Ultrasound-based breast nodule malignancy diagnosis using CNN enhanced by knowledge distillation.
Authors
Affiliations (7)
Affiliations (7)
- Taizhou Key Laboratory of Minimally Invasive Interventional Therapy & Artificial Intelligence, Taizhou Campus of Zhejiang Cancer Hospital (Taizhou Cancer Hospital), Taizhou, Zhejiang 317502, China.
- School of Life Science and Technology, Xi'an Jiaotong University, Xi'an, Shaanxi 710049, China.
- Affiliated Xiaoshan Hospital, Hangzhou Normal University, No.728 North Yucai Road, Hangzhou, Zhejiang 311202, China.
- Center of Intelligent Diagnosis and Therapy (Taizhou), Zhejiang Cancer Hospital, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Taizhou, Zhejiang 317502, China.
- Wenling Institute of Big Data and Artificial Intelligence in Medicine, Taizhou, Zhejiang 317502, China.
- Department of Diagnostic Ultrasound Imaging & Interventional Therapy, Zhejiang Cancer Hospital, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Hangzhou, Zhejiang 310022, China.
- Wenzhou Medical University, Wenzhou, Zhejiang 325035, China.
Abstract
Accurate discrimination of benign and malignant breast lesions is essential for early diagnosis, yet remains challenging due to lesion heterogeneity and interobserver variability in ultrasound assessment. We developed a deep learning (DL) framework integrating dual inverted bottleneck blocks (Dual-IB), adaptive hierarchical feature fusion, and knowledge distillation (KD) for breast ultrasound lesion classification. The model was trained and evaluated on public datasets and validated on an independent external cohort to assess its generalizability. Results demonstrated superior performance compared with baseline methods and resident radiologists, while maintaining robust performance across diverse datasets and imaging conditions. On the external validation dataset, the model achieved the best performance, with an accuracy (Acc) of 92.64% and an F1-score of 92.37%. Furthermore, class activation mapping (CAM) visualizations revealed focused and anatomically plausible attention patterns within lesion regions, aiding interpretation and suggesting its potential to assist in breast ultrasound lesion classification.