	0.258
ALB (>35 g/L)	0.604(0.317,1.151)	0.126
GLU (>6.1 mmol/L)	1.562(0.907,2.690)	0.107
Hb (<120 g/L)	1.178(0.659,2.104)	0.580
RBC (<4.3 × 10^9/L)	1.099(0.698,1.731)	0.682
PLT (<100 × 10^9/L)	1.604(1.012,2.542)	0.044
PT (<13s)	0.904(0.548,1.491)	0.692
First technical success	1.211(0.739,1.984)	0.448
Completely ablation (no)	1.613(0.647,4.020)	0.305
Temperature monitoring (yes)	0.867(0.520,1.446)	0.585
Ethanol injection (yes)	1.240(0.765,2.009)	0.382
Hydrodissection (yes)	1.255(0.542,2.906)	0.595
CEUS (yes)	0.956(0.579,1.581)	0.862

DFS: disease-free survival; PSM: propensity score matching; HR: hazard ratio; CI: confidence interval; aCCI: age-adjusted Charlson comorbidity index; BCLC: Barcelona Clinic Liver Cancer; HBV: hepatitis B virus; HCV: hepatitis C virus; AFP: alpha-fetoprotein; ALT: alanine aminotransferase; AST: aspartate aminotransferase; GGT: gamma-glutamyl transferase; TBIL: total bilirubin; DBIL: direct bilirubin; ALB: albumin; GLU: blood glucose; Hb: hemoglobin; RBC: red blood cell; PLT: platelet; PT: prothrombin time; CEUS: contrast-enhanced ultrasound.

Table 3. Univariate and multivariate analyses of OS after PSM.

	Univariate analysis		Multivariate analysis
Characteristics	HR (95% CI)	p value	HR (95% CI)	p value
Sex (male)	1.165(0.580,3.342)	0.668
Age (≥65 year)	1.906(1.062,3.420)	0.031	1.977(1.095,3.570)	0.024
aCCI (>3)	2.341(1.254, 4.370)	0.008
Etiology (HBV)	0.527(0.288,0.962)	0.037
Etiology (HCV)	1.348(0.527,3.450)	0.533
Cirrhosis (yes)	0.570(0.281,1.157)	0.120
Tumor number (>1)	1.599(0.892, 2.865)	0.115
Tumor size (>3.0 cm)	1.037(0.499,2.155)	0.922
Adjacent to gallbladder(yes)	0.925(0.522,1.639)	0.789
AFP (>200μmg/L)	0.961(0.406,2.275)	0.928
ALT (>40U/L)	1.565(0.797,3.076)	0.194
AST (>40U/L)	2.374(1.298,4.340)	0.005
GGT (>50 U/L)	2.015(1.120,3.626)	0.019	1.930(1.063,3.504)	0.031
TBIL (>21μmol/L)	1.273(0.627,2.584)	0.504
DBIL (>8.6 μmol/L)	2.409(1.146,5.061)	0.020
ALB (>35 g/L)	0.242(0.116,0.504)	0.000	0.258(0.123,0.543)	0.000
GLU (>6.1 mmol/L)	2.086(1.070,4.064)	0.031
Hb (<120 g/L)	1.802(0.932,3.485)	0.080
RBC (<4.3 × 10^9/L)	1.862(1.051,3.298)	0.033
PLT (<100 × 10^9/L)	2.078(1.172,3.684)	0.012
PT (<13s)	0.716(0.332,1.544)	0.394
First technical success	1.166(0.614,2.216)	0.639
Completely ablation (no)	22.404(0.103,4876.800)	0.258
Local tumor progression (yes)	1.090(0.429,2.769)	0.856
Intrahepatic metastasis (yes)	0.904(0.456,1.790)	0.771
Extrahepatic metastasis (yes)	1.544(0.845,2.822)	0.158
Temperature monitoring (yes)	1.219(0.665,2.237)	0.522
Ethanol injection (yes)	0.820(0.619,2.691)	0.192
Hydrodissection (yes)	1.090(0.336,3.540)	0.885
CEUS (yes)	1.334(0.727,2.446)	0.352

OS: overall survival; PSM: propensity score matching; HR: hazard ratio; CI: confidence interval; aCCI: age-adjusted Charlson comorbidity index; BCLC: Barcelona Clinic Liver Cancer; HBV: hepatitis B virus; HCV: hepatitis C virus; AFP: alpha-fetoprotein; ALT: alanine aminotransferase; AST: aspartate aminotransferase; GGT: gamma-glutamyl transferase; TBIL: total bilirubin; DBIL: direct bilirubin; ALB: albumin; GLU: blood glucose; Hb: hemoglobin; RBC: red blood cell; PLT: platelet; PT: prothrombin time; CEUS: contrast-enhanced ultrasound.

