Predicting Recurrence and Progression in Contemporary Patients with NMIBC Undergoing Blue Light Cystoscopy-Aided Transurethral Resection of Bladder Tumor

BACKGROUND: Although blue light cystoscopy (BLC) is recommended by clinical practice guidelines to improve cancer detection, predictive models for patients undergoing BLC are lacking. We therefore evaluated the associations of clinicopathologic features with recurrence and progression in a contemporary cohort of patients treated with BLC to develop a risk prediction model. METHODS: We identified patients aged 18-89 years in the multi-institutional Blue Light Cystoscopy with Cysview Registry who underwent BLC-aided TURBT between 2014-2023. Associations of baseline characteristics with recurrence and progression were evaluated using Cox regression, and predictive models were developed using lasso regression. RESULTS: 1040 patients underwent BLC-TURBT. Tumor stage was Ta in 578 (56%), T1 in 166 (16%), and pure CIS in 190 (18%) patients. On multivariable analysis, HGTa (HR 1.60, 95%CI 1.10-2.35) and HGT1 (HR 1.94, 95%CI 1.25-3.02) tumors were associated with increased risk of recurrence. HGT1 (HR 12.4, 95%CI 4.52-34.0), HGTa (HR 3.08, 95% CI 1.03-9.18), pure CIS (HR 4.51, 95% CI 1.44-14.1), Ta+CIS (HR 4.98, 95%CI 1.17-21.1), and T1+CIS (HR 7.65, 95%CI 1.97-29.7) tumors, as well as lymphovascular invasion (HR 3.27, 95%CI 1.22-8.83) were associated with increased risk of progression. In contrast, adjuvant BCG was associated with decreased risk of progression (HR 0.40, 95%CI 0.22-0.72). Model discrimination was 0.64 for RFS and 0.73 for PFS. CONCLUSIONS: Using a multi-institutional cohort, we developed predictive models for recurrence and progression in patients undergoing BLC-TURBT. Although they require external validation, these models reflect to contemporary treatment paradigms and can inform personalized, risk-adapted management of NMIBC.