Enhancing recurrence risk prediction for bladder cancer using multi-sequence MRI radiomics.

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Bladder cancer (BCa) recurrence poses a significant challenge in patient management.
• Accurate prediction of recurrence-free survival (RFS) is crucial for tailoring treatment strategies.
• Existing clinical models have limitations in precisely predicting BCa recurrence.

Purpose of the Study:

• To develop and validate a radiomics-clinical nomogram for predicting RFS in BCa patients.
• To assess the superiority of the developed nomogram compared to traditional clinical models.
• To leverage multi-sequence MRI for enhanced prognostic accuracy in BCa.

Main Methods:

• Retrospective analysis of 229 BCa patients with preoperative multi-sequence MRI.
• Extraction and selection of radiomics features using LASSO regression.
• Development of a nomogram integrating radiomics and clinical factors.
• Validation using Kaplan-Meier analysis, NRI, and decision curve analysis.

Main Results:

• Radiomics features significantly correlated with RFS and were independent of clinical factors (p < 0.001).
• The combined radiomics-clinical model achieved high prognostic value (C-index: 0.853 training, 0.832 validation).
• The radiomics-clinical nomogram demonstrated superior calibration and classification over the clinical model (NRI: 0.6768, p < 0.001).

Conclusions:

• A multi-sequence MRI-based radiomics-clinical nomogram effectively predicts BCa recurrence risk.
• This novel nomogram outperforms both radiomics-only and clinical-only models.
• The nomogram offers a promising tool for personalized treatment and surveillance in bladder cancer.