Development and multi-observer validation of a histopathological scoring system for interstitial cystitis/bladder pain syndrome diagnosis

PURPOSE: To develop and validate a pragmatic histopathological scoring system for interstitial cystitis/bladder pain syndrome (IC/BPS) that distinguishes Hunner-type interstitial cystitis (HIC) from bladder pain syndrome (BPS) and inflammatory non-HIC chronic cystitis mimicking Hunner lesions ("other cystitis"). METHODS: Bladder biopsies from 96 patients were analyzed: HIC (n = 32), BPS (n = 42), and other cystitis (n = 22; bacillus Calmette-Guérin-related, follicular, and chronic bacterial cystitis). Four pathologists (A-D; A as index/model observer) independently scored hematoxylin-eosin sections using five histopathological features (inflammation, plasma cell infiltration, eosinophil infiltration, tertiary lymphoid structures, urothelial denudation; total 0-8). ROC curves were generated for HIC vs. BPS and HIC vs. other cystitis. Youden index-derived thresholds were compared across observers to select simple integer cutoffs and define an operational three-tier classification (0-2: BPS, 3-4: indeterminate, 5-8: HIC). Diagnostic performance was assessed using AUC and contingency-table metrics. Inter-observer agreement for the three-tier classification was evaluated using Fleiss' kappa. RESULTS: HIC vs. BPS discrimination was excellent across pathologists (AUC 0.966-0.991; pooled AUC 0.983). HIC vs. other cystitis discrimination was moderate-to-good (AUC 0.723-0.808; pooled AUC 0.738). Overall agreement for the three-tier classification was 60.4% with Fleiss' κ = 0.657; agreement was highest for BPS and modest for other cystitis. CONCLUSIONS: This hematoxylin-eosin-based scoring system enables standardized IC/BPS phenotyping, robustly separating HIC from BPS and providing a structured framework for inflammatory overlap cases, with reproducible performance across international pathologists.