ARTICLE ABSTRACTPurpose: Intermediate/high-risk operated patients with head and neck cancer may benefit from the addition of EGF receptor (EGFR) inhibitor gefitinib to chemoradiation. This study was designed to assess improved outcomes and identify predictive biomarkers.Experimental Design: Patients provided informed consent for tumor biomarker analyses and, when eligible, were further enrolled in the therapeutic CARISSA multicenter randomized phase II trial of postoperative irradiation with cisplatin + gefitinib (GORTEC 2004-02-NCT00169221).Results: Seventy-nine patients were included in the biomarker study, whereas 27 did not meet prerequisites for randomization between gefitinib and placebo. Two-year disease-free survival (DFS) rate was 65.0% and did not differ between randomized patients treated with gefitinib or placebo (P = 0.85). The similarity of DFS curves between nonrandomized patients (n = 27), randomized patients without gefitinib (n = 27), and randomized patients receiving gefitinib (n = 25), and similar histoclinical parameter distributions for all groups, allowed us to conduct statistical analyses on the entire population. On multivariate analysis, elevated expression of PAK1 by Western blotting, CD31 and membranous insulin-like growth factor 1 receptor (IGF1R) both by immunohistochemistry was significantly associated with shorter DFS. There was a significant interaction between IGF1R and gefitinib. Gefitinib abolished the prognostic discriminative power of high IGF1R expression; patients with elevated IGF1R expression benefited from gefitinib whereas those with low IGF1R fared worse.Conclusion: Gefitinib treatment affords no significant clinical benefit on DFS in an unselected population of patients with head and neck cancer. Our results point to the potential advantage of personalizing treatment for gefitinib based on tumoral IGF1R expression. This should foster confirmatory analyses in trials involving EGFR-targeting agents. Clin Cancer Res; 18(18); 5123–33. ©2012 AACR.