Abstract
Background/Objectives: Non-muscle-invasive bladder cancer (NMIBC) recurs frequently, and clinicopathological tools predict recurrence imperfectly. We assessed whether six preoperative blood-count-derived inflammatory indices (SII, SIRI, PIV, NLR, PLR, and LMR) are associated with adverse histology and 12-month recurrence and whether SIRI adds prognostic information to established risk factors. Methods: We retrospectively analysed 180 consecutive NMIBC patients treated by transurethral resection of bladder tumour (TURBT) and immediate instillation. Indices were modelled per 1-SD in logistic (adverse histology: high grade, T1 or carcinoma in situ) and Cox (recurrence) models adjusted for established risk factors. Results: Adverse histology occurred in 121 patients (67.2%); 88 recurred (12-month Kaplan–Meier estimate: 49.2%). All six indices were univariably associated with both endpoints, with similar effects (SIRI: adverse histology OR: 2.35, 95% CI: 1.41–3.90; recurrence HR: 1.41, 1.19–1.67). After adjustment, SIRI was not independently associated with adverse histology (OR: 1.69, 0.92–3.11) or recurrence (HR: 1.10, 0.86–1.40) and did not improve discrimination (ΔC: 0.001, −0.006 to 0.016). Results were robust across sensitivity analyses, including propensity weighting for BCG (limited by positivity violations). Conclusions: SIRI added no demonstrable independent prognostic value. Given modest overfitting (optimism-corrected calibration slope: 0.79), neither the indices nor the model should be used clinically without external validation and recalibration.