Can machine learning help predict risks before surgery for bladder cancer?

Machine learning (ML) is a type of artificial intelligence that can analyze complex patterns in medical data. For bladder cancer, researchers are testing whether ML can help predict risks before surgery, such as how long a patient might survive or whether the cancer will spread. Current evidence shows that ML models, especially those using CT scans (radiomics) and clinical information, can predict outcomes like overall survival and distant metastasis with good accuracy. These tools could help doctors and patients make more informed treatment decisions.

What the research says

Several studies have developed and tested ML models for preoperative risk prediction in bladder cancer. A 2024 multicenter study used a deep learning model based on preoperative CT scans to predict overall survival after radical cystectomy (bladder removal surgery) in patients with muscle-invasive bladder cancer. The model's deep learning score outperformed traditional radiomics and clinical models, with a C-index of 0.690 in internal validation and 0.658 in external validation ⁸. Another 2025 study used data from over 43,000 patients to build ML models that predict distant metastasis in muscle-invasive bladder cancer. The best model (CatBoost) achieved an AUC of 0.956 in the training set and 0.839 in an external test set, meaning it was highly accurate at distinguishing patients who would develop metastasis from those who would not ⁹. A meta-analysis of radiomics-based ML studies found that these models can effectively stratify risk before surgery, though the authors noted that systematic evidence is still limited ¹. Additionally, a 2025 study integrating multi-omics data and ML identified an eight-gene signature that predicted overall survival and immune profiles in bladder cancer, suggesting ML can also help personalize treatment ⁷. Overall, the research indicates that ML models can provide useful risk predictions, but they are not yet standard clinical tools and require further validation.

What to ask your doctor

• Are there any machine learning tools available at your center to help predict my surgical outcomes?
• How accurate are current risk prediction models for bladder cancer surgery, and what factors do they consider?
• Could a CT-based deep learning model help estimate my survival or risk of cancer spread after surgery?
• What are the limitations of using machine learning predictions for my personal treatment plan?
• Should I consider any additional tests or imaging to improve risk assessment before surgery?