Clinical Impact of the Feasibility of Epidermal Growth Factor Receptor (egfr) Mutations Detection in Cytological Samples of Patients (pt) with Non-Small Cell Lung Cancer (nsclc) Treated with Erlotinib (E) at First Line

e18051 Background: EGFR mutations have been reported to be a predictor of efficacy of new drugs in NSCLC first line treatment. Recently, two phase III trials have shown the efficacy of gefinitib and erlotinib in this context, but the majority of pt were randomized after large biopsy samples. This study try to evaluate the efficacy and safety of E in untreated pt with advanced NSCLC and mutated EGFR (most of them detected in cytological specimens). Methods: Since 2007 we have screened 270 patients and we have found 40 patients with EGFR mutations (14%). 23 patients were treated with E at 150 mg per day at first line treatment. 13 of 23 pt were identify with cytological samples. Results: Median age was 72 years (33-90). Sex ratio (F/M): 74%/26%. 65% were non-smokers and 91% of them had ECOG 0-1. According with histology, 73.9% adenocarcinomas, 8.7% squamous cell carcinomas, 4.3% bronchioloalveolar carcinomas (13% non-specific). Oncogenic EGFR mutations detected were deletion of exon 19 (52.2%) and mutation of exon 21 (47.8%). Response has been evaluated in 18 p so far: complete response, 1 p; partial response, 11 p; stable disease, 3 p; progressive disease, 3 p; overall response rate, 67%; clinical benefit, 84%. In the cytological subgroup: overall response rate, 64%; clinical benefit, 82%. 11 patients continue under treatment with E. Progression free survival for all pt were 11 months, and 10 months for pt with mutations in EGFR gene detected in cytological specimens. Toxicity profile was predictable (mild to moderate rash, diarrhea and stomatitis). Conclusions: Our results suggest that detection of EGFR mutations in cytological samples is feasible and could have clinical impact when these pt are treated with tirosin-kinase inhibitors. A study comparing histological and cytological samples is on going.