EGFR Variables Do Not Correlate With Response to Treatment With Gefitinib Plus Platinum-Based Chemotherapy

WILMINGTON, Delaware-Epidermal growth factor receptor(EGFR) variables "do not predict forsurvival benefits with gefitinib in combinationwith platinum-based chemotherapy,"for non-small-cell lung cancer (NSCLC), according to Lisa ReneeBailey, PhD, MPH, of AstraZeneca,Wilmington, Delaware. Dr. Bailey wasreporting on an analysis of two phaseIII trials showing that gefitinib (Iressa)combined with first-line platinumbasedchemotherapy did not improvesurvival in NSCLC patients (abstract7013). The study objectives were todetermine whether EGFR is prognosticfor survival, independent of treatmentregimens, and whether it predictsoutcomes for patients treated withgefitinib plus chemotherapy."EGFR expression appears to be aprognostic factor but not a predictorof response to treatment," she said."These results seem to contradict thebelief in the literature."Tumor SamplesAssessedPretreatment tumor biopsies of 516patients from the two trials that didnot show a survival benefit (INTACT1 and 2) were analyzed by immuno-histochemistryfor EGFR. The percentageof tumor cells was assessedusing four levels of intensity: no staining,0; weak, 1+; moderate, 2+; strong,3+; as well as the presence of completemembrane staining. A restricted backwardselimination Cox regressionanalysis was conducted to identify independentEGFR factors, and significantvariables were also tested for treatmentinteraction to determine if theyserved as predictive factors.Dr. Bailey reported that two EGFRfactors were prognostic for survival inpatients treated with gefitinib andthose receiving placebo. A glandular,rather than solid, tumor growth patternand membrane staining were bothassociated with significantly longersurvival. "Perhaps EGFR has a differentrole in early vs late disease?" Dr.Bailey suggested.Neither of these two variables wasuseful for predicting who would respondto gefitinib treatment. "Giventhe overall trial results for INTACT 1and 2 showing no survival benefitbetween gefitinib and placebo, it washighly unlikely that EGFR variableswould identify a subset of patientswho receive survival benefit. The datapresented support this hypothesis,"Dr. Bailey concluded. She also pointedout that clinical use will requireeasier methods for measuring EGFRexpression such as monoclonal antibodiesor fluorescent in-situ hybridization(FISH) assays.