Local ablation therapy (LAT), such as radiofrequency ablation (RFA), microwave ablation, percutaneous ethanol injection, and cryoablation, is primarily recommended for solitary hepatocellular carcinoma (HCC) ≤3 cm in diameter, especially when surgical resection or liver transplantation cannot be considered.^1-5 However, only 60% of uninodular small HCCs are suitable for LAT, despite recent advances in procedures and imaging techniques.^6,7 The major reasons for this are: (1) inability to distinguish tumor from surrounding liver parenchyma or cirrhotic nodules;⁸ (2) a potentially inadequate ablative zone with respect to the heat sink effect;⁹ (3) possibility of thermal damage to adjacent anatomic structures, including diaphragm and bowel;⁹ (4) inappropriateness of needle passage through nearby organs;⁹ and (5) hemostatic problems associated with cirrhosis-related coagulopathy.⁹

If an HCC of ≤3 cm that is theoretically indicated for LAT is not in practice ablatable, transarterial chemoembolization (TACE) is considered the next-best option by current guidelines.^1-5 The 5-year survival rate of TACE in patients with early-stage HCC is reported to be 40% to 70%, which may not be inferior to that of LAT.^10-13 However, TACE is still classified as a modality with palliative intent and not formally recognized as an alternative strategy to LAT. Radiotherapy (RT) is another method used for locoregional control of small HCCs, and cumulative evidence ranks RT close to LAT in the hierarchy for therapeutic decision-making, with promising outcome data from multiple comparisons.^14-16 Nevertheless, RT is not widely recommended as an alternative to LAT and tends to be underestimated even as an experimental treatment.

In order to establish the second-best treatment for small HCCs, we aimed to investigate and compare the effects of TACE and RT on tumor progression and survival in patients with single small HCCs considered unsuitable for LAT after planning sonography.

1. Study samples

A total of 804 patients with single HCCs of ≤3 cm who underwent pre-procedure planning for LAT using sonography at the Asan Medical Center, South Korea, between January 2017 and December 2019 were consecutively enrolled in this retrospective study. None of the patients experienced extrahepatic spread or vascular invasion of the tumor. HCC diagnosis was based on contrast-enhanced dynamic imaging or histologic confirmation.^1,3-5,17 The following individuals were excluded: (1) 98 patients lost to follow-up immediately after sonographic planning or after subsequent treatment; (2) two patients who had undergone liver transplantation prior to planning sonography; (3) 47 patients who underwent hepatectomy after planning sonography; (4) four patients with Child-Pugh class C liver function; and (5) two patients with histories of other malignancies (Fig. 1). Finally, 801 nodules were analyzed in 651 patients.

Figure 1. Flowchart of the treatment of 651 patients with single HCCs of ^≤3 cm meeting the predefined criteria. The numbers for the treatments in the boxes indicate the numbers of patients receiving the treatments after each planning sonography. HCC, hepatocellular carcinoma; LAT, local ablation therapy; TACE, transarterial chemoembolization; RT, radiotherapy.

Of these 651 patients, 121 had two or more nodules at different times, at least 1 month apart, over the study period, all of which met the above-mentioned inclusion and exclusion criteria: two nodules in 97 patients, three in 19 patients, and four in five patients. Of the nodules detected, 619 were treatment-naïve and 182 were recurrent or residual (Fig. 1). The Institutional Review Board of Asan Medical Center approved this study (IRB number: 2021-0416) and waived the informed consent of individual patients.

2. Definition of treatment groups

As per the nodule-based analyses, the LAT group included 561 tumors treated with RFA (n=552), microwave ablation (n=4), percutaneous ethanol injection (n=4), or cryotherapy (n=1). The TACE group included 177 tumors treated with conventional TACE and eight treated with drug-eluting bead TACE. The RT group included 55 HCC lesions, consisting of 48 tumors treated with stereotactic body RT and seven with hypo-fractionated RT. For the patient-based analyses, the patients were categorized according to the treatment modality received after the initial planning sonography. The choice of second-line treatment in patients with unablatable HCCs was based on the location and enhancing pattern of the tumor, adjacent organs, and biliary-vascular anatomy.

