Conversion hepatectomy and survival in initially unresectable hepatocellular carcinoma: a systematic review and meta-analysis with quantitative bias analysis

BACKGROUND: Conversion hepatectomy has become feasible for selected patients with initially unresectable hepatocellular carcinoma (HCC), with comparative studies reporting longer survival after surgery. We performed a systematic review, meta-analysis, and quantitative bias analysis (QBA) to assess whether residual selection and time-dependent treatment bias could account for the reported benefit. METHODS: We systematically reviewed comparative studies of conversion or salvage hepatectomy versus non-surgical management in initially unresectable HCC and pooled endpoint-specific HRs for OS and PFS/EFS using random-effects models. For each estimate, we extracted the time origin, exposure definition, and adjustment model. Risk of bias and certainty of evidence were assessed using ROBINS-I and GRADE. Robustness to unmeasured confounding, immortal time bias, and small-study effects was evaluated using E-values, tipping-point analysis, probabilistic QBA, timing-bias scenario analyses, and funnel-plot diagnostics. RESULTS: Fourteen studies contributed 24 h: 14 for OS and 10 for PFS/EFS. Hepatectomy was associated with lower hazards of death (HR 0.33, 95% CI 0.22-0.48) and progression or event (HR 0.41, 95% CI 0.28-0.59), although heterogeneity was substantial and both prediction intervals crossed 1.0. E-values suggested robustness to moderate unmeasured confounding, whereas extreme surgical-candidacy confounding attenuated the median bias-adjusted HR to 0.82 for OS and 1.02 for PFS/EFS. Immortal-time bias scenarios did not fully account for the observed associations. OS small-study-effect diagnostics suggested funnel-plot asymmetry, although sample-size-based Peters-type regression was not significant. Certainty of evidence was very low for both endpoints. CONCLUSIONS: Conversion hepatectomy was associated with better observed survival outcomes in selected patients with initially unresectable HCC after conversion-intent therapy. The association was robust to moderate, but not extreme, assumptions about unmeasured surgical-candidacy confounding. These findings should be interpreted as a composite post-conversion association reflecting surgical clearance, treatment response, favorable tumor biology, and patient selection rather than as the isolated causal effect of hepatectomy. Future studies should report surgical eligibility landmarks, time zero, time to hepatectomy, and reasons for non-resection to clarify the independent contribution of surgery within conversion pathways. TRIAL REGISTRATION: PROSPERO CRD420261382843 (https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=1382843).