The loss of secondary mutation T
The loss of secondary mutation T790M mutation after AR and SCLC transformation was also reported for Rociletinib. Piotrowska et al. using the MGH NGS platform recently reported in 13 biopsies among 12 EGFR T790M+ patients progressing after treatment with Rociletinib the loss of EGFR T790M mutation in 6 (with evidence of SCLC transformation plus RB1 loss in 2/6 cases) and persistence of T790M mutation in 7 samples (with EGFR amplification in 3/7 tissue biopsies), respectively. No C797S positive cases were reported, probably due to the small sample size and the population evaluated in the study: both primary and secondary resistant patients to Rociletinib were included and mostly with poor response to the treatment. Interestingly, using a patient-derived cell line from an Afatinib-resistant NSCLC, the authors also demonstrated the intratumor heterogeneity of T790M mutations within the same sample “T790M positive”, providing a possible explanation for the apparent “loss” of this mutation at the AR to Rociletinib: it is likely expression of selection of pre-existing T790M-wild type clones rather than mutation/deletion of T790M AMI-1 synthesis (Piotrowska et al., 2015). Moreover, a recent study using a CAPP-seq ctDNA, analysis, allowing simultaneously study of single-nucleotide variants (SNVs), insertions/deletions, rearrangements, and somatic copy-number alterations (SCNAs), revealed a largely underestimate heterogeneity of mechanisms of acquired resistance to 3rd generation EGFR TKIs, with co-existence of multiple mechanisms in the same patient at higher frequency than previously reported (46% in T790M-mutant vs. 5–15%) (Chabon et al., 2016). These findings have important clinical implications since intrapatient heterogeneity may negatively affect clinical response to 3rd generation EGFR TKis.
Future perspectives The development of 3rd generation EGFR TKIs poses novel challenges in the therapeutic management of EGFR-mutant NSCLCs (Fig. 1). One of the possible future scenarios is the use of mutant-selective EGFR TKIs in the 1st line setting. Front-line use of use of these agents might provide several advantages over the current approved EGFR TKIs: activity against EGFR mutants, including T790M, while sparing EGFR wild type and delay of acquired resistance, as demonstrated in in vitro models (Ramalingam et al., 2014), with a better central nervous system (CNS) penetration. Indeed, Ballard et al. recently reported in a mouse model that Osimertinib is distributed in the CNS at greater extent than Gefitinib, Rociletinib and Afatinib (Ballard et al., 2016), providing the rationale for the use of Osimertinib in EGFR-mutant patients with brain metastases as first line option. Several different clinical trials are evaluating the role of EGFR-mutant selective EGFR TKIs in TKI-naïve patients [Table 4]. Preliminary data from two cohorts of TKI-naïve patients enrolled into the phase I AURA trial treated with Osimertinib at 80mg or 160mg daily were recently reported, showing a promising activity in this subset of patients: 67% ORR in the 80mg cohort and 77% in the 160mg cohort, 97% DCR overall and a median PFS of 19.3 months in the overall population (median not reached in the 80mg cohort and 19.3 months in the 160mg cohort) (Ramalingam et al., 2016). The FLAURA trial (NCT02296125) is an ongoing double blind, randomized, phase III study comparing Osimertinib 80mg/d with standard 1st generation EGFR TKIs, Gefitinib and Erlotinib. Rociletinib is being evaluating in both treatment-naïve and pretreated patients in the phase II part of TIGER1 trial and in combination with the anti-PDL1 inhibitor Atezolizumab in a phase Ib/II (NCT02630186). ASP8273 is also being compared with first generation EGFR TKIs in the 1st line setting in the open label phase III trial SOLAR (NCT02588261). Another challenge posed by 3rd generation EGFR TKIs is the re-biopsy of patients with acquired resistance to 1st/2nd generation EGFR TKIs. Albeit tumor tissue genotyping is the gold standard, a re-biopsy is not always feasible in clinical practice because of scheduling problems, costs, risks of complications and issues related to tissue acquisition and preservation (LA Jr and Bardelli, 2014). Moreover, tumor heterogeneity may represent an obstacle for tumor genotyping, since single region sampling may underestimate the genomic complexity of a tumor (Gerlinger et al., 2012 8, de Bruin et al., 2014b). Liquid biopsy may overcome some of the limits of traditional tissue biopsy and in particular plasma genotyping of circulating tumor DNA (ctDNA) has shown promising results.