3. Treatment procedures

During the pre-procedural planning sonography, an operator determined the feasibility of ablation for the index tumor based on the following criteria: conspicuity of the index tumor, presence of a safe path, and prospect of collateral thermal damage to adjacent organs. The procedures were carried out in accordance with the standard protocols of our institution.^11,14,18-20 During LAT, the patients were administered local or general anesthetic, and a radiologist advanced a single electrode with a 3 to 3.5 cm exposed tip to the focal HCC lesion under guidance of an imaging modality, primarily ultrasonography. The target lesion was ablated for 12 to 15 minutes at maximal power using a 200 W generator in the impedance control mode until complete removal of the visible tumor and 0.5 to 1 cm of the surrounding normal parenchyma.^11,18 Before TACE, arteriography was performed to determine the anatomy of the superior mesenteric and celiac arteries and the vascularity of the HCC. We initially infused 2 mg/kg of cisplatin or 50 mg of adriamycin for 15 minutes and then injected 5 to 10 mL of iodized oil into the feeding artery until it reached the portal branches. Finally, absorbable gel-foam sponge particles (1-mm diameter) were inserted to embolize the feeding artery.¹⁹ For drug-eluting bead TACE, DC bead 100–300 µm in diameter (Boston Scientific, Marlborough, MA, USA) or HepaSphere 20–40 µm in diameter (Merit Medical, South Jordan, UT, USA) loaded with chemo-agents was slowly loaded until near stasis.²¹ TACE was performed in two or three sessions if a viable portion of the target lesion remained without any evidence of disease progression. Lastly, four-dimensional free-breathing computed tomography images were obtained and sorted into 10 computed tomography series corresponding to the respiratory cycle in the RT procedures. The target tumor volume included the whole HCC lesion along with the surrounding 0.5 cm of normal liver parenchyma. Stereotactic body RT was performed using a three-dimensional RT planning system, with a total dose of 45 Gy delivered over three fractions.¹⁴ For hypo-fractionated RT, a total dose of 25 to 50 Gy was delivered over 10 fractions to reduce liver and gastrointestinal tract damage.²⁰ All the above treatments were performed by radiologists or radiation oncologists with ≥7 years of experience in the individual procedures.

4. Assessment of response, and study endpoints

Treated patients were regularly evaluated by dynamic liver computed tomography or magnetic resonance imaging at intervals of 1 to 3 months for the first 2 years after treatment, and subsequently at intervals of 3 to 6 months, until tumor recurrence or progression occurred. Recurrent or progressive lesions were treated based on standard Korean practice guidelines.⁴ The primary outcome of this study was local tumor progression (LTP)-free survival of the target lesion after individual treatment. Progression of the target lesion was defined as a ≥20% increase in the longest dimension of the treated lesion, or appearance of a new intrahepatic lesion adjacent to the treated zone based on the Response Criteria in Solid Tumors Criteria version 1.1 (RECIST 1.1). The uptake of iodized oil in the tumor treated with TACE was calculated and expressed as tumor diameter.²² Overall progression (OP) was defined as appearance of a new lesion or progression of an existing lesion. LTP-, or OP-free survival was calculated from the date of the index treatment to the date of progression of the target or any lesion, or of the last dynamic imaging evaluation. Responses were initially evaluated by two abdominal radiologists with ≥6 years of experience, along with retrospective review by three hepatologists (J.L., E.K., and J.H.S). Overall survival (OS) was measured from the date of treatment performed immediately after the first planning procedure to the date of death from any cause, liver transplantation, or last follow-up.

5. Clinical and pathological variables

The following baseline variables were collected from the electronic medical records at the time of the first planning: sex, age, diabetes, hypertension, body mass index, alcohol consumption, smoking status, and liver cirrhosis based on imaging or histological findings, and laboratory data (platelet count, prothrombin time, aspartate aminotransferase level, alanine aminotransferase level, total bilirubin level, albumin level, creatinine level, and alpha-fetoprotein level). The characteristics of the included HCC nodules were recorded at the time of each sonographic planning event.

