Randomized phase II trial of capecitabine (X) versus X plus erlotinib (E) in patients (pts) with metastatic colorectal cancer (mCRC): Differential impact of KRAS.

Abstract

e14031^ Background: Palliative X monotherapy of mCRC is generally tolerated by elderly, and/or medically unfit pts. Epidermal growth factor receptor (EGFR) inhibitors improve efficacy when added to combination chemotherapy for mCRC. It is unknown whether biologic agents add benefit to alone in pts who either opt against, or are not candidate for, combination chemotherapy. We conducted a randomized phase II trial to investigate the novel “all-oral” combination of low-dose X and E in this pt population. Methods: Eligible pts with mCRC were either deemed unfit for, or chose against, 1st-line combination chemotherapy. Pts were randomized to X (1000 mg/m2 PO BID x 14 days) alone or in combination with E (150 mg PO OD) on a 3-week schedule. A 3rd arm (E alone) was ineffective, and discontinued after a preplanned interim analysis. Imaging was performed every 9 weeks until progression. Primary endpoint was time to disease progression (TTP); secondary endpoints included: objective response rate (ORR), overall survival (OS), safety and quality of life. Results: From June 2004 to June 2008, 95 pts [median age: 67; range 40-84; ECOG PS: 0 (45pts), 1 (38 pts), or 2 (22 pts)] were randomized to X alone (40 pts), XE (42 pts), or E alone (13 pts). KRAS status, where possible, was analyzed for pts in the first 2 arms only. As compared to X alone, XE was associated with significantly more Gr 3+ toxicity: (diarrhea: 29% vs 5%; fatigue: 12% vs 0%; anorexia 7% vs 0%; stomatitis 7% vs 0%). One pt discontinued treatment in the X arm vs 2 in XE. There was 1 toxic death in XE. TTP was not different for the whole trial (X=7.9 m, XE=9.2 m; p=0.8897), but exploratory subgroup differences emerged (below). Conclusions: The addition of E to X is associated with additional toxicity. E may benefit pts with KRAS-wild-type CRC and may harm those with KRAS-mutated CRC. Further study of XE in elderly/unfit pts with KRAS-wild-type CRC is warranted. Supported by grant from Roche. Median TTP (m) X XE p All pts (n = 40 / 42) 7.9 9.2 0.8897 KRAS wt (n =18 / 24) 8.4 11.7 * KRAS mut (n = 18 / 12) 7.4 1.9 * *Cox regression p=0.0236 for interaction of study arm by Kras status.