[MRI-based deep learning model for preoperative prediction of urothelial carcinoma with variant histology of bladder: a retrospective, multicenter study].

Authors

Abstract

<b>Objectives:</b> To develop a deep learning model based on magnetic resonance imaging (MRI) for the preoperative prediction of urothelial carcinoma with variant histology (VUC), and to evaluate its predictive performance and prognostic stratification value. <b>Methods:</b> This is a retrospective cohort study. Data from 1 221 patients with bladder cancer across seven centers of China between June 2013 and May 2025 were retrospectively analyzed. There were 1 032 male cases and 189 female cases, with an age (<i>M</i>(IQR)) of 68.0 (15.0) years (range: 19 to 96 years). Data from Jiangsu Provincial People's Hospital (<i>n=</i>959) were randomly assigned to a training set (<i>n=</i>767) and a validation set (<i>n=</i>192) in an 8∶2 ratio, while data from the other six centers (<i>n=</i>262) served as an external test set. Patients were categorized into pure urothelial carcinoma (PUC) group and VUC group. A deep learning prediction model for VUC (DeepVUC) was developed using a 3D ResNet50-based framework, integrating MRI T2WI and diffusion-weighted imaging (DWI) through a self-attention mechanism. The comparison of data between groups was conducted using Mann-Whitney <i>U, χ</i>² test or Fisher's exact probability method. Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC), and the prognostic value was assessed through Kaplan-Meier analysis. <b>Results:</b> Among the 1 221 patients with bladder cancer, 1 078 cases were PUC and 143 cases were VUC, with a proportion of VUC at 11.7%. Comparative analysis revealed that relative to the PUC group, the proportions of lymph node metastasis, high-grade urothelial carcinoma, carcinoma in situ, and lymphovascular invasion were significantly higher in the VUC group (all <i>P<</i>0.01). Bonferroni pairwise comparison showed that the proportion of VUC in pT3 stage patients was significantly higher than that in pT2 stage (<i>χ</i>²=14.30, <i>P<</i>0.01), pT1 stage (<i>χ</i>²=56.76, <i>P<</i>0.01) and pTa stage (<i>χ</i>²=174.02, <i>P<</i>0.01). The VUC group also exhibited significantly higher vesical imaging-reporting and data system(VI-RADS) scores (<i>χ</i>²=141.48, <i>P<</i>0.01). Specifically, the percentage of VUC in patients with VI-RADS scores of 1 to 5 was 3.5% (4/113),4.2% (22/524), 9.1% (25/275), 25.0% (37/148) and 34.2% (55/161), respectively. Squamous differentiation was the most common variant, accounting for 51.7% (74/143) of all VUC patients, with the highest frequency observed in patients with a VI-RADS score of 5 (69.1%, 38/55). The DeepVUC model achieved AUCs of 0.972 (95<i>%CI</i>: 0.955 to 0.989), 0.940 (95<i>%CI</i>: 0.902 to 0.977), and 0.925 (95<i>%CI</i>: 0.889 to 0.961) in the training, validation, and external test sets, respectively. Accuracies were 95.2%, 90.1%, and 84.4%; sensitivities were 92.1%, 90.9%, and 84.4%; and specificities were 95.6%, 90.0%, and 84.3%. Prognostic analysis showed that the DeepVUC-defined high-risk group had significantly worse overall survival compared to the low-risk group in the total cohort (<i>HR</i>=4.51, 95<i>%CI</i>: 3.24 to 6.26, <i>P<</i>0.01), the PUC group (<i>HR</i>=2.13, 95<i>%CI</i>: 1.42 to 3.18, <i>P<</i>0.01), and the VUC group (<i>HR</i>=3.17, 95<i>%CI</i>: 1.75 to 5.76, <i>P<</i>0.01). <b>Conclusions:</b> The DeepVUC model enables non-invasive preoperative prediction of VUC in bladder cancer and demonstrates significant value in prognostic stratification. As a non-invasive imaging biomarker, this model provides a valuable reference for the development of individualized treatment strategies for bladder cancer patients.

Tags

Topics