6. Statistical analysis

Continuous variables are expressed as mean±standard deviation and were analyzed using the t-test or Mann-Whitney U test. Categorical variables are expressed as numbers with percentages and were analyzed using the chi-square test or Fisher exact test. Cumulative incidences of LTP-, and OP-free survival, and OS were estimated by the Kaplan-Meier method and compared using the log-rank test. Cox proportional hazards models were used to identify the risk factors for each outcome in the study cohort. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated. Propensity score (PS) matching and inverse probability treatment weighting (IPTW) were employed to minimize confounders between the TACE-treated and RT-treated groups. PSs were calculated using logistic regression models with the following variables as predictors: sex, age, Child-Pugh class, nodule size, alanine aminotransferase level, and history of previous treatment per person and per nodule. Weights were calculated as the reciprocal of the probability of receiving RT. They were stabilized and truncated at the 1st and 99th percentiles, with the overall weights ranging from 0.48 to 2.29, using 1:2 greedy nearest neighbor matching without replacement within the specified caliper widths. The Wald test with robust variance estimation was used to compare clinical outcomes in the matched sets. For the PS-matched set, a Cox proportional hazards regression model with robust variance estimator was used to account for clustering. For the IPTW set, an inverse probability weighted Cox proportional hazards regression model was used. A p-value of <0.05 was considered statistically significant. Statistical analysis was conducted using R software, version 3.6.0 (R Foundation for Statistical Computing, Vienna, Austria).

1. Baseline characteristics of patients and tumors

Patients’ baseline characteristics by treatment modality after planning sonography are summarized in Table 1. The mean age of the patients was 61.3±9.4 years, and 504 (77.4%) were male, of whom 426 (65.4%) were infected with hepatitis B virus. Over one-third of the total of 651 patients had received treatment for HCC prior to planning, TACE (28.3%) being the most frequent method. Of the patients, 467 (71.7%), 148 (22.7%), and 36 (5.5%) were treated with LAT, TACE, and RT, respectively, after the first planning event during the study period. In the RT group, 77.8% of the patients had previously received HCC treatment, compared with only 39.2% of the patients in the LAT group (p<0.001). Most of the other pre-planning parameters were similar across treatment groups.

Table 1

Pre-Procedure Characteristics of Patients According to Treatment Modality.

Variable	Total (n=651)	LAT (n=467)	TACE (n=148)	RT (n=36)	p for trend
Age, yr	61.3±9.4	61.1±9.6	61.7±8.4	62.8±10.0	0.491
Male sex	504 (77.4)	363 (77.7)	114 (77.0)	27 (75.0)	0.923
Hypertension	258 (39.6)	184 (39.4)	61 (41.2)	13 (36.1)	0.839
Diabetes	310 (47.6)	231 (49.5)	60 (40.5)	19 (52.8)	0.136
Body mass index, kg/m2	25.1±3.6	25.1±3.6	24.8±3.5	26.1±3.6	0.177
Etiology of liver disease					0.473
HBV infection	426 (65.4)	311 (66.6)	94 (63.5)	21 (58.3)
HCV infection	45 (6.9)	32 (6.9)	12 (8.1)	1 (2.8)
Other	180 (27.6)	124 (26.6)	42 (28.4)	14 (38.9)
Platelets, ×103/μL	14.4±10.8	13.9±9.1	16.3±15.7	13.4±1.7	0.050
PT, INR	1.1±0.1	1.1±0.1	1.1±0.1	1.1±0.1	0.771
AST, IU/L	47.2±43.7	49.4±46.4	40.9±38.1	44.1±21.5	0.105
ALT, IU/L	33.3±27.7	34.6±29.5	29.4±22.8	32.4±19.3	0.134
Bilirubin, mg/dL	0.9±0.6	0.9±0.5	1.0±0.8	0.9±0.5	0.030
Albumin, g/dL	3.7±0.5	3.7±0.5	3.6±0.5	3.7±0.5	0.896
Creatinine, mg/dL	0.9±0.8	0.9±0.9	0.9±0.5	1.1±1.1	0.307
Serum AFP, ng/mL	128.8±1,093	100.4±955.5	211.1±1,493	158.9±795.8	0.555
Liver cirrhosis	423 (65.0)	296 (63.4)	104 (70.3)	23 (63.9)	0.307
Child-Pugh score	5.8±0.9	5.9±0.9	5.8±1.0	5.7±0.9	0.331
Previous treatment history of the patient*	255 (39.2)	159 (34.0)	68 (45.9)	28 (77.8)	<0.001
LAT	38 (5.8)	20 (4.3)	13 (8.8)	5 (13.9)
TACE	184 (28.3)	114 (24.4)	47 (31.8)	23 (63.9)
Radiotherapy	33 (5.1)	19 (4.1)	8 (5.4)	6 (16.7)
Surgical resection	51 (7.8)	34 (7.3)	15 (10.1)	2 (5.6)

