LOS ANGELESThe identification of biomarkers to aid in patient selection for treatment with agents targeting epidermal growth factor receptor (EGFR) has become a high-priority research goal. French investigators report that mutations in the KRAS oncogene may serve this purpose, as they were predictive of a lack of response and shortened progression-free survival (PFS) in colorectal cancer patients receiving cetuximab (Erbitux). [See also page 1.] The findings were presented at the 2007 American Association of Cancer Research annual meeting (abstract 5671).
Among the population of EGFR+ patients, approximately 40% display KRAS mutations, which affect the signaling pathway. Cetuximab is likely, therefore, to be ineffective in patients with the mutation, said senior author Professor Pierre Laurent-Puig, of the Hospital European Georges Pompidou, Paris, who discussed the findings at a press conference. The study was presented at the meeting by Dr. Astrid Lievre, of University Rene Descartes, Paris. "We found we can classify with two simple markers several groups of patients. Maybe for those with 'bad' markers, we need an agent other than cetuximab," Professor Laurent-Puig said.
The retrospective multicenter study evaluated 89 EGFR+ metastatic colorectal cancer patients treated with cetuximab alone or with irinotecan (Camptosar) as a single agent or as part of FOLFIRI. The overall response rate to cetuximab was 29%, including 1 complete response and 25 partial responses. The median duration of response was 35 weeks.
KRAS mutations, identified in 24 patients (27%), strongly correlated with lack of response to cetuximab. None of the 24 patients with mutations responded, and only 10 of these patients (41.7%) achieved stable disease. In contrast, among the 64 patients without KRAS mutations the response rate was 76.8%, and 25 patients (38.4%) had stable disease. The difference between mutated and nonmutated patients was highly significant (P = .001).
Progression-free and overall survival also correlated with mutation status. Median PFS was 10.1 weeks with mutations vs 31.4 weeks without mutations (P = .0001); median overall survival was 10.1 and 14.3 months, respectively (P = .026).
In a pooled analysis of 114 patients that included these 89 subjects and 25 from a previous study, the same highly significant differences were shown. No patients with the mutation had a response vs 41% without the mutation. Progression-free survival was 9 weeks with the mutation and 32 weeks without the mutation. Median overall survival was 10.1 vs 14.3 months (P = .0017).
Among the 89 patients in Professor Laurent-Puig's study, acneiform rash was seen in 85% of patients, including grade 1 toxicity in 43%, grade 2 in 49%, and grade 3 in 8%. Skin toxicity was highly correlated with tumor response and overall survival. Median survival was 8.2 months in patients with no skin toxicity or grade 1 rash vs 13.9 months in patients with grade 2-3 skin toxicity (P = .029).
The prognostic value of KRAS mutations remained significant in a multivariate Cox model that included age, sex, and skin toxicity. When both KRAS mutation status and skin toxicity were considered, a highly predictive scenario emerged (see Table). Patients with both good-prognosis markers had a survival time that was three times higher than patients with no good-prognosis markers (P = .0008).