Integrating artificial intelligence across the bladder cancer continuum: progress, promise, and pitfalls

Comprehensive Summary

This study conducted by Manivasagam et al. examines the role of artificial intelligence (AI) in improving the diagnosis, risk stratification, and management of bladder cancer. The authors performed a literature search across major databases from 2005 to 2025 and included 49 peer-reviewed studies. The review found that AI models, particularly deep learning systems, improve tumor detection during cystoscopy, enhance CT/MRI staging accuracy, automate histopathologic grading, and predict recurrence, treatment response, and survival. Several systems showed high sensitivity, sometimes outperforming clinicians in controlled settings. AI also showed promise in integrating imaging, genomic, and clinical data for personalized treatment decisions. However, most studies were retrospective and lacked prospective multi-center validation or proven survival benefit. The authors concluded that although AI shows strong potential to standardize and personalize bladder cancer care, broader adoption requires external validation, workflow integration, and evidence of cost-effectiveness and clinical outcome improvement.

Outcomes and Implications

This research is important because bladder cancer has high recurrence rates and costly long-term surveillance, creating a strong need for better risk prediction and treatment selection tools. AI could support real-time diagnosis, automate grading, and guide therapy decisions, but it remains investigational. Clinical implementation will likely require several more years, pending prospective validation and guideline endorsement