Abstract

TKIs have significantly improved the survival of NSCLC pts carrying sensitive mutations. However, heterozygous responses were observed. We conducted a prospective multicenter clinical trial to explore factors associated with the efficacy of EGFR-TKI, and assess the mutation and TMB concordance between plasma and tissue NGS. Paired tumor and plasma samples were obtained from treatment naïve advanced NSCLC pts whenever applicable. DNA was sequenced by target-capture deep sequencing of 1021 tumor-related genes (pan-cancer panel). PFS was estimated using Kaplan-Meier method and compared using log-rank test. Tissue TMB (tTMB) and plasma TMB (bTMB) analysis interrogated SNVs/Indels with VAF ≥3% and ≥0.5%, respectively. TMB-H pts were identified with ≥9 muts/Mb. From Feb. 2017 to Jan. 2019, 262 advanced NSCLC pts were enrolled from 12 centers. In 224 pts with paired tumor and plasma samples, 144 had EGFR sensitive mutations in tumor samples (L858R, 46%; Ex19Indel, 42%), of whom, 106 (74%) had the identical mutations detected in plasma. The detection rate of tissue EGFR mutations in paired plasma was significantly higher in pts with extrathoracic metastasis (81% vs. 61%, p = 0.03). In 38 pts lacking paired samples, 20 pts had EGFR sensitive mutations detected. Thus, 164 pts were identified as EGFR positive by either plasma or tissue NGS. One hundred of them were treated EGFR TKIs (ORR: 70%, mPFS: 20 mo). The ORR was affected by EGFR subtypes (Ex19Indel vs. L858R: 72% vs. 45%, p = 0.02), concomitant CNV/fusion (with vs. without: 11% vs. 68%, p = 0.002) and CDKN2A mutations (with vs. without: 0% vs. 66%, p = 0.007). Mutations in p53 pathway (p = 0.02), CDK12/13 (p = 0.0002), concomitant CNV/fusion (p = 0.003), and high number of alterations (≥ 5) (p = 0.003) significantly shortened mPFS. tTMB was correlated with bTMB (rPearson = 0.9, p <0.0001), with a concordance rate of 90% for TMB-H and TMB-L classification. Interestingly, 9.8% of the EGFR positive pts were bTMB-H, and mPFS was shorter in bTMB-H pts (6 mo, 95% CI: 5 - NR) than in bTMB-L pts (NR, 95%: 13 - NR) (p = 0.2). Deep sequencing with the pan-cancer panel effectively detected mutations and evaluated TMB in both tissue and plasma with a high consistence. Moreover, the presence of structure variation, high tumor heterogeneity and concomitant mutations in genes such as CDKN2A were associated with worse prognosis. Further studies of predictive factors are ongoing (NCT03059641).

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call