6041 Background: Dacomitinib is an irreversible pan-HER TKI with preclinical (EGFRvIII+ cell lines, SCCHN xenografts) and clinical activity (phase II recurrent/metastatic SCCHN; Razak et al, Ann Oncol 2012). However, little is known about predictive markers of efficacy related to EGFR signalling in this setting. Methods: Of69 pts treated with 1st-line dacomitinib in a phase II trial for recurrent/metastatic SCCHN, 48 pts had archival tumor specimens obtained before treatment and 13 had paired biopsies (days 0 and 7 of therapy, FFPE and snap frozen). EGFRvIII and PTEN (IHC), HPV genotyping and human genomic mutations (Sequenom OncoCarta Panel – 19 genes, 238 mutations) were evaluated on archival tissue. IHC expression of AKT, CC3, EGFR, ERK, HER2, HER3, MET, Ki67, pAKT, pEGFR, pERK, pHER2 and pMET was evaluated in paired specimens. The presence/absence or expression level of these markers was correlated with response (RR)/clinical benefit (CB), PFS and OS. Results: In pts with archival tissue, no statistically significant difference was found in RR/CB or PFS based on HPV, EGFRvIII, PTEN or presence of mutation. There was a trend to increased OS in HPV+ pts (HR 0.47, 95% CI 0.21–1.07, P=0.068). In paired biopsies, some expression variation was seen for cytoplasm AKT, membrane EGFR, nuclear ERK and pAKT. There was no correlation between basal expression of these markers and RR/CB or PFS. Variations in ratio to baseline of EGFR, pAKT, pERK and MET were qualitatively associated with RR/CB. No statistically significant correlations could be established for PFS, but there were interesting qualitative variations in the levels of expression of some molecules, eg, EGFR, pAKT. Conclusions: No predictive efficacy marker was identified. It cannot be determined if increased OS in HPV+ cases is due to prognostic or predictive effects. Paired biopsies demonstrated that dacomitinib was associated with variation in expression of multiple elements in signalling pathways linked to EGFR. Given the small number of paired biopsies, and large amount of data generated, descriptive study of cases is required. Further data will be presented at the meeting. Clinical trial information: 00768664.