Dual-channel deep learning captures intratumoural heterogeneity on CECT for preoperative risk stratification of thymic epithelial tumors.
Authors
Affiliations (5)
Affiliations (5)
- Department of Intensive Care Unit, Sixth Hospital of Shanxi Medical University, Taiyuan, Shanxi 030001, China.
- Department of Medical Imaging, First Hospital of Shanxi Medical University, Taiyuan, Shanxi 030001, China.
- Department of Medical Imaging, Shanxi Provincial People's Hospital Affiliated to Shanxi Medical University, Taiyuan, Shanxi 030001, China.
- Department of Medical Imaging, Shanxi Cancer Hospital, Shanxi Hospital Affiliated to Cancer Hospital, Chinese Academy of Medical Sciences, Cancer Hospital Affiliated to Shanxi Medical University, Taiyuan, Shanxi 030013, China.
- Department of Ultrasound Medicine, Inner Mongolia Autonomous Region People's Hospital, Hohhot, Inner Mongolia Autonomous 010000, China.
Abstract
Accurate preoperative risk stratification is critical for treating thymic epithelial tumors (TETs). This study developed a deep learning framework that combines a dual-channel convolutional neural network (CNN) with an adaptive dynamic clustering algorithm. The model was trained on contrast-enhanced CT (CECT) images from 336 multicenter TET patients. It first automates the segmentation of tumor subregions. Then, it constructs dual-channel input data containing the largest cross-sectional ROI and its corresponding habitat masks. Using transfer learning, we trained four CNN architectures for risk stratification. The DenseNet121-based dual-channel CNN achieved an AUC of 0.74-0.76 on an external test set. This performance surpassed conventional radiomics, single-channel CNN, and radiologists' visual assessment. Our framework effectively captures intratumoral heterogeneity, improves risk stratification accuracy, and aids in rapid identification of high-risk patients. This approach can support personalized treatment planning.