A foundation model with weak experiential guidance in detecting muscle invasive bladder cancer on MRI

Yu Gong; Xiaodong Zhang; Yi-Fan Xia; Yi Cheng; Jie Bao; Nan Zhang; Rui Zhi; Xue-Ying Sun; Chen-Jiang Wu; Fei-Yun Wu; Yu-Dong Zhang

doi:10.1016/j.canlet.2025.217438

A foundation model with weak experiential guidance in detecting muscle invasive bladder cancer on MRI

Cancer Lett. 2025 Jan 2:217438. doi: 10.1016/j.canlet.2025.217438. Online ahead of print.

Authors

Yu Gong¹, Xiaodong Zhang², Yi-Fan Xia³, Yi Cheng⁴, Jie Bao⁵, Nan Zhang⁶, Rui Zhi⁷, Xue-Ying Sun⁸, Chen-Jiang Wu⁹, Fei-Yun Wu¹⁰, Yu-Dong Zhang¹¹

Affiliations

¹ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
² Department of Radiology, Peking University First Hospital, Beijing, P.R. China. Electronic address: [email protected].
³ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
⁴ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
⁵ Department of Radiology, The First Affiliated Hospital of Soochow University, Suzhou, Jiangsu Province, P.R. China, 215006. Electronic address: [email protected].
⁶ Department of Urology, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang Province, P.R. China, 310009. Electronic address: [email protected].
⁷ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
⁸ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
⁹ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
¹⁰ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029. Electronic address: [email protected].
¹¹ Department of Radiology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu Province, P.R. China, 210029; The Affiliated Suqian First People's Hospital of Nanjing Medical University, Suqian, Jiangsu Province, China. Electronic address: [email protected].

PMID: 39755362
DOI: 10.1016/j.canlet.2025.217438

Abstract

Preoperative detection of muscle-invasive bladder cancer (MIBC) remains a great challenge in practice. We aimed to develop and validate a deep Vesical Imaging Network (ViNet) model for the detection of MIBC using high-resolution T_2-weighted MR imaging (hrT₂WI) in a multicenter cohort. ViNet was designed using a modified 3D ResNet, in which, the encoder layers were pretrained using a self-supervised foundation model on over 40,000 cross-modal imaging datasets for transfer learning, and the classification modules were weakly supervised by an experiential knowledge-domain mask indicated by a nnUNet segmentation model. Optimal ViNet model was trained in derivation data (cohort 1, n = 312) and validated in multicenter data (cohort 2, n = 79; cohort 3, n = 44; cohort 4, n = 56) across a multi-ablation-test for model selection. In internal validation, ViNet using hrT₂WI outperformed all ablation-test models (odds ratio [OR], 7.41 versus 1.85 - 2.70; all P < 0.05). In external validation, the performance of ViNet using hrT₂WI versus ablation-test models was heterogeneous (OR, 1.31 - 3.89 versus 0.89 - 9.75; P = 0.03 - 0.15). In addition, clinical benefit of ViNet was evaluated between six readers using the Vesical Imaging-Reporting and Data System (VI-RADS) versus ViNet-adjusted VI-RADS. As a result, ViNet-adjusted VI-RADS upgraded 62.9% (17/27) of MIBC missed in VI-RADS score 2, while downgraded 84.1% (69/84), 62.5% (35/56) and 67.9% (19/28) of non-muscle-invasive bladder cancer (NMIBC) overestimated in VI-RADS score 3-5. We concluded that ViNet presents a promising alternative for diagnosing MIBC using hrT2WI instead of conventional multiparametric MRI.

Keywords: Deep learning; Multiparametric MRI (mpMRI); Muscle-invasive bladder cancer; Self-supervised learning; Vesical imaging-reporting and data system (VI-RADS).