Subgroup analysis after PSM

Tumor size plays a significant role in both the curability and prognosis of HCC patients [29]. Therefore, the present study conducted a subgroup analysis based on tumor size (<3 cm or 3-5 cm) to explore the impact of HCC adjacent to the gallbladder on patients’ PFS and OS. After PSM, in the subgroup of tumor size <3 cm (65 cases), the PFS of the adjacent group was not significantly different from that of the nonadjacent group, with a median PFS of 27.0 months (95% CI 22.7-31.4) and 24.0 months (95% CI 16.6-31.4), respectively. Both groups had a similar PFS rate at 1, 3, and 5 years (76% vs. 78%, 41% vs. 40%, and 23% vs. 33%, respectively), with an HR of 1.074 (95% CI 0.616-1.875; p = 0.800) (Figure 3A). Similarly, no significant difference was found in OS between the adjacent and nonadjacent groups in the <3 cm subgroup, with a median OS of 86.1 months (95% CI 61.8-110.4) and 114.9 months (95% CI 51.1-178.6), respectively. The OS rates at 1, 3, and 5 years were also comparable between the two groups (100% vs. 94%, 87% vs. 90%, and 71% vs. 71%, respectively), with an HR of 1.087 (95% CI 0.523-2.262; p = 0.822) (Figure 3C).

Figure 3. Kaplan–meier curves of HCC patients with tumor <3 cm (A, C) and 3-5 cm (B, D) after PSM. (A)PFS of HCC patients with tumor <3cm; (B)PFS of HCC patients with tumor 3-5 cm; (C) OS of HCC patients with tumor <3 cm; (D) OS of HCC patients with tumor 3-5 cm. HCC: hepatocellular carcinoma; PSM: propensity score matching; PFS: progression-free survival; OS: overall survival.

In the subgroup with a tumor diameter of 3-5 cm (31 cases), the median PFS of the adjacent group was 27.4 months (95% CI 13.3-41.5), while that of the nonadjacent group was 26.6 months (95% CI 6.3-46.9). There was no statistically significant difference in PFS between the adjacent and nonadjacent groups at 1 year (80% vs. 75%), 3 years (40% vs. 31%), and 5 years (20% vs. 24%) (HR 0.792; 95% CI 0.363-1.725; p = 0.556) (Figure 3B). The median OS of adjacent patients was 64.0 months (95% CI 0-136.0), whereas that of patients in the nonadjacent group was 69.9 months (95% CI 37.9-101.9). Our study revealed similar OS between the adjacent and nonadjacent groups at 1 year (93% vs. 100%), 3 years (67% vs. 81%), and 5 years (53% vs. 60%) (HR 0.804; 95% CI 0.319-2.025; p = 0.643) (Figure 3D).

Safety profile, ablation technology, and intervention time after PSM

As shown in Table 4, the use of assistive technologies in the adjacent group was significantly higher than that in the nonadjacent group (p < 0.001). These technologies included temperature monitoring (22 vs. 3), alcohol injection (23 vs. 4), hydrodissection (7 vs. 1), and CEUS (16 vs. 13). Although the use of assistive technologies was different, there were no significant differences between the adjacent group and nonadjacent group (all p > 0.05) in terms of the TS rate (89.6% vs. 97.9%, p = 0.204), incidence of LTP (10.4% vs. 8.3%, p = 1.000), ablation duration (median 600 s vs. 600 s, p = 0.965) or postoperative hospital stay (median 6 days vs. 6 days, p = 0.581).

Table 4. Perioperative treatment status of patients in the adjacent and nonadjacent groups after PSM.

