New AI model forecasts bladder cancer survival outcomes
AI model predicts survival outcomes for bladder cancer patients with newfound accuracy
image: AI Model Predicts Survival Outcomes for Bladder Cancer Patients with Newfound Accuracy
Credit: Freepik Licensed by: JMIR Publications
(Toronto, March 11, 2026) Researchers at Fondazione Policlinico Universitario Agostino Gemelli IRCCS have developed a promising machine learning algorithm capable of predicting survival and cause of death for patients with bladder cancer undergoing radical cystectomy. The study, published in JMIR Perioperative Medicine, highlights a significant shift toward using artificial intelligence (AI) to personalize post-surgical care and improve prognostic accuracy.
Radical cystectomy is a critical intervention for bladder cancer, yet roughly 50% of patients develop metastases within two years of the procedure. Traditional statistical models often struggle to capture the complex, non-linear variables that dictate whether a patient will remain disease-free. This new AI-driven approach provides clinicians with a more nuanced tool to identify high-risk individuals.
Key Findings from the AI Analysis
The research team utilized the CatBoost algorithm to analyze a comprehensive dataset of clinical, pathological, and inflammatory markers. Key insights include:
-
Predictive Power: The model successfully predicted Disease-Free Survival (DFS) and Overall Survival (OS), identifying Clinical Tumor Stage and Pathological Tumor Classification as the most dominant predictors of patient outcomes.
-
The "SII" Threshold: The study highlighted the Systemic Immune-Inflammation Index (SII) as a vital biomarker. The AI revealed a "threshold effect," where SII values above 1,000 correlated with a sharp decline in survival, suggesting that systemic inflammation plays a larger role in cancer progression than previously quantified.
-
Anatomical Influence: For the first time, bladder tumor position was identified as a top-tier predictor for determining the specific cause of death, with tumors in the posterior and right lateral walls showing higher correlations with tumor-related mortality.
-
The BMI Paradox: The model observed a U-shaped relationship with Body Mass Index, indicating that both underweight and obese patients faced higher risks, while moderate BMI ranges were associated with more favorable outcomes.
Bridging the Gap to Personalized Medicine
"The integration of AI offers a promising avenue for enhancing prognostic accuracy and personalizing treatment strategies," the study notes. While the researchers emphasize that the model is currently best suited for clinical trial stratification rather than immediate emergency bedside use, the ability to identify 11 out of 14 tumor-related deaths in the test group marks a significant step forward in AI-assisted urology.
By making the code publicly available, the team aims to encourage further validation across larger, multi-center datasets to refine the algorithm’s precision.
About the Research Team
This study was made possible through the collaborative efforts of a distinguished multidisciplinary team: Vittorio De Vita, Andrea Nappi, Melissa Sawaya, Bernardo Rocco, Nazario Foschi, Giuseppe Maioriello, and Pierluigi Russo. The team is particularly proud to highlight the contributions of Francesco Andrea Causio from Università Cattolica del Sacro Cuore, Italy. Causio was recently recognized as one of the five winners of the 2025 JMIR Publications Early Career Researcher Award, an honor that underscores the caliber and impact of the research presented in this study.
Original article: Causio F, De Vita V, Nappi A, Sawaya M, Rocco B, Foschi N, Maioriello G, Russo P. Survival Prediction in Patients With Bladder Cancer Undergoing Radical Cystectomy Using a Machine Learning Algorithm: Retrospective Single-Center Study. JMIR Perioper Med 2026;9:e86666
URL: https://periop.jmir.org/2026/1/e86666
DOI: 10.2196/86666
###
About JMIR Publications
JMIR Publications is a leading open access publisher of digital health research and a champion of open science. With a focus on author advocacy and research amplification, JMIR Publications apartners with researchers to advance their careers and maximize the impact of their work. As a technology organization with publishing at its core, we provide innovative tools and resources that go beyond traditional publishing, supporting researchers at every step of the dissemination process. Our portfolio features a range of peer-reviewed journals, including the renowned Journal of Medical Internet Research.
To learn more about JMIR Publications, please visit jmirpublications.com or connect with us via X, LinkedIn, YouTube, Facebook, and Instagram.
Head office: 130 Queens Quay East, Unit 1100, Toronto, ON, M5A 0P6 Canada
Media contact: communications@jmir.org
The content of this communication is licensed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, published by JMIR Publications, is properly cited.