Phase II study of cisplatin (C), gemcitabine (G) and gefitinib for advanced urothelial carcinoma (UC): Analysis of the second cohort of CALGB 90102.

4578 Background: The epidermal growth factor receptor (EGFR) is frequently expressed in UC and carries a poor prognosis. C has preclinical synergy with gefitinib and GC plus gefitinib was safe in lung cancer patients (pts). In a UC trial of GC using fixed dose rate infusion of G plus gefitinib, Gr 4 and 5 toxicities led to premature closure of the study after 27 patients. Because toxicity was felt to be related to the fixed dose rate infusion of G, accrual was resumed with a standard G dosing schedule in a second cohort. METHODS: Eligible pts had measurable N2, N3 or M1 disease; PS of 0-2; CrCl of > 50 ml/min; adequate organ function; no prior systemic combination chemotherapy. Treatment consisted of C 70 mg/m(2) D1, G 1000 mg/m(2) D1+8 given over 30 min every 3 weeks concurrent with gefitinib 500 mg/day PO for a maximum of 6 cycles. Responders received maintenance gefitinib 500 mg/day until progression. RESULTS: In 55 eligible pts with a median age of 64 years, 67% (of 49) had visceral metastases and 91% had PS of 0-1. Objective response (CR+PR) was observed in 51% (95% CI = 37-65). With a median follow up of 13.2 mo, the median time to progression was 8 mo (95% CI = 6.8-9.2) on the basis of 45 events, and the median overall survival was 14.4 mo (95% CI = 10.7-20.9) on the basis of 26 deaths. No lethal toxicity was seen. Gr 4 toxicities included Gr 4 neutropenia (20%) and Gr 4 metabolic/electrolyte disorders in 13%. Gr 3 and 4 diarrhea was observed in 25% and 2 % of pts respectively and Gr 3 skin rash in 16% of pts. CONCLUSIONS: The combination of GC and gefitinib has acceptable toxicity in advanced UC. However, the relative contribution of gefitinib to the efficacy of this regimen remains uncertain. These preliminary results do not suggest a substantial improvement upon historical results with GC alone in advanced UC. No significant financial relationships to disclose.

Duke Scholars

Author Susan Halabi Biostatistics & Bioinformatics, Division of Biostatistics