Development of a MRI-Based Radiomics Nomogram for Prediction of Response of Patients With Muscle-Invasive Bladder Cancer to Neoadjuvant Chemotherapy

• DOI: 10.3389/fonc.2022.878499
• Publication Date: 2022-05-11T18:55:17Z
• Authors: Xinxin Zhang; Yichen Wang; Jin Zhang; Lianyu Zhang; Sicong Wang; Yan Chen
• Keywords: Nomogram

To develop and evaluate the performance of a magnetic resonance imaging (MRI)-based radiomics nomogram for prediction response patients with muscle-invasive bladder cancer (MIBC) to neoadjuvant chemotherapy (NAC).A total 70 clinical T2-4aN0M0 MIBC were enrolled in this retrospective study. For each patient, 1316 features extracted from T2-weighted images (T2WI), diffusion-weighted (DWI), apparent diffusion coefficient (ADC) maps. The variance threshold algorithm Student's t-test or Mann-Whitney U test applied select optimal features. Multivariate logistic regression analysis was used eliminate irrelevant features, retained incorporated into final single-modality model. Combined radiomic models generated by combining models. A nomogram, incorporating signatures independent risk factors, developed determine whether model predicting tumor NAC could be further improved.Based on pathological T stage post-surgery, 36 (51%) classified as good responders (GR) 34 (49%) non-good (non-GR). In addition, 3 4 combined established. Among all models, based T2WI_Score, DWI_Score, ADC_Score yielded highest area under receiver operating characteristics curve (AUC) (0.967, 95% confidence interval (CI): 0.930-0.995). integrating higher AUC (0.973, 95%CI: 0.934-0.998) than other models.The proposed MRI-based has potential non-invasive tool quantitatively MIBC.