Phase 2 study of the efficacy and safety of erdafitinib in patients (pts) with bacillus Calmette-Guérin (BCG)-unresponsive, high-risk non–muscle-invasive bladder cancer (HR-NMIBC) with FGFR3/2 alterations (alt) in THOR-2: Cohort 2 interim analysis results.

503 Background: Pts presenting with NMIBC carcinoma in situ (CIS) have a high risk of progression. FGFR inhibition may benefit CIS pts with FGFRalt who are unresponsive to first-line BCG, for whom treatment (tx) options, other than radical cystectomy, are limited. Erdafitinib (erda), an oral selective pan-FGFR tyrosine kinase inhibitor, is approved for locally advanced or metastatic urothelial cancer in adults with FGFR3/2alt who have progressed during or after ≥1 line of platinum-containing chemotherapy. THOR-2 (NCT04172675) is a multicohort phase 2 study of erda in pts with HR-NMIBC. Here we report results from an exploratory cohort of pts with BCG-unresponsive CIS with FGFRalt with or without papillary disease (Cohort 2). Methods: Inclusion criteria: age ≥18 y, with histologically confirmed, BCG-unresponsive HR-NMIBC with FGFR3/2alt (by local/central testing) presenting as CIS, with or without a papillary tumor and who refused or were not eligible for cystectomy. In this cohort, pts received continuous oral erda 6 mg once daily without uptitration in 28-d cycles (dose selected to improve tolerability while maintaining activity to prevent disease recurrence, for this population). Erda was discontinued if no complete response (CR) was observed within 3 mos. Exploratory efficacy end points are CR rates at the Cycle 3 Day 1 (C3D1) disease evaluation and the Cycle 6 Day 1 (C6D1) disease evaluation; safety was a key secondary end point. Results: As of the data cutoff (Sep 2022) (median follow-up of 10 mos), 10 pts have received erda (enrolled population; median age 72 y [range 52-83]; 90% CIS [1 pt with Ta was mis-enrolled]). Pts received erda for a median duration 5.9 mos (range 1.1-17.0). Of 10 enrolled pts, the CR rates at first evaluation (C3D1) and second evaluation (C6D1) were 100% (9/9 evaluable pts) and 75% (6/8 evaluable pts), respectively. The median duration of response was 3.0 mos. The most common tx-emergent adverse events (TEAEs) were dry mouth (60%; n=6), hyperphosphatemia (50%; n=5), dysgeusia (50%; n=5), and diarrhea (50.0%; n=5). 1 pt (10%) had Gr 2 retinal detachment which led to tx discontinuation and 1 pt (10%) had Gr 1 subretinal fluid; both reported as resolved. Gr ≥3 tx-related TEAEs occurred in 3 pts (30%), and included dry mouth, stomatitis, nail disorder, onychomadesis, acute kidney injury, chronic kidney disease, sepsis, and hypotension in 1 pt each. 1 pt (10%) had serious tx-related TEAEs (dry mouth, hypotension, pneumonitis, acute kidney injury, and sepsis), and 1 pt (10%) discontinued tx due to a tx-related TEAE. No tx-related deaths were observed. Conclusions: Data from Cohort 2 of THOR-2 demonstrate efficacy at C3D1 and C6D1 evaluations in pts with HR-NMIBC with FGFRalt. Safety data were consistent with the known safety profile of erda. Clinical trial information: NCT04172675 .