ObjectiveTo compare the efficacy of liver resection (LR), radiofrequency ablation (RFA), and radiofrequency ablation combined with transcatheter arterial chemoembolization (RFA+TACE) in the treatment of very-early- and early-stage hepatocellular carcinoma (HCC).MethodsWe systemically searched the PubMed, Embase, and Cochrane Library databases. Randomized controlled trials (RCTs) and observational analyses with propensity score-matched cohort analyses (PSMs) comparing any two of the three treatments were included in this study. The primary result was overall survival (OS) and the secondary result was recurrence-free survival (RFS), which were analyzed by calculating the hazard ratio (HR) and 95% confidence intervals (CI).ResultsA total of 25 studies (4249 patients), including 10 RCTs and 15 PSM observational studies, met the inclusion criteria. Although there was no significant difference between LR and RFA in terms of one-year OS, though LR showed superior performance for three- and five-year OS (at three years, HR: 0.74, 95% CI: 0.56-0.96; at five years, HR: 0.73, 95% CI: 0.55-0.94). In addition, significantly higher rates of RFS at one-, three- and five-year follow-up were found for LR than for RFA alone (at one year, HR: 0.68, 95% CI: 0.51-0.92; at three years, HR: 0.67, 95% CI: 0.55-0.81; at five years, HR: 0.61, 95% CI: 0.48-0.78). The combination of RFA+TACE was superior to RFA alone based on one-year RFS (HR: 0.57, 95% CI: 0.34-0.96), while there were no significant differences in OS at one, three, and five years, and in RFS at three and five years.ConclusionsFor very-early- and early-stage HCC, this systematic review and network meta-analysis showed that the efficacy of LR is superior to that of RFA alone, regardless of whether the evaluation is based on either OS or RFS. The advantages of RFA+TACE compared to RFA alone are limited, and further studies are needed to determine whether combination therapy is necessary, i.e., results in significantly improved outcomes. Systematic Review RegistrationThe study was registered with http://www.crd.york.ac.uk/PROSPERO, identifier: CRD42022299269