The significance of micro-EGFR T790M mutation on EGFR-TKI efficacy in patients with NSCLC: The WJOG13119L study

e21177 Background: The significance of micro-pretreatment EGFR T790M mutation on EGFR-tyrosine kinase inhibitor (TKI) efficacy in patients with non-small cell lung cancer (NSCLC) remains uncertain. In addition, there has been no method to examine the micro-T790M mutation after excluding formalin-fixed and paraffin-embedded (FFPE)-derived artificial mutations. Therefore, we used a novel method to examine the variant allele frequency (VAF) of T790M mutation in the FFPE samples after excluding the artificial mutations. Subsequently, we retrospectively investigated the association between micro-T790M mutations and time to treatment failure (TTF) in patients treated with 1st, 2nd, or 3rd generation (1G, 2G, 3G) EGFR-TKIs. Methods: We enrolled patients with NSCLC who harbored EGFR exon 19 deletion (Del19) or L858R without T790M mutation using conventional methods. Patients who received 1G or 2G EGFR-TKIs until June 2018 or 3G EGFR-TKI until June 2019 from 31 institutions were included. Droplet digital PCR was performed using pretreatment FFPE tumor samples. We defined cases as micro-T790M positive when the VAF of T790M was higher than that of F795F, owing to the FFPE-derived artificial mutations. The ratio of micro-T790M mutation was calculated as follows: {(VAF of T790M) – (VAF of F795F)} /VAF of (Del19 or L858R). The lowest value was set as 0. Results: A total of 110, 102, and 103 patients received 1G, 2G, and 3G EGFR-TKIs, respectively. Of these patients, 48.1%, 47.0%, and 47.6% were classified as micro-T790M positive, respectively. The ratios of micro-T790M mutation, represented as median (%) (interquartile range), in 1G, 2G, and 3G EGFR-TKI groups were 0.27 (0.12-0.87), 0.28 (0.13-0.47), and 0.42 (0.20-0.68), respectively. The VAF of T790M or F795F in all patients, represented as median (%) (interquartile range), was 0.26 (0.19-0.35) or 0.25 (0.19-0.33), respectively. The TTF was not significantly different between the T790M-positive and negative groups in analysis based on each generation TKIs. In the Del19 group, the TTF of the 1G EGFR-TKI was significantly shorter, while that of the 3G EGFR-TKI was significantly longer in the T790M-positive group than in the negative group ( p= 0.02, 0.04, respectively). In the T790M-positive group, the TTF of the 2G or 3G EGFR-TKI group was significantly longer than that of the 1G EGFR-TKI group ( p= 0.04, 0.01, respectively). This result was observed in the Del19 group but not in the L858 group. Furthermore, TTF did not significantly differ between the 1G, 2G, and 3G EGFR-TKIs in the T790M-negative group. Conclusions: This study showed that micro- EGFR T790M mutations influence the EGFR-TKI efficacy. Detection of micro-T790M mutations will facilitate selection of the optimal EGFR-TKI for patients with EGFR mutated NSCLC.