Multimodal MR Radiomics for Preoperative Prediction of Perineural Invasion and Prognostic Stratification in Intrahepatic Cholangiocarcinoma

Background To develop and validate a multimodal model integrating multiphasic contrast-enhanced MRI radiomics with clinical-imaging information for the preoperative prediction of perineural invasion (PNI) in intrahepatic cholangiocarcinoma (ICC) and for postoperative risk stratification. Methods Patients with ICC treated at the Eastern Hepatobiliary Surgery Hospital between September 2014 and February 2025 were retrospectively enrolled ( n = 455) and randomly assigned to training ( n = 232), testing ( n = 99) and validation ( n = 124) cohorts. Radiomic features were extracted with PyRadiomics from tumor regions on all four phases and selected by LASSO regression; the clinical-imaging, radiomics and multimodal models were built with logistic regression. Discrimination, calibration and clinical net benefit were assessed by the area under the receiver operating characteristic curve (AUC), calibration curves and decision curve analysis (DCA). Overall survival (OS) and recurrence-free survival (RFS) were compared by the Kaplan-Meier