Abstract B32: Retrospective correlative study between EGFRvIII, HPV, p16, c‐MET and response to EGFR inhibitors in patients with recurrent or metastatic squamous cell carcinoma of the head and neck (R/M SCCHN)

Background: No validated biomarkers exist to predict the response to EGFR inhibitors in SCCHN. A constitutively activated mutant, EGFRvIII, confers oncogenicity in human cancers and resistance to EGFR blockade, and is detected in 42% of 33 SCCHN tumors (Sok et al. Clin Cancer Res 2006). The aim of this study is to confirm the prevalence of EGFRvIII, HPV, p16, c‐MET in R/M SCCHN and evaluate their potential prognostic and predictive role. Materials and Methods: Archival tumor specimens of 53 patients (pts) who were treated in 4 phase I/II trials for R/M SCCHN at Princess Margaret Hospital from 2000–2005 were examined. Two of the 4 trials involved the EGFR inhibitor erlotinib (tumor specimens available in 35 of 48 pts) whereas the remaining 2 trials involved non‐EGFR targeted agents (tumor specimens available in 18 of 37 pts). EGFRvIII mutation was determined by quantitative RT‐PCR (positive result determined by decreased expression of exon 4 compared with exon 9 of the EGFR gene), presence of HPV DNA by Linear Array Genotyping, p16 and c‐MET expression by immunohistochemistry, using p16CINtec (Westborough, MA) and SP44 (Ventana, Tuczon, AZ) antibodies, respectively. Results: Demographics of the 53 pts were: median age 56 (range 15–78), F:M (%) = 23:77, ECOG 0:1:2 (%) = 28:64:8, locoregional recurrence = 85%, metastatic disease = 36%, oropharyngeal primary = 38%. Overall response rate (CR+PR) of the entire cohort to study treatment = 4/53 (7.5%), median TTP = 1.8 months, median OS = 5.9 months. Univariate analyses were significant for erlotinib‐treated pts compared to non‐erlotinib treated pts in both TTP and OS. EGFRvIII was detected in 22 pts (42%); median fold change was 6.8 (0.56–576.36). The presence of EGFRvIII mutation was associated with better disease control (PR+SD) on univariate analysis (p = 0.01), but no difference was seen between erlotinib‐treated versus non‐erlortinib treated pts. Median EGFRvIII fold changes were higher for pts with PR+SD than pts with PD (11.11 vs 3.16, p=0.04). The presence of EGFRvIII mutation was not associated with TTP or OS. HPV DNA (16, 6 or 33) was detected in 20 pts (38%), p16 immunostaining was present in 17 pts (32%) and c‐MET was highly expressed (IHC score >2) in 31 pts (58%). HPV, p16 and c‐MET were not associated with response, TTP or OS. Conclusions: This retrospective study confirms the presence of EGFRvIII mutation in about 40% of SCCHN, and it appears to be a prognostic biomarker associated with better disease control in R/M SCCHN regardless of treatment with erlotinib. Interestingly, the presence of activating mutations conferring a better prognosis has been reported with EGFR mutations in NSCLC (Eberhand et al. J Clin Oncol 2005) and with PIK3CA mutations in breast cancer (Kalinsky et al. Clin Cancer Res 2009). The positive prognostic value of EGFRvIII in R/M SCCHN is unexpected and large prospective studies are required to validate its significance. Citation Information: Mol Cancer Ther 2009;8(12 Suppl):B32.