Objective EGFR-TKIs have been the first-line treatment for advanced non-small cell lung cancer patients with EGFR mutations.Chemotherapy remains the standard treatment for patients without T790M mutation after the progression of first-line EGFR-TKIs.Previous retrospective studies showed that NSCLC patients with L858R mutation seemed to benefit more from pemetrexed-based chemotherapy than patients with 19Del mutation,but little literature reported the optimal chemotherapy regimen for NSCLC patients with common EGFR mutations.Methods 4,021 lung adenocarcinoma patients treated at the Guangdong Lung Cancer Institute from June 2007 to June 2014 were screened in the present study.EGFR common mutations were defined as EGFR exon 19 deletion(19Del)or EGFR exon 21 point mutation(L858R mutation).Inclusion criteria: 1.Neither less than 18 years old nor more than 70 years old; 2.Stage ⅢB-Ⅳ patients without indications for radical surgery or radical chemoradiotherapy; 3.Had received systemic chemotherapy; 4.Had at least one assessable lesion.Exclusion criteria: 1.Patients without efficacy evaluation; 2.Had other cancer history; 3.Concurrently treated with bevacizumab or other antiangiogenic drugs; 4.Patients with wild type EGFR or uncommon EGFR mutations.The effect of chemotherapy on chemotherapy-naive patients with EGFR common mutations was analyzed in this retrospective study.Kaplan-Meier curves and the log-rank test were used to analyze progression-free survival(PFS)and overall survival(OS)according to EGFR mutation subtypes and treatment regimens.The variables age,gender,PS,smoking history,EGFR mutation subtypes,treatment regimens and brain metastasis status were entered into Coxs proportional hazard models to predict the hazard ratios for PFS.Differences in objective response rates(ORR)and disease control rates(DCR)according to the EGFR mutation subtypes and treatment regimens were compared by the chi-squared test.The statistical significance level was set at P <0.05.Results Two hundred and five lung adenocarcinoma patients detected with common EGFR mutations were included in the retrospective analysis(Table 1).Among the 205 patients,124 were 19Del(124/205,60.5%),81 were L858R mutation(81/205,39.5%).Patients were treated with gemcitabine-based(G,n=119),pemetrexed-based(P,n=60)and taxane-based(T,n=26)chemotherapy,only 10(10/205,4.9%)patients received maintenance chemotherapy.Baseline characteristics of patients were similar except for brain metastasis,which was more common in Group L858R mutation(P=0.035).The ORR were 26.4%(19/72),34.2%(13/38),28.6%(4/14)and 23.4%(11/47),9.1%(2/22),33.3%(4/12)for 19Del+G,19Del+P,19Del+T and L858R+G,L858R+P,L858R+T respectively.No statical significance difference was observed in ORR and DCR according to mutation subtypes and treatment regimens(P< 0.395 for ORR; P< 0.745 for DCR).The median PFS in the L858R mutation group(5.8 months,95%confidence interval [CI]: 5.2-6.4)was slightly longer than 19Del group(5.3 months,95%CI,4.7-5.9)with no statistical difference(P = 0.614).No statistically significant difference was observed in PFS among three treatment groups(P = 0.947).In the multivariate analysis,the PFS was significantly longer in patients without brain metastasis than in patients with brain metastasis at baseline(P = 0.005,HR: 1.715,95%CI 1.174-2.505),other than age,gender,PS,smoking history,treatment regimen and EGFR mutation types.OS was similar between 19Del and L858R mutation(median overall survival: 30.7 months vs.32.1 months,P = 0.979).Conclusions Similar PFS duration was observed among gemcitabine-based,pemetrexedbased and taxane-based chemotherapy in lung adenocarcinoma patients with EGFR common mutations,regardless of EGFR mutation subtypes.Brain metastasis was an independent risk factor for prognosis.