Response-durability phenotypes and clinical benefit with erdafitinib: a THOR trial post hoc analysis

To determine whether response durability refines the clinical interpretation of objective response (OR) to erdafitinib in patients with fibroblast growth factor receptor (FGFR)-altered metastatic urothelial carcinoma.

This post hoc exploratory analysis used de-identified individual participant-level data from THOR trial Cohort 1 (ClinicalTrials.gov identifier: NCT03390504) The safety population included 247 patients treated with erdafitinib (n = 135) or chemotherapy (n = 112). Response-durability phenotypes were defined by integrating best overall response with a 6-month progression-free survival threshold: durable OR, durable disease control, non-durable OR, and limited benefit. Phenotype distributions were compared between treatment arms. Overall survival (OS) was evaluated using Kaplan-Meier and Cox analyses, including 6-month landmark analyses to reduce guarantee-time bias.

Erdafitinib shifted patients toward more favourable response-durability phenotypes compared with chemotherapy. Limited benefit was less frequent with erdafitinib than with chemotherapy (45.9% vs 75.9%; P < 0.001), whereas OR phenotypes were more common. In erdafitinib-treated patients, 6-month landmark OS differed significantly across phenotypes (P < 0.001). Durable benefit was associated with longer landmark OS than non-durable benefit (non-durable vs durable benefit: hazard ratio [HR] 4.09, 95% confidence interval [CI] 2.24-7.49). Among objective responders, non-durable OR was associated with substantially worse landmark OS than durable OR (HR 8.71, 95% CI 3.30-22.96).

In patients with FGFR-altered metastatic urothelial carcinoma treated with erdafitinib, Response Evaluation Criteria in Solid Tumours (RECIST) OR alone did not fully capture clinically meaningful benefit. Incorporating response durability provided stronger survival stratification and may improve the interpretation of treatment benefit after FGFR-targeted therapy.