Development and external validation of a prognosis model to predict outcomes after curative resection of early-stage intrahepatic cholangiocarcinoma

BACKGROUND: Early-stage intrahepatic cholangiocarcinoma (ESICC) with curative resection and lymph node-negative still has the risk of poor prognosis, and there lacks prognosis-assessing tools for these patients. The objective of this study was to develop a prognosis model to predict outcomes and identify risk stratification for ESICC after resection. METHODS: Totally 263 patients with ESICC after hepatectomy from January 2012 to January 2022 were analyzed. Clinicopathological factors were selected using multivariable Cox regression analysis and a prognosis model was developed. The performance of the model was evaluated by concordance index (C-index), calibration plots, decision curves analysis (DCA), and net reclassification index (NRI). Kaplan-Meier curves were analyzed for risk stratification of overall survival (OS) and recurrence-free survival (RFS) based on the prognosis model. RESULTS: The clinicopathological features that were independently associated with OS of ESICC included carbohydrate antigen19-9, carcinoembryonic antigen, tumor size, tumor differentiation, and T stage. The prognosis model based on these prognostic factors demonstrated excellent discriminatory performance in both derivation cohort (C-index, 0.71) and external validation cohort (C-index, 0.78), which outperformed the TNM staging system (C-index, 0.59) and individual prognostic factors (all C-index < 0.7). Calibration plots, DCA and NRI also showed superior predictive performance. According to the risk for survival, the model stratified patients into low risk (median OS, 66.6 months; median RFS, 24.3 months) and high risk (median OS, 24.0 months; median RFS, 6.4 months) (P < 0.001). CONCLUSIONS: Our prognosis model can robustly predict the outcomes of ESICC after curative resection and provide precise evaluation on prognosis risk, facilitating clinicians to develop individualized postoperative treatment options.