Locoregional Therapies for Hepatocellular Carcinoma: A Systematic Review and Meta-Analysis

Abstract

Importance: Several locoregional therapies (LRTs) for nonmetastatic hepatocellular carcinoma (HCC) are available; however, a global comparison of the relative efficacy of each is needed.

Objective: To conduct a systematic review and direct, pairwise meta-analytic comparison of all identified randomized clinical trials evaluating the treatment of nonmetastatic HCC.

Data sources: A comprehensive search of PubMed and the proceedings of the American Society of Clinical Oncology and American Society for Radiation Oncology annual meetings from January 1, 2010, to November 1, 2023, was performed.

Study selection: Randomized clinical trials using a form of LRT (surgery with or without adjuvant therapy, radiofrequency ablation [RFA], microwave ablation [MWA], radiotherapy [RT], hepatic arterial infusion chemotherapy [HAIC], transarterial bland embolization [TAE], transarterial chemoembolization [TACE], or transarterial radioembolization [TARE]).

Data extraction and synthesis: Study eligibility and data extraction were each reviewed by 2 authors independently. Random-effects meta-analyses were used to compare treatment categories.

Main outcomes and measures: Progression-free survival (PFS) was the primary outcome; overall survival (OS) was the secondary outcome.

Results: Forty randomized clinical trials reporting on comparative outcomes of 11 576 total patients with localized HCC treated with LRT were included. The median follow-up was 30.0 (IQR, 18.5-40.8) months. Direct pooled comparisons between treatment classes suggested improved outcomes for surgery combined with adjuvant therapy over surgery alone (PFS: hazard ratio [HR], 0.62 [95% CI, 0.51-0.75]; P < .001; OS: HR, 0.61 [95% CI, 0.48-0.78]; P < .001), surgery over RFA (PFS: HR, 0.74 [95% CI, 0.63-0.87]; P < .001; OS: HR, 0.71 [95% CI, 0.54-0.95]; P = .02), RT over TACE (PFS: HR, 0.35 [95% CI, 0.21-0.60]; P < .001; OS: HR, 0.35 [95% CI, 0.13-0.97]; P = .04), and HAIC over TACE (PFS: HR, 0.57 [95% CI, 0.45-0.72]; P < .001; OS: HR, 0.58 [95% CI, 0.45-0.75]; P < .001). No substantial heterogeneity was noted for any pairwise comparison with the exception of RT-based regimens compared with tyrosine kinase inhibitor therapy.

Conclusions and relevance: The findings of this systematic review and direct, pairwise meta-analysis suggest that all LRTs are not equivalent for the treatment of localized HCC. The efficacy of LRTs appears hierarchical, with surgery-based management outcomes associated with the best treatment outcomes and embolization-based treatment options associated with the worst treatment outcomes.

Figure 1.. Size of Each Included Study

This diagram represents each study included in the progression-free survival (30 studies, 8537 patients) (A) and overall survival (37 studies, 10 301 patients) (B) meta-analyses. Each chord represents a single study, with arrow direction and intensity of shading indicating the favored treatment and relative magnitude of treatment effect observed in that study. Arc lengths represent the number of patients treated with a specific therapy included in this analysis. CT indicates systemic therapy without specification of drug; HAI, hepatic arterial infusion; ICI, immune checkpoint inhibitor; MWA, microwave ablation; RFA, radiofrequency ablation; RT, radiotherapy-based treatment; TACE, transarterial chemoembolization; TAE, transarterial bland embolization; TARE, transarterial radioembolization; and TKI, tyrosine kinase inhibitor.

Figure 2.. Trial-Level Relative Treatment Effects From Each Profiled Randomized Clinical Trial

Box areas are proportional to weights in each meta-analysis, and diamond widths are proportional to the precision of the pooled estimate from a given meta-analysis. HAIC indicates hepatic arterial infusion chemotherapy; HR, hazard ratio; ICI, immune checkpoint inhibitor; MWA, microwave ablation; RFA, radiofrequency ablation; RT, radiotherapy; TACE, transarterial chemoembolization; TAE, transarterial bland embolization; TARE, transarterial radioembolization; and TKI, tyrosine kinase inhibitor. ^aPooled outcome with or without the addition of adjunctive therapy.

Figure 3.. Treatment Effect Estimates of Each Profiled Locoregional Therapy

The relative treatment effect (ie, hazard ratio) from each trial is included in this analysis and shows the emergence of a 4-tier ordinal structure of locoregional therapies (LRTs) at the trial level, where treatment effects favor LRTs either within the same tier or in a lower-numbered tier (ie, tier 1 > 2 > 3 > 4), and this is congruent with the findings from the pairwise meta-analysis. Arrows point to the LRT with a more favorable outcome in a randomized comparison, and shading intensity is proportional to the magnitude of the treatment effect. Arc lengths are proportional to the amount of information informing a specific trial comparison. CT indicates chemotherapy; HAI, hepatic arterial infusion; ICI, immune checkpoint inhibitor; MWA, microwave ablation; RFA, radiofrequency ablation; RT, radiotherapy-based treatment; TACE, transarterial chemoembolization; TAE, transarterial bland embolization; TARE, transarterial radioembolization; and TKI, tyrosine kinase inhibitor.