Clinically actionable TURBT-derived bladder cancer organoid pharmacotyping as a predictor of postoperative outcomes in a prospective multicenter study.

Abstract

e16613 Background: Post-TURBT chemotherapy is largely empiric despite marked inter-patient heterogeneity in bladder cancer drug sensitivity. We evaluated a multicenter TURBT-derived PDO pharmacotyping workflow and its concordance with postoperative outcomes. Methods: In a prospective multicenter study (NCT06662071), TURBT bladder cancer specimens were used to generate patient-derived organoids (PDOs), with histologic fidelity assessed by H&E and immunohistochemistry (CK5/6, CK7, CK20, Ki-67, p53, p63). PDOs underwent 3D drug testing (serial dilutions; 72-hour exposure) across eight clinically used agents/regimens; viability (CellTiter-Glo 3D) was used to derive IC50 and nAUC (plus viability at Cmax). Six response-classification strategies (Jenks vs tertiles × IC50/nAUC/Cmax viability) were compared, and concordance with postoperative chemotherapy outcomes (recurrence/progression on surveillance follow-up) was summarized by specificity and accuracy. Results: From 4 centers, 63 TURBT specimens yielded 41 PDO lines from 37 patients (65.1% success). Successful establishment correlated with higher post-digestion viability (p < 0.01) and cell yield (p < 0.05). Drug profiling was completed in 35/41 lines (85.4%). IC50 values spanned submicromolar medians for anthracyclines/pirarubicin (pirarubicin 0.080 µM; doxorubicin 0.540 µM; epirubicin 0.508 µM) to higher micromolar medians for platinum/antimetabolites (cisplatin 20.245 µM; gemcitabine 16.913 µM; methotrexate 61.205 µM), with gemcitabine showing the widest range (0.004–369.500 µM). Of six strategies, IC50-based Jenks natural breaks (Sensitive/Intermediate/Resistant) best matched clinical outcomes (specificity 0.933; accuracy 0.882). Clinical concordance was assessed in 15 treated patients (17 PDO–patient pairs); 2/15 (13.3%) recurred/progressed. Illustrative concordant cases included pirarubicin sensitivity (IC50 0.12 µM) with recurrence-free follow-up and gemcitabine resistance (IC50 227.7 µM) with early recurrence/progression after GC-based management. Conclusions: This multicenter TURBT-derived bladder cancer PDO study establishes a clinically actionable pharmacotyping workflow. PDOs can be generated from routine specimens, capture marked inter-patient drug-response heterogeneity, and provide drug-specific sensitivity tiers that align with postoperative outcomes with high specificity for identifying potentially ineffective therapy. These data support prospective validation of PDO-guided regimen selection to reduce futile treatment and individualize postoperative bladder cancer care.