Variables	Adjacent (n = 48)	Nonadjacent (n = 48)	p value
Assistive technologies, n (%)	35(72.9)	15(31.3)	0.000
Temperature monitoring, n (%)	22(46.8)	3(6.3)	0.000
Ethanol injection, n (%)	23(47.9)	4(8.3)	0.000
Hydrodissection, n (%)	7(14.6)	1(2.1)	0.059
Contrast-enhanced ultrasound, n (%)	16(33.3)	13(27.1)	0.505
Technique success rate, n (%)	43(89.6)	45(97.9)	0.204
Local tumor progression, n (%)	5(10.4)	4(8.3)	1.000
Complications, n (%)	29(60.4)	29(60.4)	1.000
Immediate complications	29(60.4)	29(60.4)	1.000
Abdominal pain, n (%)	13(27.1)	11(22.9)	0.637
NRS pain score, Median (P25,P75)	0(0,1)	0(0,0)	0.549
Days of abdominal pain, Median (P25,P75)	0(0,1)	0(0,0)	0.937
Fever, n (%)	23(47.9)	26(54.2)	0.540
Maximum temperature, Median (P25,P75)	37.2(36.5,38.4)	37.6(36.4,38.2)	0.909
Days of fever, Median (P25,P75)	0(0,3)	1(0,2)	0.789
Delayed complications	2(4.2)	0(0)	0.495
Abscess, n (%)	1(2.1)	0	1.000
Needle tract implantation metastases, n (%)	1(2.1)	0	1.000
Operative time, Median Seconds (P25,P75)	600(398,915)	600(360,960)	0.965
Hospital stay, Median Days (P25,P75)	6(4,8)	6(5,8)	0.581

PSM: propensity score matching; NRS: numeric rating scale; P₂₅: 25th percentile; P₇₅: 75th percentile.

Table 4 shows that there was no significant difference in the incidence of immediate complications (60.4% vs. 60.4%, p = 1.000) or delayed complications(4.2% vs. 0%, p = 0.495), including abdominal pain (27.1% vs. 22.9%, p = 0.637), fever (47.9% vs. 54.2%, p = 0.540), abscess (2.1% vs. 0%, p = 1.000), and needle tract implantation metastases (2.1% vs. 0%, p = 1.000), between the adjacent group and the nonadjacent group. Similarly, there were no significant differences in the intensity (p = 0.549) and duration (p = 0.937) of abdominal pain or the highest temperature (p = 0.909) or duration (p = 0.789) of fever between the two groups (all p > 0.05).

Discussion

Ultrasound-guided percutaneous thermal ablation is an increasingly popular and effective minimally invasive treatment for liver tumors. Among the available thermal ablation technologies, MWA is particularly popular due to the ‘heat sink’ effect being less of an issue and a more uniform ablation zone in both shape and size [30,31]. However, the guidelines do not explicitly recommend MWA for the treatment of HCCs in high-risk sites due to a lack of sufficient evidence [9,32,33]. Previous studies have shown that tumors located at high-risk sites were independent risk factors for LTP and were associated with the prognosis of HCC patients [34,35]. Moreover, smooth puncture and thermal ablation of tumors adjacent to critical structures are challenging procedures, as they may result in incomplete necrosis or additional damage to the surrounding organs [36]. Therefore, it is crucial to explore the efficacy and safety of MWA for treating HCC in high-risk locations, such as those adjacent to the gallbladder.

Percutaneous ablation for HCCs that are adjacent to the gallbladder is a controversial procedure for the following reasons: (1) there are risks of gallbladder perforation and acute cholecystitis accompanying thermal ablation of tumors adjacent to the gallbladder [37], and (2) to avoid thermal damage to the gallbladder, complete ablation of the tumor and the surrounding 5-10 mm of the hepatic parenchyma are difficult to achieve, which may increase the incidence of LTP [38–40]. Although a gallbladder injury is rarely life-threatening, it can be somewhat painful for the patient. With the rapid development of ablation techniques, guidance methods, and auxiliary technologies, recent studies have shown that MWA is a safe and effective treatment for HCCs near the gallbladder [13,22]. However, these studies were single-center with short follow-up periods and did not balance the basic clinical characteristics of patients in the adjacent and nonadjacent groups, which made it difficult to draw definitive conclusions. It is vital to further evaluate the efficacy and safety of MWA for the treatment of HCCs adjacent to the gallbladder, particularly in multicenter studies with longer follow-up periods and better-balanced patient selection criteria.

