Evaluation of never-smoking status and male gender as clinical predictive factors for survival with erlotinib treatment of non-small cell lung cancer.

Abstract

e18134 Background: In non-small cell lung cancer (NSCLC), EGFR-mutation status predicts response to treatment with erlotinib. However, often EGFR-mutation analysis is not available. Therefore, clinical response predictors are needed. Available data suggest that never-smokers, females and patients with adenocarcinoma (AC) histology may benefit most from treatment with erlotinib. Methods: Using Cox proportional-hazard regression, we retrospectively analyzed 121 consecutive patients with advanced inoperable NSCLC treated with erlotinib to test whether these clinical parameters predict response and survival in a routine setting. Results: In our Caucasian population, partial remission (PR) was achieved in 16.5% with higher PR rates for never-smokers (46.9%), females (23.1%, males 13.4 %) and AC histology (25.0%). Progression free survival (PFS) was 3.0 months and overall survival (OS) was 7.5 months. Male patients had slightly longer survival than female patients (OS 8.5 vs. 7.0 months). After adjustment for smoking and histology, the gender difference was significant (adjusted hazard ratio (HR) 0.57, CI 0.34-0.97). This unexpected finding may reflect the large proportion of males achieving stable disease (41.5 %, females 17.9 %). Histology had no effect on PFS and OS. Never smokers had a longer PFS and OS (7 and 13 months) than smokers and ex-smokers (PFS 1.75 and 3.5 months, OS 5.5 and 7.5 months). The duration of ex-smoking had no significant effect on PFS (p for trend 0.80). The HR for OS of ever-smokers vs. never-smokers was 2.73 (OS, CI 1.53-4.86). Most useful prognostic parameters during erlotinib treatment were radiological response (OS 23 months, HR for PR vs. SD or PD 0.19, CI 0.09-0.40) and skin toxicity (OS 13.5 months, HR for grade 2 or 3 toxicity vs. grade 0, 1 toxicity 0.54, CI 0.33-0.88). Conclusions: Our results indicate that the clinical predictors (never-smoking status, female gender, AC histology) are useful to predict PR to erlotinib. However, only never-smoking status and male gender were associated with an improved OS. During treatment, objective response and higher degree skin toxicity predict prolonged OS. No significant financial relationships to disclose.