Patient-Derived Organoids for Predicting Chemotherapy Response in Bladder Cancer: A Multicenter, Real-world...

Responses to chemotherapeutic agents vary substantially among patients with bladder cancer. The lack of preclinical models that accurately predict the clinical efficacy of intravesical chemotherapy remains a major barrier to precision oncology.

In this prospective multicenter study (NCT06662071), 63 patients with bladder cancer were enrolled at four medical centers. Patient-derived organoids (PDOs) were established using specimens obtained by transurethral resection of bladder tumor. A total of 35 PDOs underwent high-throughput screening with a panel of eight clinically relevant chemotherapeutic agents. To identify the optimal classification strategy, we evaluated six distinct approaches by combining two statistical methods (the Jenks natural breaks method and the tertile method) with three sensitivity indicators (viability, nAUC, and IC₅₀) and assessed the concordance between drug sensitivity profiles and real-world clinical follow-up data.

We successfully established 41 PDOs. Comparative analysis indicated that post-digestion cell viability and total cell count may influence the success of PDO establishment. The PDOs showed high morphological and molecular concordance with the primary tumors. Among the six candidate strategies, IC₅₀-Jenks showed the highest numerical concordance with clinical outcomes, with accuracy of 88.2% (95% CI, 63.6-98.5%) and specificity of 93.3% (95% CI, 68.1-99.8%). Representative case studies confirmed the utility of PDOs in identifying therapeutic futility and supporting the feasibility of bladder-sparing strategies.

This multicenter feasibility study suggests that bladder cancer PDOs may support exploratory assessment of chemotherapy sensitivity. The proposed classification approach requires validation in larger, prospectively designed, clinically homogeneous cohorts before it can inform routine treatment decisions.