EGFR T790M mutation testing within a phase I study with AZD9291.

Abstract

e19051 Background: EGFR tyrosine kinase inhibitors (TKIs) have demonstrated clinical benefit in patients (pts) with advanced EGFR mutation-positive NSCLC. Eventually pts develop acquired resistance and disease progression, most commonly associated with the EGFR T790M mutation in 50-60% of tumors. AZD9291 is a potent, oral, third-generation, irreversible inhibitor of EGFR sensitising (EGFRm) and resistance (T790M) mutations with a reduced potency against the wild-type EGFR. Within the current phase I AURA study (NCT01802632), patients who have progressed following an EGFR-TKI are enrolled into expansion cohorts based on T790M mutation status. Determination of T790M and EGFRm mutation status were performed either at a central testing laboratory (lab), or at a local lab with central lab confirmation of the result. Methods: Currently, 3 central labs are performing T790M testing using the Roche Molecular Diagnostics cobas EGFR Mutation test. Local testing was performed using a variety of mutation detection methods. Testing is performed on fresh biopsy material taken following progression on the latest line of therapy. In order to maximise assay success rates, the sample area is used to estimate the number of tissue sections required for analysis. As of DCO, 135 samples had been tested centrally; of these 102 had both central and local results. Results: 97% (131/135) of samples were analysed successfully despite biopsy samples being small, with a median area of 12mm2. A median of five 5mm sections were used for DNA extraction. The prevalence of centrally tested T790M mutations in this study was 67% (88/131) with no meaningful difference between Asian (67% [57/85]) and Caucasian (68% [31/46]) patients. Overall, concordance between central and local testing was 91% (93/102). Conclusions: Despite limited sample material, the cobas EGFR test proved to be a reliable testing platform in our study. The prevalence of T790M mutations in post EGFR-TKI patients was 67%, and were consistent between Asian and Caucasian subgroups. Whilst local lab testing was largely concordant with the central result, the variability in the sensitivity of local testing methods led to some inconsistencies when compared with the central result. Clinical trial information: NCT01802632.