Retrospective Histopathological Aspects in the Recurrence of Bladder Tumours Following TUR-B: Insights into Progression and Regression Patterns, Computational Aspects and the Relevance of AI

Abstract

Background: Bladder cancer is one of the most common urological malignancies worldwide, often requiring multiple transurethral resections of the bladder (TUR-B) due to high recurrence rates. This retrospective study analyses 389 histopathological (HP) reports from 117 patients to evaluate the evolution of tumour grade (G1, G2, G3) and invasion level (non-invasive vs. invasive) in recurrent bladder tumors. Methods: We included patients who underwent two or more  TUR-B interventions between 2009 and 2024, had complete HP data for each resection, and were followed to assess recurrence, progression, and regression. Descriptive statistics and chi-square tests were used to examine differences in tumour behaviour among subgroups. Endpoints included recurrence (any new tumour), progression (increase in grade or stage to ≥T2), stagnation (no change in grade or stage), and regression (downgrade or downstage from invasive to non-invasive). Results: Nearly 15% of non-invasive tumours progressed to invasive disease, whereas over 30% of tumours initially classified as invasive showed partial or complete regression during subsequent resections. G2 tumours were most prevalent (around 55%), with more than half exhibiting stagnation. Some high-grade (G3) lesions demonstrated notable regression rates, highlighting possible responsiveness to intravesical therapy. Conclusions: These findings underscore the heterogeneity of bladder tumour evolution following TUR-B and the importance of vigilant surveillance and adjuvant therapies. While a subset of invasive tumours may regress, others progress despite initial low-grade presentation. Future prospective or multivariate analyses are needed to identify precise predictors of progression and regression in non-muscle-invasive bladder cancer (NMIBC). Also considering our previous expertise in the connections with AI and both urology and cancer, this could open further discussions and relevance in this growing area of research.