Résumé
PURPOSE: Epidermal growth factor receptor (EGFR) overexpression is associated with poor prognosis in head and neck squamous cell carcinoma (HNSCC). Despite intensive biomarker studies, a consensual method for assessing EGFR protein expression is still lacking. Here we set out to compare three EGFR detection methods in tumor specimens from HNSCC patients. EXPERIMENTAL DESIGN: Tumors were prospectively excised from a series of 79 high-risk HNSCC patients enrolled in a GORTEC-sponsored clinical trial. EGFR expression was determined using a ligand-binding assay on membranes, Western blotting (WB) on membranes and total homogenates, and immunohistochemistry (IHC) on tissue microarrays. In addition, phosphorylated EGFR (pEGFR) was measured by WB on membranes. RESULTS: Distributions and ranges of tumor EGFR expression were method dependent. Moderate positive correlations (Spearman coefficient r ≈ 0.50) were observed between EGFR expression measured by the binding assay and WB or IHC. pEGFR levels positively and significantly correlated with total EGFR expression measured by WB or ligand binding, but not by IHC. The highest correlation (r = 0.85) was observed between EGFR and pEGFR levels, both measured by WB on membranes. Interestingly, the fraction of phosphorylated receptor (pEGFR/EGFR both measured by WB on membranes) significantly declined with increasing tumor EGFR expression, by all assessment methods used. CONCLUSION: This study shows significant correlations between EGFR detection methods. The observed relationships between EGFR and pEGFR indicate that high-throughput pEGFR/EGFR analyses merit further investigations and consideration for routine use in patient samples.