Abstract
BACKGROUND: Hepatectomy represents the cornerstone therapeutic approach for intrahepatic cholangiocarcinoma (ICCA); however, research pertaining to the prognosis of ICCA patients utilizing competing risk models remains scarce. This study aimed to construct a prognostic model utilizing competing risk analysis to predict cancer-specific survival (CSS) among ICCA patients posthepatectomy. METHODS: This study retrospectively analyzed ICCA patients from the Surveillance, Epidemiology, and End Results (SEER) database (2004-2015) who underwent hepatectomy. Patients were randomly allocated to the training (70%) and validation (30%) cohorts, with baselines balanced via propensity score matching (PSM). Prognostic factors were ascertained through both univariate and multivariate analyses of competing risks, facilitating the development of pertinent risk models and nomograms. The efficacy of the model was assessed via receiver operating characteristic (ROC) curves, area under the curve (AUC), and calibration plots, with clinical utility appraised through decision curve analysis (DCA). The X-tile program facilitated the categorization of participants into low-, intermediate-, and high-risk groups on the basis of their scores derived from the nomogram. RESULTS: Among the 1,131 participants included in the analysis after PSM, 65.34% (n=739) died from ICCA, and 13.97% (n=158) died from other causes. The 1-, 2-, and 3-year overall survival (OS) rates for ICCA patients after hepatectomy were 79.4%, 59.8% and 46.4%, respectively; the corresponding CSS rates were 82.5%, 64.0%, and 51.3%, respectively. Multivariate analysis revealed that hypodifferentiation, advanced T stage, lymph node invasion, and distant metastasis were significant risk factors. The AUCs for predicting CSS in the training cohort were 0.668, 0.711, and 0.710 for 1, 2, and 3 years, respectively. similarly, the AUCs for the test cohort were 0.709, 0.718, and 0.721 for 1, 2, and 3 years, respectively. The AUC demonstrated that the developed nomogram model exhibited moderate discriminatory power. The calibration curve demonstrated that the predicted values closely matched the actual data. DCA demonstrated greater clinical utility for the nomogram than the tumor node metastasis (TNM) classification system. Patients were divided into three risk groups according to the nomogram, which revealed substantial differences in survival rates between the groups (P<0.001). CONCLUSIONS: The prognostic nomogram developed based on the competitive risk model demonstrates moderate predictive accuracy for the specific survival rate of ICCA patients after hepatectomy, offering a practical tool for individualized prognostication and treatment planning.