Review Girard 33.3 vs 21.1 months, p = 0.0015; LUX-Lung 6: median, 31.4 vs 18.4 months, p = 0.023 [ Table 1 ]) [26] . It was hypothesized that the observed OS benefit may be related to broad inhibition of the ERBB family by afatinib compared with reversible inhibition of EGFR only by first-generation EGFR TKIs. In addition, given the availability of first-generation TKIs at the time those trials were conducted, a significant proportion of patients did receive additional lines of TKIs after the study treatment, which may have impacted OS. However, until recently, there were no head-to-head trials comparing first- and second-generation EGFR TKIs to substantiate the hypothesis that second-generation EGFR TKIs were more effective. The exploratory Phase IIB LUX-Lung 7 trial, which compared first-line afatinib with gefitinib in EGFR mutation-positive (Del19 or L858R) NSCLC, was the first prospective, randomized trial to compare first- and second-generation TKIs [28] . Patients received treatment until disease progression, or beyond if deemed beneficial by the investigator. Given the exploratory nature of the trial, no formal hypothesis was specified. Three coprimary end points were selected: PFS by independent central review, OS and time to treatment failure (TTF). TTF, which was defined as the time from randomization until treatment discontinuation for any reason, including progression, AEs or death, was included to reflect clinical practice where many patients receive treatment beyond radiological progression in the absence of clinical deterioration. In LUX-Lung 7, afatinib significantly prolonged PFS (median, 11.0 vs 10.9 months; hazard ratio [HR]: 0.73; p = 0.017) and TTF (median, 13.7 vs 11.5 months; HR: 0.73; p = 0.0073) versus gefitinib. A significantly higher proportion of patients achieved an objective response with afatinib (70%) versus gefitinib (56%; Table 1 ) [28] . Median OS was not significantly different between treatment arms (27.9 months with afatinib vs 24.5 months with gefitinib; HR: 0.86; p = 0.2580). However, LUX-Lung 7 was not powered for OS, as the sample size was based on controlling the width of the CI for the HR of PFS [29] . As in LUX-Lung 3 and LUX-Lung 6, the efficacy findings were generally consistent across patient subgroups of gender, age and ethnicity. Furthermore, the efficacy benefits of afatinib over gefitinib were not affected by EGFR mutation type [28 , 29] . More recently, another second-generation EGFR TKI, dacomitinib, has been assessed in the first-line setting versus gefitinib in the Phase III ARCHER 1050 trial [34] . This study enrolled patients with EGFR mutation-positive (Del19 or L858R) NSCLC, but excluded patients with brain metastases. Dacomitinib significantly prolonged PFS compared with gefitinib (14.7 vs 9.2 months; HR: 0.59; p < 0.0001; Table 1 ). While the proportion of patients who achieved an objective response was similar between the two groups (75% in the dacomitinib group and 70% in the gefitinib group), duration of response was longer with dacomitinib than gefitinib (HR: 0.55; p < 0.0001). OS data were not mature at the time of data cutoff. Although cross trial comparisons of LUX-Lung 7 and ARCHER 1050 should be cautioned, especially given differences among patient populations, the cumulative evidence suggests that second-generation EGFR TKIs may be more effective than first-generation EGFR TKIs. Third-generation EGFR TKIs in patients with acquired resistance Despite the efficacy of first- and second-generation EGFR TKIs, resistance inevitably develops, usually within 10–14 months after starting treatment. The predominant mechanism of resistance appears to be acquisition of EGFR T790M, occurring in 50–70% of patients [6 , 7] . The third-generation EGFR TKIs were developed to target this mechanism, showing high potency for T790M and EGFR TKI-sensitizing mutations, although the pharmacological potency for the common Del19 / L858R mutations is lower with osimertinib than second- generation EGFR TKIs [35 , 36] . A number of third-generation TKIs are in development, but osimertinib is the only agent approved in Europe and the USA for treatment of advanced EGFR T790M mutation-positive NSCLC, based on impressive data from the Phase I / II AURA, Phase II AURA2 and Phase III AURA3 studies. AURA enrolled patients with advanced NSCLC that had progressed after treatment with EGFR TKIs. In the Phase I part, the response rate for osimertinib was 61%, with a median PFS of 9.6 months among patients who had centrally confirmed EGFR T790M [8] . These results were confirmed in the Phase II AURA extension, in patients with EGFR TKI-pretreated EGFR T790M-positive disease, where the response rate was reported as 62% and median PFS was 12.3 months [7] . Similarly, in the AURA2 study in patients who had progression on previous EGFR TKI therapy, the response rate was 70% [37] . Pooled OS analysis from AURA2 and the AURA extension showed median OS of 26.8 months from initiation of osimertinib in patients with T790M-positive NSCLC that had progressed after a first- or second-line EGFR TKI [38] . Osimertinib was subsequently assessed in the randomized Phase III AURA3 trial in patients with T790M-positive advanced NSCLC who had disease progression after first-line EGFR TKI therapy. Patients were randomized to osimertinib or platinum-based therapy plus pemetrexed. Osimertinib significantly prolonged investigator-assessed PFS versus chemotherapy (10.1 vs 4.4 months; HR: 0.30; 95% CI: 10.2217/fon-2017-0636 Future Oncol. (Epub ahead of print) future science group