Data are presented as mean±SD or number (%)..

LAT, local ablation therapy; TACE, transarterial chemoembolization; RT, radiotherapy; HBV, hepatitis B virus; HCV, hepatitis C virus; PT, prothrombin time; INR, international normalized ratio; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein..

*It represents the patients with any anti-hepatocellular carcinoma treatment even once..

Of the HCC nodules with a mean size of 1.5±0.6 cm, 610 nodules (76.2%) were located in the right lobe (segments 5, 6, 7, and 8), 189 (23.6%) in the left lobe (segments 2, 3, and 4), and two (0.2%) in segment 1 (Supplementary Table 1). LAT was the most frequently used modality for HCCs with involvement of both lobes of the liver. Of the patients with right lobe nodules, 74.3% were treated with LAT, 20.3% with TACE, and 5.4% with RT. Of those with tumors in the left lobe, 56.6%, 32.3%, and 11.1% were treated with LAT, TACE, and RT, respectively. Of the two segments 1 HCCs, one was treated with LAT and the other with RT. Meanwhile, 45.5% of the RT-treated nodules had previously received HCC treatment (p<0.001), and they had the smallest mean diameter of 1.3±0.5 cm (p=0.043).

Of the total of 801 nodules, 561 were treated with LAT, whereas TACE and RT were used to treat 185 and 55 nodules, respectively (Supplementary Table 2). Reasons that LAT was considered unsuitable were: (1) inability to position the ablation catheter in an effective manner because the lesion could not be detected (n=131, 54.6%); (2) proximity to large vessels or bile ducts (n=37, 15.4%); (3) subcapsular tumor location (n=19, 7.9%); (4) location adjacent to other organs, such as gallbladder, colon, esophagus, lung, or heart (n=32, 13.3%); and (5) other factors (n=21, 8.8%). No significant difference was observed between the TACE and RT groups in terms of the causes of LAT unsuitability (p=0.439).

2. Per nodule-based analyses

1) LTP-free survival of target lesions and OP-free survival in the entire cohort of nodules

The 801 treated nodules were followed up for a median duration of 20.6 months (range, 8.7 to 31.9 months). During the observation period, 124 target and 359 non-target lesions progressed or recurred. The 1- and 3-year cumulative rates of LTP-free survival were 8.0% and 17.1%; 15.0% and 42.1%; and 7.9% and 13.9% in the LAT, TACE, and RT groups, respectively (p<0.001) (Fig. 2A). Similar trends for Kaplan-Meier estimates were observed for OP-free survival (p<0.001) (Fig. 2B). The multivariate Cox regression analysis revealed that TACE (adjusted HR, 2.36; 95% CI, 1.61 to 3.47; p<0.001), high serum bilirubin level (adjusted HR, 1.44; 95% CI, 1.12 to 1.85; p=0.005), and previous history of HCC treatment (adjusted HR, 1.70; 95% CI, 1.09 to 2.63; p=0.018) were significant risk factors for LTP-free survival (Table 2). TACE was also an independent predictor (adjusted HR, 1.64; 95% CI, 1.30 to 2.07; p<0.001) (Supplementary Table 3) in the multivariate analysis of OP-free survival. The treatment-naïve tumors yielded the same results for LTP-free survival (p<0.001) and OP-free survival (p<0.001) (Supplementary Tables 4 and 5, Supplementary Fig. 1).

