Relationship between recurrence-free survival after curative resection for hepatocellular carcinoma and prophylactic tyrosine kinase inhibitor use

BACKGROUNDPostoperative recurrence remains common after curative resection for hepatocellular carcinoma (HCC), whereas the effectiveness of prophylactic tyrosine kinase inhibitor (TKI) therapy in routine clinical practice remains uncertain. AIMTo investigate the association between postoperative prophylactic oral TKI therapy and recurrence-free survival (RFS) after curative resection for HCC. METHODSWe retrospectively included 143 patients who underwent their first R0 resection for HCC at a single center.TKI therapy was modeled as a time-dependent exposure from the adjudicated initiation date.The primary model incorporated verified demographic, virological, liver-function, surgical, pathological, tumor-stage, and postoperative prophylactic transarterial chemoembolization (TACE) variables.We additionally performed 31-, 60-, and 90-day landmark analyses, propensity-weighted analyses in the 90-day cohort, and prespecified sensitivity analyses.RESULTS Among 143 patients, 76 received prophylactic TKI therapy and 67 did not.During a median follow-up of 53.7 months, 91 RFS events occurred.In the primary time-dependent Cox model, TKI initiation was not significantly associated with RFS [adjusted hazard ratio (HR) = 1.29; 95% confidence interval (CI): 0.76-2.20;P = 0.338].The direction was consistent across landmark, weighting, doubly robust, time-dependent TACE, competing-event, and alternative-model analyses.Cirrhosis (adjusted HR = 1.62; 95%CI: 4 / 35 1.05-2.50;P = 0.030) and American Joint Committee on Cancer stage IIIA vs IA (adjusted HR = 3.26; 95%CI: 1.33-8.02;P = 0.010) were associated with worse RFS.CONCLUSION In this single-center retrospective cohort, postoperative prophylactic TKI initiation was not significantly associated with improved RFS after accounting for treatment timing and measured confounders.The wide confidence interval and residual confounding preclude firm conclusions regarding benefit, harm, equivalence among agents, or responsive subgroups.