Concurrent L858R and uncommon EGFR exon 21 mutations can be observed in NSCLC. Nevertheless, the detection characteristics and treatment response of these patients remain elusive. In this study, the most common EGFR exon 21 mutations that coexisted with L858R mutations were V834L (24.07%, 13/54), K860I (18.52%, 10/54) and L833V (14.81%, 8/54), all of which could not be detected by ARMS-PCR. Especially, PCR negative but NGS positive alterations were observed for concurrent EGFR L858R and K860I mutations, and NSCLC patients with these mutations identified by NGS are able to benefit from EGFR-TKIs.
ARMS-PCR is a classical approach to selectively amplify targeted DNA sequences and is widely applied to identify EGFR mutations. However, it is not able to identify uncommon or unknown mutations(Cai et al. 2019; Sousa et al. 2020). It was reported that more than 40% of NSCLC patients harboring EGFR ex20ins mutations may be undetected by ARMS-PCR assays(He et al. 2022; Ou et al. 2023). Our previous study found that around 5% NSCLC patients carried uncommon EGFR mutations that could not be identified by ARMS-PCR(Li et al. 2018). Nevertheless, NGS enables ambitious large-scale genomic sequencing efforts with its massively parallel sequencing capacity and high sensitivity(Koboldt 2020, 2021). In this study, we observed that NSCLC patients with concurrent uncommon EGFR exon 21 mutations identified by NGS could not be identified by ARMS-PCR. Interestingly, completely negative results were observed for patients with concurrent L858R and K860I mutations identified by NGS. Moreover, high concordance was observed between the MAFs of L858R and K860I, indicating that L858R and K860I mutations might be on the same allele from the same mutant clone. Thus, the base substitution sites of c.2573 T > G and c.2579 A > T are too close for the primers of ARMS-PCR to bind, leading to a completely false negative result. However, The hybrid capture-based NGS analyzes gene mutations with probes can detect both L858R and K860I.
Some previous studies explored the association of complex EGFR mutations with the efficacy of EGFR-TKI treatment. For example, Zhang’s study analyzed the treatment response of EGFR-TKIs at first-line setting for advanced adenocarcinoma patients with concurrent EGFR Del-19/21 L858R and uncommon EGFR mutations, the median PFS was 9.7 months and ORR was 60.0%(Zhang et al. 2018). However, the treatment efficacy of EGFR-TKIs among patients with concurrent L858R and uncommon EGFR exon 21 mutations is largely unknown. Here, our study revealed that NSCLC patients with concurrent L858R and uncommon EGFR exon 21 mutations are able to benefit from EGFR-TKIs. Importantly, patients with concurrent L858R and K860I mutations, which may be detected by NGS but not by ARMS-PCR, are associated with good response to EGFR-TKIs with a median PFS of 11.0 months and a median DFS of 10.0 months. Furthermore, we noticed that one patient with concurrent L858R and K860I mutations had a 11-month partial response to Dacomitinib before acquiring T790M, which is the most common resistant mechanism among patients with common EGFR mutations.
Our study has some limitations: On one hand, the number of NSCLC patients with concurrent L858R and uncommon EGFR exon 21 mutations in our study is relatively small, and further studies in a larger cohort are warranted to confirm the conclusions. On the other hand, the time of follow-up of these patients treated with second or third generation EGFR-TKIs is relatively short.
In summary, our study found that ARMS-PCR failed to identify concurrent L858R and K860I mutations in NSCLC, while NGS provided comprehensive information to optimize clinical outcomes. In clinical practice, if we merely rely on the results of ARMS-PCR, patients with concurrent L858R and K860I mutations may miss the chance to benefit from EGFR-TKIs, confirmed to be highly effective in this study. Thus, NGS should be recommended for NSCLC patients to identify driver alterations such as EGFR mutations, especially when negative results are obtained by ARMS-PCR.