Figure 2. Kaplan-Meier curves for (A) LTP-free survival and (B) OP-free survival after LAT, TACE, and RT in all patients (801 hepatocellular carcinoma nodules). LTP, local tumor progression; OP, overall progression; LAT, local ablation therapy; TACE, transarterial chemoembolization; RT, radiotherapy.

Table 2

Analysis of Factors Affecting Local Tumor Progression-Free Survival in All Patients (801 Nodules).

Variable	Univariate analysis			Multivariable analysis
	HR (95% CI)	p-value		HR (95% CI)	p-value
Age	1.01 (0.99–1.03)	0.342
Male sex	1.27 (0.81–1.99)	0.290
Hypertension	1.06 (0.74–1.51)	0.758
Diabetes	0.92 (0.64–1.31)	0.635
Body mass index	1.03 (0.98–1.08)	0.188
Etiology of liver disease
HBV infection	1 (reference)
HCV infection	1.46 (0.76–2.83)	0.256
Others	1.05 (0.70–1.57)	0.802
Platelets	0.91 (0.83–1.00)	0.057
PT INR	2.77 (0.89–8.65)	0.079
AST	1.00 (1.00–1.01)	0.589
ALT	1.00 (1.00–1.01)	0.491
Bilirubin	1.47 (1.12–1.93)	0.005		1.44 (1.12–1.85)	0.005
Albumin	0.85 (0.60–1.21)	0.373
Creatinine	0.98 (0.74–1.29)	0.882
Serum AFP	1.00 (1.00–1.00)	0.039		1.00 (1.00–1.00)	0.079
Cirrhosis	0.98 (0.68–1.41)	0.904
Previous treatment history of the patient	1.85 (1.29–2.66)	0.001		1.70 (1.09–2.63)	0.018
Previous treatment history of the target nodule	1.68 (1.15–2.47)	0.008		1.24 (0.79–1.96)	0.355
Tumor location (hepatic segment)
S1	1 (reference)
S2	0.29 (0.03–2.45)	0.254
S3	0.33 (0.04–2.50)	0.280
S4	0.36 (0.05–2.80)	0.331
S5	0.29 (0.04–2.17)	0.228
S6	0.24 (0.03–1.76)	0.158
S7	0.38 (0.05–2.85)	0.348
S8	0.29 (0.04–2.15)	0.228
Nodule size (>2 cm)	0.96 (0.63–1.47)	0.851
Treatment modality
LAT	1 (reference)			1 (reference)
TACE	2.56 (1.75–3.75)	<0.001		2.36 (1.61–3.47)	<0.001
RT	1.09 (0.50–2.37)	0.822		0.83 (0.38–1.83)	0.643

HR, hazards ratio; CI, confidence interval; HBV, hepatitis B virus; HCV, hepatitis C virus; PT, prothrombin time; INR, international normalized ratio; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AFP, alpha-fetoprotein; LAT, local ablation therapy; TACE, transarterial chemoembolization; RT, radiotherapy..

2) LTP-free survival and OP-free survival in the sub-cohort of LAT-unsuitable nodules

After PS-based matching and weighting in the subgroup of target lesions not suitable for LAT, covariates were balanced between the TACE-treated and RT-treated nodules (Supplementary Table 6). The RT-treated nodules had a longer LTP-free survival than the TACE-treated nodules in both PS-matched and IPTW pairs (HR, 2.91; 95% CI, 1.13 to 7.47; p=0.027 and HR, 3.07; 95% CI, 1.11 to 8.48; p=0.031) whereas OP-free survival did not differ for either the PS-matched or the IPTW pairs (HR, 1.24; 95% CI, 0.82 to 1.87; p=0.314 and HR, 1.46; 95% CI, 0.91 to 2.33; p=0.114) (Table 3, Fig. 3).

Figure 3. Propensity score-matched analyses of (A) LTP-free survival and (B) OP-free survival after TACE versus RT in local ablation therapy-unsuitable hepatocellular carcinomas. LTP, Local tumor progression; OP, overall progression; TACE, transarterial chemoembolization; RT, radiotherapy.