Our study used multicenter data with longer follow-up periods and used PSM balance to reduce selection bias and confounding factors between the two study groups. To the best of our knowledge, our study represents the first multicenter PSM analysis to compare the therapeutic outcomes of MWA as a first-line therapy in patients with small HCCs located adjacent to the gallbladder and those with nonadjacent HCCs. Our research showed that the position of the tumor adjacent to the gallbladder led to the application of adjuvant technology more frequently (p < 0.05), but the position near the gallbladder had no significant effect on the long-term efficacy of MWA in the treatment of HCCs (PFS or OS), the incidence of adverse events, the efficiency of the ablation technique, or postoperative hospital stay (all p > 0.05). Previous studies have documented that the size and location of HCC critically affect the possibility and difficulty of radical cure [7]. Therefore, after PSM, a subgroup analysis was carried out according to tumor diameters <3 cm and 3-5 cm. The results showed that there was no significant difference in PFS or OS between the adjacent group and the nonadjacent group (all p > 0.05).

Tumors adjacent to the gallbladder were not found to be an independent risk factor for PFS (p = 0.963) or OS (p = 0.789). Furthermore, there was no significant difference in the TS rate (p = 0.204) between the adjacent group and the nonadjacent group, which may be related to the use of more auxiliary techniques in the adjacent group. The combination of multiple adjuvant techniques, including ethanol injection, hydrodissection, temperature monitoring, and CEUS, offers greater possibilities for the radical treatment of HCCs adjacent to the gallbladder. Ethanol injection was used to chemically ablate tumor edge cells to achieve complete necrosis of the tumor nodule [23]. Hydrodissection and real-time temperature monitoring technology can avoid thermal damage to adjacent vital organs during ablation [25,41]. CEUS can show lesions near the gallbladder more clearly than conventional ultrasound. In addition to using assistive technology, it may also be related to the use of straight needle electrodes to improve visibility and control of the tip [42]. The two groups had a comparable complication rate, including immediate and delayed complications (p > 0.05). Therefore, our results indicated that MWA could be a first-line alternative for ≤5 cm HCC adjacent to the gallbladder with the use of advanced assistive technologies.

Despite the major advantages of our study, including the multicenter research design and the use of PSM to balance baseline characteristics, there are several limitations. First, potential selection and indication bias are inevitable due to the retrospective nature of the study design. Moreover, the different guiding methods used, including CT and US from multiple medical centers, may have contributed to the artificial discrepancies in ablation techniques. Finally, pharmacological therapies such as angiotensin receptor blockers may affect patient outcomes after ablative treatment for HCC [43]. Therefore, a large prospective multicenter cohort study is needed to further validate these conclusions.

Conclusion

In summary, this 12-year multicenter study provided strong evidence for the efficacy of percutaneous MWA in treating HCCs adjacent to the gallbladder, with the longest follow-up period. With appropriate patient selection, preoperative planning, and assistive technology, there was no significant difference in the long-term effects (PFS and OS) and the incidence of complications of percutaneous MWA as first-line treatment of HCCs adjacent and nonadjacent to the gallbladder.

Article category

Research Articles (Cancer therapy and prevention).

Abbreviations
MWA	=	microwave ablation
HCC	=	hepatocellular carcinoma
PSM	=	Propensity score matching
GB	=	gallbladder
OS	=	overall survival
PFS	=	progression-free survival
LTP	=	local tumor progression
BCLC	=	Barcelona Clinic Liver Cancer
RFA	=	radiofrequency ablation
EM	=	extrahepatic metastasis
TACE	=	transcatheter arterial chemoembolization
CEUS	=	contrast-enhanced ultrasound
MRI	=	magnetic resonance imaging
TS	=	technical success
IDM	=	intrahepatic metastasis
aCCI	=	age-adjusted Charlson codisease index
AST	=	aspartate aminotransferase
ALT	=	alanine aminotransferase
ALB	=	albumin
AFP	=	alpha-fetoprotein
TBIL	=	total bilirubin
GLU	=	glucose
HB	=	hemoglobin
PLT	=	platelet count
PT	=	prothrombin time
RBC	=	red blood cell

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

We understand the terms of the share upon reasonable request data policy, and our data generated or analyzed during the study are available from the corresponding author by request.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.