Clin Radiol. 2025 Jun 11. pii: S0009-9260(25)00190-4. [Epub ahead of print]88 106985
AIM: To evaluate the value of preoperative magnetic resonance imaging (MRI) in predicting perineural invasion (PNI) of mass intrahepatic cholangiocarcinoma (MICC) and construct a nomogram.
MATERIALS AND METHODS: This retrospective study included 228 patients with pathologically confirmed MICC who underwent preoperative MRI between January 2015 and November 2022 at two institutions. Patients were randomly divided into a training cohort (n = 160) and validation cohort (n = 68) in a 7:3 ratio based on PNI presence. Two radiologists independently analysed imaging features. Significant predictors were identified using univariate and multivariate logistic regression analyses. Predictive performance was validated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). A nomogram was constructed based on the optimal model.
RESULTS: Age, presence of bile duct stones, T2-weighted imaging (T2WI) signal intensity, diffusion-weighted imaging (DWI) signal intensity, intratumoural exponential apparent diffusion coefficient (eADC) values, and the tumour-to-spleen eADC ratio were significant independent predictors of PNI (all P < 0.05). The nomogram performed well in the training cohort (AUC: 0.839; 95% CI: 0.776-0.901) and validation cohort (AUC: 0.771; 95% CI: 0.658-0.883).
CONCLUSION: The T2WI signal intensity, DWI signal intensity, intratumoural eADC values, and the tumour-to-spleen eADC ratio may serve as novel noninvasive biomarkers for predicting PNI in MICC patients. The proposed nomograms can be selectively applied to enhance the accuracy of preoperative PNI predictions in patient with ICC, thereby aiding in the development of surgical strategies.