Table 3

Cox Regression Analysis of TACE versus RT for LAT-Unsuitable Hepatocellular Carcinomas in PS-Matched and IPTW-Weighted pairs.

Endpoint	PS-matched sample			IPTW-weighted sample
	HR (95% CI)*	p-value		HR (95% CI)†	p-value
LTP-free survival
RT	1 (reference)			1 (reference)
TACE	2.91 (1.13–7.47)	0.027		3.07 (1.11–8.48)	0.031
OP-free survival
RT	1 (reference)			1 (reference)
TACE	1.24 (0.82–1.87)	0.314		1.46 (0.91–2.33)	0.114

TACE, transarterial chemoembolization; RT, radiotherapy; LAT, local ablation therapy; PS, propensity score; IPTW, inverse probability treatment weighting; HR, hazards ratio; CI, confidence interval; LTP, local tumor progression; OP, overall progression..

*The Cox proportional hazards regression model was used with the robust variance estimator to account for the clustering within matched sets; ^†An inverse probability weighted Cox proportional hazards regression model was used..

3. Patient-based OS analysis

During the median follow-up of 26.9 months (range, 17.2 to 37.6 months) after the first course of treatment of the 651 patients, 17 died from various causes, while 37 underwent liver transplantation. Univariate and subsequent multivariate analyses did not identify any significant difference in OS across treatment modalities (HR, 0.66; 95% CI, 0.21 to 2.02; p=0.463 for TACE and HR, 3.04; 95% CI, 0.88 to 10.51; p=0.079 for RT compared with LAT) (Supplementary Table 7, Supplementary Fig. 2). The same was true for OS in the matched pairs (log-rank p=0.300) (Supplementary Fig. 3).

Current guidelines recommend LAT, especially RFA, as the first-line treatment if a single small HCC tumor is unresectable or if the patient is not eligible for liver transplantation. However, there is no consensus regarding second-line treatment for unablatable hepatic nodules.^1-5 In our matched evaluation of solitary HCC lesions that underwent initial assessment by sonography for pre-procedural planning, TACE and RT proved to have comparable OS, while RT gave better local tumor control. The relevant results of treatment by RT were not inferior to those of LAT restricted to ablatable tumors.

A previous Korean study assessing the suitability for ablation of 256 HCCs revealed that half were not suitable for ablation after planning sonographic examination due to invisibility of the nodules (55.8%), inadequate electrode path (23.9%), possible thermal injury to nearby organs (10.1%), and the heat sink effect (3.6%).⁹ Another Korean study showed that 45 out of 136 HCCs were not suitable for ablation on sonographic examination, with undetectability (71.1%) being the most common cause.⁸ These observations are consistent with the results of our study, where out of 801 nodules, 131 were unablatable due to ultrasonic invisibility. Inconspicuous HCCs are more difficult to locate when they are small or when no anatomical landmarks are found near the tumor.⁶ Otherwise, HCCs close to large vessels and those close to bile ducts or adjacent organs (e.g., gallbladder, colon, diaphragm, or heart) are not suitable for LAT, because of a potential heat-sink effect and thermal injury, respectively.^9,23 Moreover, subcapsular HCCs that were found in 19 unablatable nodules might not be good candidates for LAT due to technical difficulty and the greater risk of LTP.²⁴ Surprisingly, location of HCCs in the liver was not associated with LAT feasibility in the present study (data not shown).

