Résumé : The prognostic value of epidermal growth factor receptor (EGF-R) for survival of patients with lung cancer remains controversial. The authors performed a systematic review of the literature in order to clarify its impact. Published studies were identified using an electronic search in order to aggregate the available survival results, after a methodological assessment using a scale specifically designed by the European Lung Cancer Working Party (ELCWP). To be eligible, a study had to have dealt with EGF-R assessment in lung cancer patients on the primary site and to have analysed survival according to EGF-R expression. Among the 16 eligible studies, 14 assessed any nonsmall-cell lung cancer (NSCLC) subtype, one adenocarcinoma only and one squamous-cell carcinoma only. The overall median quality score was 56.3%, with no significant difference either between studies assessable or not assessable for meta-analysis or between studies with significant and nonsignificant results. One individual trial reported a survival benefit for patients with EGF-R expression, three a survival disadvantage and 12 no statistically significant difference. Eleven studies (2,185 patients) provided sufficient data to allow a meta-analysis of the survival results. EGF-R expression positivity was determined according to the cut-off as determined by the authors. The meta-analysis showed that EGF-R expression was not a statistically significant prognostic factor for survival in NSCLC. In the subgroup of studies using immunohistochemistry, statistical tests reached a significant level against EGF-R. Epidermal growth factor receptor might be a poor prognostic factor for survival in nonsmall-cell lung cancer. The amplitude of the impact is small, however, and may be subject to publication bias.