Oncogenic EGFR amplification in EGFR-mutated NSCLC: EGFR subtype–dependent association with benefit from first-line osimertinib.

Abstract

e20534 Background: De novo oncogenic amplification, including EGFR and MET represents a biologically relevant layer of genomic heterogeneity in EGFR -mutated non–small cell lung cancer (NSCLC) and may influence the clinical efficacy of first line (1L) osimertinib. However, the clinical significance of baseline de novo oncogenic amplification at treatment initiation remains incompletely understood. Methods: We retrospectively identified patients with advanced EGFR -mutated NSCLC who underwent The Oncomine Dx Target Test (DxTT) at National Cancer Center Hospital, Japan (Jan 2020–Aug 2025) and received 1L osimertinib. We reanalyzed the DxTT-generated BAM files using Ion Reporter Software v5.18 (NCCH Oncomine Focus v1.2 workflow). Copy number variants (CNV) amplification such as EGFR and MET was defined as a 5th-percentile estimated copy number ≥4 (≥2 copies above diploid) . PFS was assessed using the Kaplan–Meier method and the Cox proportional hazards model. Data cutoff: January 11, 2026. Results: A total of 109 patients were included in the study. The median age (range) was 66 years (30-86); 76 patients (69.7%) were female and 64 (58.7%) were never smokers. The primary EGFR mutation was ex19 del in 51 (46.8%) and L858R in 50 (45.9%). EGFR amplification was the most frequent (26/109, 23.9%), whereas other potentially actionable amplifications were uncommon, including FGFR1 (3/109, 2.8%), MET (2/109, 1.8%), and ERBB2 (2/109, 1.8%). Overall, PFS did not differ significantly by EGFR amplification status (median PFS [ EGFR amp+ vs. amp-]: 18.4 vs. 15.6 months; p=0.63). In subtype stratified analyses, EGFR amplification was associated with longer PFS in the ex19del subgroup but shorter PFS in the L858R subgroup. An adjusted Cox model showed a significant EGFR amplification-by-subtype interaction (p=0.031). The effect of EGFR amplification (amp+ vs. amp−) differed by subtype: ex19del HR 0.41 (95% CI, 0.16-1.04; p=0.061) versus L858R HR 1.72 (95% CI, 0.66-4.45; p=0.266). Among EGFR amplification positive patients, EGFR copy numbers (CN) were widely distributed (median 11.1; IQR 7.8-22.8; max 106.9), with CN 4-<10 in 42.3%, CN 10-<20 in 26.9%, and CN ≥20 in 30.8%. No clear difference in PFS was observed according to EGFR CN high vs. low as the cutoff 11.1 (HR, 0.75; 95% CI, 0.26–2.19; p = 0.60). In this EGFR amplification subgroup, 8 (30.8%) harbored co-occurring gene amplifications ( MYC [n=3]), and PFS did not differ significantly between EGFR amplification only and EGFR amplification plus other amplifications (median PFS, 13.1 months [95% CI, 8.8-NR] vs. 32.2 months [95% CI, 4.4-NR]; p=0.48). Conclusions: The prognostic association of de novo EGFR amplification with 1L osimertinib benefit appears to be dependent on EGFR subtype, which supports an interpretation and molecular stratification that integrates EGFR subtype and amplification status, beyond EGFR mutation status alone.