To the best of our knowledge, this study is the first to report on alternative therapies for HCC with complete data on pre-procedural planning sonography carried out to evaluate the feasibility of LAT. A few Asian studies have compared the efficacy of LAT and TACE in early HCC lesions.^11,25 A Korean study compared the treatment efficacy of TACE and RFA in HCC lesions of ≤2 cm and showed that a group of 122 patients with TACE-treated tumors (18.0±2.9 months) had a shorter mean LTP-free survival than a group of 165 patients with RFA-treated tumors (27.0±3.8 months), although they had comparable OS results.¹¹ Additionally, a PS-matched analysis of cases from China with HCC tumors meeting the Milan criteria found similar trends in outcomes between the two treatments.²⁵ Although these studies did not clarify whether the included HCC lesions were evaluated for potential suitability for LAT, their findings are consistent with the data for our patients with unablatable nodules who were treated with TACE. As early-stage HCC lesions receive their blood supply from both hepatic artery and portal vein, this may be related to the reduction in the beneficial effect of TACE.²⁶ On the other hand, the effect of RT, particularly when using the stereotactic approach, was similar to or better than that of RFA in terms of local tumor control for small HCCs.^15,16 More recently, proton beam therapy, another type of RT, also yielded significantly longer 2-year LTP-free survival for recurrent small HCCs than RFA.²⁷ Explant liver evaluation of within-Milan tumors showed that pathologic tumor necrosis occurred in 56% and 68% of HCC lesions treated with stereotactic body RT and TACE, respectively.²⁸ These findings, together with ours, suggest that TACE and RT can both be considered rescue therapies in patients with small HCCs who are found to be not suitable for LAT after undergoing planning sonography.

Another crucial decision when caring for patients with unablatable HCCs is determining which type of procedure, TACE or RT, would be more effective for individual lesions. A Chinese study of 95 matched pairs of TACE-treated and stereotactic body RT-treated HCC lesions of Stage A according to the Barcelona Clinic Liver Cancer staging system reported that stereotactic body RT had a superior cumulative 5-year local control rate (56.9% vs 36.6%) and comparable 5-year OS results (62.8% vs 50.4%), in line with our results.²⁹ Another study found that TACE could be recommended as second-line treatment for tumors previously treated with RT and lesions adjacent to radiosensitive organs.³⁰ RT could be used instead in patients not suitable for TACE because of technical contraindications, allergies to contrast media or chemotherapeutic agents, or renal impairment.^31,32 Unfortunately, we could not identify specific indications for the use of TACE versus RT. However, we found that TACE-treated HCCs were frequently located nearer other organs than RT-treated HCCs. For the RT-treated lesions, we considered liver function and proximity to other organs. Hypo-fractionated RT was used to reduce the risk of hepatic dysfunction or other organ damage.²⁰ Hence, prospective studies are warranted to establish the indications for this choice.

This study has some limitations. First, although we attempted to reduce the effect of biases and confounders by means of various statistical adjustments such as multivariate analysis, PS matching, and IPTW, our study did not completely mimic a realistic randomized design. Second, due to the small size of the individual treatment samples, we could not compare outcomes across subtypes of LAT (RFA, microwave ablation, and cryoablation), TACE (conventional TACE and drug-eluting bead TACE), and RT (conventional RT, stereotactic body RT, and proton beam therapy). The specific method of treatment may have impacted the clinical outcomes. For HCCs ≤3 cm, conventional TACE exerted better local tumor control than drug-eluting bead TACE.³³ However, only 5.4% of the HCCs were treated with drug-eluting bead TACE so they will have had little impact on the results. The duration of follow-up in our study may not have been long enough to determine which treatment strategy was superior considering the expected survival of patients with single HCC ≤3 cm. Further work on the effects of these treatment strategies on OS is needed.

In conclusion, our comparative analysis suggests that TACE and RT have similar abilities to control single HCCs of ≤3 cm in diameter that are unsuitable for LAT, and to prolong survival. Moreover, RT may be as beneficial for ablatable tumors as LAT. Hence, the optimal indications for TACE and RT need to be established in order to clarify their application in clinical practice.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230080

This study was supported by grants from the Basic Science Research Program through the National Research Foundation of Korea (NRF-2022R1A2C3008956) and the Asan Cancer Institute of Asan Medical Center, Seoul, Korea (2022IP0046).

No potential conflict of interest relevant to this article was reported.

Study concept and design: J.L., E.K., J.H.S. Data acquisition: J.L., E.K. Data analysis and interpretation: J.L., E.K., S.K., J.H.K. Drafting of the manuscript: J.L., E.K., S.K., S.Y.K., J.H.K., S.M.Y., J.H.S. Critical revision of the manuscript for important intellectual content: S.Y.K., J.H.K., S.M.Y. Obtained funding: J.H.S. Study supervision: J.H.S. Approval of final manuscript: all authors.