Artificial intelligence in the diagnosis and pathological assessment of bladder cancer: current evidence and clinical applications

Abstract Bladder cancer represents the ninth most commonly diagnosed malignancy worldwide, with substantial morbidity and mortality. Conventional diagnostic modalities—including cystoscopy, urine cytology, cross-sectional imaging, and histopathological examination—are limited by operator-dependent variability, modest sensitivity for early-stage disease, and significant interobserver discordance. Artificial intelligence (AI), and particularly deep learning (DL)-based approaches, has emerged as a transformative paradigm to enhance diagnostic accuracy and standardize assessment across the complete diagnostic continuum. This narrative review critically synthesizes contemporary evidence on AI applications in bladder cancer diagnosis, encompassing cystoscopic tumor detection, urine cytology analysis, radiomics-based staging, computational pathology, multimodal fusion architectures, and intraoperative guidance. Across validation cohorts, AI-enhanced cystoscopy achieves sensitivity of 91–99% and specificity of 87–99%. AI-augmented urine cytology demonstrates substantial sensitivity improvements, with the VisioCyt system achieving 84.9% overall sensitivity compared with 43% for conventional cytology. Radiomics and deep learning approaches for imaging analysis achieve area under the curve values ranging from 0.834 to 0.997 for staging and muscle-invasion prediction. Computational pathology systems achieve diagnostic accuracy meeting or exceeding that of experienced pathologists while providing standardized, reproducible assessments. Notwithstanding these advances, challenges including data standardization, model interpretability, prospective clinical validation, regulatory harmonization, and health-economic evaluation must be addressed to enable widespread clinical implementation. This review identifies critical research priorities and discusses pathways for responsible translation of AI innovations into routine urologic practice.