842 Background: Luminal and basal bladder cancers exhibit differential chemosensitivity, yet intratumoral heterogeneity limits treatment prediction. We hypothesized that tumors arising through papillary versus carcinoma in situ pathways show differential neoadjuvant chemotherapy (NAC) response. We developed a dual CK20/GATA3 scoring system to test this hypothesis in a pilot study. Methods: Pre-treatment transurethral resection specimens from 92 consecutive radical cystectomy patients receiving cisplatin-based NAC underwent dual CK20/GATA3 staining. Three patterns were identified: Luminal-Papillary (CK20-/GATA3+, grade 3), Luminal-CIS (CK20+/GATA3±, grade 4), Basal-CIS (CK20-/GATA3-, grade 5). Using Gleason-like methodology, a genitourinary pathologist assigned primary grade (most prevalent pattern) plus secondary grade (highest additional pattern present in any amount); pure single-pattern tumors received doubled grade, yielding scores 6-10. Pre-specified binary classifier: Low (6-7, papillary-enriched) versus High (8-10, CIS-predominant). Primary endpoint was pathologic downstaging (ypT1 or less). Results: Low (n = 48) and High (n = 44) groups balanced for age, sex, clinical T stage (cT2: 38% versus 33%). Downstaging occurred in 26 of 48 (53.1%) Low versus 15 of 44 (33.3%, absolute difference ≈20 percentage points; Fisher p = 0.066) and the direction of effect persisted after adjustment for clinical T stage (logistic OR 0.52 for High vs Low; 95% CI 0.21–1.25; p = 0.142). For DFS, the cT-adjusted Cox model showed HR 1.22 for High vs Low (95% CI 0.50–3.02; p = 0.662), OS was underpowered. Conclusions: This pilot study identifies a clinically meaningful downstaging signal (20 percentage-point difference) using dual-stain immunohistochemistry without sequencing and may enable preoperative enrichment for response. Prospective multi-institutional validation is warranted.
Jo et al. (Sun,) studied this question.