Multidimensional deep learning for grading and prognostic assessment of intrahepatic mass-forming cholangiocarcinoma.
Authors
Affiliations (7)
Affiliations (7)
- Department of Radiology, Affiliated Hospital of Hebei University, Baoding, People's Republic of China.
- Department of Research and Development, United Imaging Intelligence (Beijing) Co., Ltd., Beijing, People's Republic of China.
- Department of Critical Care Medicine, Baoding First Central Hospital, Baoding, People's Republic of China.
- Department of Research and Development, United Imaging Intelligence, Shanghai, People's Republic of China.
- Department of Medical Imaging, The Second Hospital of Hebei Medical University, Shijiazhuang, People's Republic of China.
- Department of Radiology, Affiliated Hospital of Hebei University, Baoding, People's Republic of China. [email protected].
- Department of Radiology, Affiliated Hospital of Hebei University, Baoding, People's Republic of China. [email protected].
Abstract
To develop an interpretable magnetic resonance imaging (MRI)-based framework for preoperative histologic grading of intrahepatic mass-forming cholangiocarcinoma (IMCC) and exploratory prognostic stratification. A retrospective analysis was conducted on preoperative MRI from 333 IMCC patients across three centers (training cohort, n = 240; external validation cohort, n = 93). An ensemble deep learning (DL) framework synergizing 2.5D and 3D ResNet-50 architectures was constructed. Significant variables from clinical-laboratory-imaging (ClinLabImag) features, radiomics, and DL outputs were integrated into a combined model. Discrimination was assessed using the area under the receiver operating characteristic curve (AUC). Model interpretability was evaluated with Gradient-weighted Class Activation Mapping (Grad-CAM) and SHapley Additive exPlanations (SHAP), and the Kaplan-Meier method was used to compare overall survival (OS) between risk groups. The DL model achieved an external validation AUC of 0.804 (95% confidence interval (CI): 0.712-0.896), significantly outperforming standalone 2.5D (p = 0.025) and 3D architectures (p = 0.030). The Combined model (AUC: 0.843 [95% CI, 0.749-0.938]) showed better external validation performance than the Radiomics (p = 0.019) and ClinLabImag models (p = 0.017), with only modest, non-significant improvement over the DL model (p = 0.355). SHAP analysis showed that DL features contributed most to model predictions. The Combined model showed exploratory OS differences between risk groups. An interpretable multidimensional MRI-based DL framework supports noninvasive preoperative grading in IMCC and provides exploratory prognostic information. This study critically evaluates an interpretable multidimensional MRI-based deep learning framework for preoperative grading and exploratory prognostic stratification of intrahepatic mass-forming cholangiocarcinoma, supporting individualized radiologic risk assessment before treatment. Reliable noninvasive MRI biomarkers are needed for preoperative grading and prognostic assessment of intrahepatic mass-forming cholangiocarcinoma to guide individualized treatment planning. The combined multidimensional MRI model achieved favorable external validation performance and showed exploratory overall survival differences between IMCC risk groups.