Multiparametric MRI Radiomics for the Prediction of Histologic Grade in Bladder Cancer
Authors:
- Anup SelvarajuASAnup Selvaraju
- Durgesh Kumar DwivediDDDurgesh Kumar Dwivedi
- Naveen Kumar PatbamniyaNPNaveen Kumar Patbamniya
- Manoj KumarMKManoj Kumar
- Amlesh SethASAmlesh Seth
- Alok Kumar DwivediADAlok Kumar Dwivedi
- Rakesh Chandra JoshiRakesh Chandra Joshi
- Seema KaushalSKSeema Kaushal
- Chandan J. DasCDChandan J. Das
- 6 August 2025
- journal article
- research article
- Published by Springer Nature in Indian Journal of Surgical Oncology
- Vol. 17 (5) , 1057-1064
Abstract
Histological grade holds great clinical significance in the management and prognosis of bladder cancer; therefore, timely and accurate prediction through non-invasive techniques such as MRI may improve health outcomes. Our objective was to create an MRI-based radiomics model that can predict the histological grade of cancer pre-operatively. In a prospective study, we gathered data from 45 bladder cancer patients who had an mpMRI scan from December 2018 to December 2022 prior to their operative procedure. Radiomics features were extracted from T2-weighted (T2W), diffusion-weighted imaging (DWI), and dynamic DCE-MRI-enhanced (DCE) MR images obtained from a 1.5 T MRI scanner. A standard 5-point VI-RADS scoring system was also assessed for each scan. The variable clustering algorithm was applied to these features, and all cluster features were univariably assessed using receiver operating characteristic (ROC) curves. Multiple predictive models were created and cross-validated based on multivariable analysis to minimize overfitting and predict the grade of the tumor. Among 45 eligible patients, 28 (62.2%) patients had high-grade tumors and the rest 17 (37.8%) were low grade. In the adjusted analysis, only DCE-MRI based (Gray Level Co-occurrence Matrix (Gray Level Co-occurrence Matrix (GLCM))-inverse variance (OR = 1.42, p = 0.028), Major Axis Length (OR = 1.04, p = 0.03)), and age (OR = 1.08, p = 0.039) were associated with the high-grade bladder cancer. Our radiomics models comprising DCE-MRI-based parameters, a T2W parameter, and age yielded the highest performance for predicting grades of bladder cancer (AUC = 0.91; 95% CI 0.82–1.00). These models demonstrated reasonably high predictive performance in bootstrap validation analysis as well. An mpMRI radiomics approach based on MRI has the potential to serve as a non-invasive imaging tool for preoperative grading of bladder cancer.Keywords
This publication has 22 references indexed in Scilit:
- VI-RADS in bladder cancer: Overview, pearls and pitfallsEuropean Journal of Radiology, 2023
- Radiomics in Oncology: A Practical GuideRadioGraphics, 2021
- The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based PhenotypingRadiology, 2020
- European Association of Urology Guidelines on Non-muscle-invasive Bladder Cancer (TaT1 and Carcinoma In Situ) - 2019 UpdateEuropean Urology, 2019
- Multiparametric MRI for Bladder Cancer: Validation of VI-RADS for the Detection of Detrusor Muscle InvasionRadiology, 2019
- Multiparametric Magnetic Resonance Imaging for Bladder Cancer: Development of VI-RADS (Vesical Imaging-Reporting And Data System)European Urology, 2018
- Development and Validation of an MRI-Based Radiomics Signature for the Preoperative Prediction of Lymph Node Metastasis in Bladder CancerEBioMedicine, 2018
- Radiomics-guided therapy for bladder cancer: Using an optimal biomarker approach to determine extent of bladder cancer invasion from t2-weighted magnetic resonance imagesAdvances in Radiation Oncology, 2018
- Radiomics: Images Are More than Pictures, They Are DataRadiology, 2016
- 3D Slicer as an image computing platform for the Quantitative Imaging NetworkMagnetic Resonance Imaging, 2012