Creative Commons source reading · Prediction and forecasting
Biparametric MRI with deep learning for the preoperative prediction of muscle invasion in patients with bladder cancer: a multicenter study
Ziyi Liu; Yilan Hu; Jiegeng Lyu; Lei Deng; Shuiqing Yang; Xiang Huang; Xiaohui Duan; Weijing Zhang; Bingsheng Huang; Jianpeng Li; Yujian Zou
Originally published in Insights Imaging · 2026-11-09
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
ObjectivesThis study aimed to develop and validate a deep learning model that integrates T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) for the noninvasive prediction of muscle invasion in bladder cancer.Materials and methodsA total of 308 bladder tumors from 254 patients were retrospectively collected from three institutions. Of these, 177 tumors from Center 1 were partitioned via stratified sampling into a training set (n = 128) and an internal validation set (n = 49). The remaining two centers contributed 59 and 72 tumors, respectively, which served as external validation cohorts. A dual-branch 3D deep convolutional neural network (Joint model) based on biparametric MRI (bp-MRI) was developed using the training data and compared with a deep learning model based on T2WI (T2WI model) and VI-RADS assessments by…
Discussion
In this study, we developed a deep learning model using bp-MRI inputs (T2WI and DWI), referred to as the Joint model, for diagnosing muscle-invasive bladder cancer (MIBC). The Joint model integrates T2WI and DWI through a low-rank Hadamard bilinear fusion module with deep supervision, and was compared with radiologists’ VI-RADS assessments and three single-sequence deep learning models trained solely on T2WI. Our findings demonstrate that the Joint model outperformed both the VI-RADS scoring and the T2WI-based model in predicting muscle invasion in patients with bladder cancer. Notably, for lesions scored as VI-RADS 2 or 3, the Joint model significantly outperformed the T2WI model in diagnostic accuracy and showed strong generalizability across external validation cohorts. Our results demonstrated strong interreader agreement of…
Conclusion
In conclusion, we propose the Joint model that integrates T2WI and DWI sequences for the diagnosis of MIBC. This model outperformed both VI-RADS and the T2WI model in multicenter validation, and showed improved performance for lesions scored VI-RADS 2 and 3 compared with the T2WI model and VI-RADS assessment. These strengths highlight the clinical value of the Joint model for the reliable and noninvasive evaluation of muscle invasion in patients with bladder cancer.
Editorial reading guide
- What data created and tested the model, and was evaluation independent?
- What is the cost of a false positive, false negative or out-of-distribution decision?
