Predicting recurrence and recurrence-free survival in initially unresectable hepatocellular carcinoma: a novel nomogram for patients undergoing conversion hepatectomy with lenvatinib, PD-1 inhibitor, and interventional therapy.

PURPOSE

This research aims to develop prognostic nomograms to predict tumor recurrence and recurrence-free survival (RFS) in individuals with initially unresectable hepatocellular carcinoma (uHCC) who were later subjected to conversion hepatectomy following lenvatinib, PD-1 inhibitors, and interventional (LPI) therapy.

METHODS

We performed a retrospective review of clinical information from 150 individuals diagnosed with HCC who underwent conversion hepatectomy following LPI therapy between November 2019 and December 2024. Independent predictors linked to recurrence and RFS were identified through comprehensive univariate and multivariate analyses, and the identified factors were subsequently integrated into nomogram models. Receiver operating characteristic (ROC) curves, calibration plots, and the concordance index (C-index) were employed to evaluate the predictive performance of the nomograms.

RESULTS

Our investigation identified several key risk factors for recurrence, including age, tumor number, tumor differentiation, preoperative prognostic nutritional index (PNI), preoperative systemic immune-inflammation index (SII), and postoperative protein induced by vitamin K absence or antagonist-II (PIVKA-II) level. For RFS, significant predictors included tumor number, tumor differentiation, preoperative SII, postoperative PIVKA-II, and postoperative alpha-fetoprotein (AFP) levels. The nomograms exhibited strong predictive performance, achieving a C-index of 0.837 (95% CI: 0.775-0.896) for recurrence prediction and 0.837 (95% CI: 0.788-0.886) for RFS. Our nomogram for recurrence prediction outperformed traditional staging systems like China Liver Cancer (CNLC) staging and Barcelona Clinic Liver Cancer (BCLC). Calibration curves and discriminative ability assessments confirmed the nomograms' reliability in predicting actual outcomes and stratifying patients into distinct prognostic subgroups with significant RFS differences across risk categories.

CONCLUSIONS

The nomogram models established in this research provide an exceptionally accurate and individualized method for predicting recurrence and RFS in initially uHCC patients undergoing LPI-based conversion hepatectomy, potentially aiding clinicians in devising tailored treatment plans and enhancing patient outcomes.