Ultrasound-guided deep learning framework with whole slide images for molecular subtyping of breast cancer: a multicentre, retrospective study.
Authors
Affiliations (17)
Affiliations (17)
- Department of Ultrasound, Zhejiang Cancer Hospital, Hangzhou, China.
- Interventional Medicine And Engineering Research Center, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Hangzhou, China.
- Zhejiang Provincial Research Center for Cancer Intelligent Diagnosis and Molecular Technology, Hangzhou, China.
- Wenling Big Data and Artificial Intelligence Institute in Medicine, Taizhou, Zhejiang, China.
- Department of Pathology, Zhejiang Cancer Hospital, Hangzhou, Zhejiang, China.
- Department of Ultrasound of the Affiliated Dongyang Hospital of Wenzhou Medical University, Dongyang, China.
- Department of Ultrasound Imaging, The First Affiliated Hospital of Wenzhou Medical University, New District of the First Affiliated Hospital of Wenzhou Medical University, Wenzhou City, China.
- Department of Medical Ultrasound, Yunnan Cancer Hospital & The Third Affiliated Hospital of Kunming Medical University, Kunming, China.
- Center of Intelligent Diagnosis and Therapy (Taizhou), Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Taizhou, Zhejiang, China.
- Department of Pathology, Zhejiang Cancer Hospital, Hangzhou, Zhejiang, China. [email protected].
- Department of Ultrasound, Zhejiang Cancer Hospital, Hangzhou, China. [email protected].
- Interventional Medicine And Engineering Research Center, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Hangzhou, China. [email protected].
- Center of Interventional Radiology and Vascular Surgery, Nurturing Center of Jiangsu Province for State Laboratory of AI Imaging & Interventional Radiology (Southeast University), Department of Radiology, Zhongda Hospital, Medical School, Southeast University, Nanjing, China. [email protected].
- Department of Ultrasound, Zhejiang Cancer Hospital, Hangzhou, China. [email protected].
- Interventional Medicine And Engineering Research Center, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Hangzhou, China. [email protected].
- Zhejiang Provincial Research Center for Cancer Intelligent Diagnosis and Molecular Technology, Hangzhou, China. [email protected].
- Wenling Big Data and Artificial Intelligence Institute in Medicine, Taizhou, Zhejiang, China. [email protected].
Abstract
Accurate molecular subtyping of breast cancer is essential for personalized therapy, yet inconsistencies between biopsy and postoperative immunohistochemistry (IHC) due to tumor heterogeneity and sampling limitations pose clinical challenges. We developed Ultrasound-Guided Pathomic Subtyping (US-GPS), a cross-modal framework enabling subtype prediction using ultrasound (US) alone, guided by IHC-stained whole-slide images (WSIs) during training. A total of 1283 patients from four centers were retrospectively included, with 454 cases (paired US and WSIs) for model training/validation and 829 US-only cases for external testing. US-GPS aligned WSI-derived patch- and region-level features (from HER2, ER, PR, Ki-67 IHC stains) with handcrafted and deep US features via co-attention and contrastive learning. Using US alone, the model achieved an AUROC of 0. 966 (95% CI 0.960-0.991) in internal validation, comparable to pathology-only (0. 971) and joint models (0.980), and superior to US-only (0.861). Performance remained robust across both internal (AUROC range: 0.930-0.942) and all external cohorts (AUROC range: 0.915-0.963). US-GPS enables scalable, non-invasive, and interpretable subtype classification, with potential for both pre-treatment decision-making and longitudinal monitoring of molecular phenotypic shifts.