Clinical significance of <i>EGFR</i> amplification in patients with <i>EGFR</i> -mutated metastatic non–small cell lung cancer receiving first-line osimertinib.

Abstract

8643 Background: With rapidly expanding first-line options for patients (pts) with EGFR -mutated non-small cell lung cancer (NSCLC), identifying biomarkers that may assist treatment selection is critical. The impact of EGFR amplification ( EGFR AMP ) on osimertinib outcomes is unclear. Methods: Pts with stage IV NSCLC and EGFR exon 19 deletions (ex19del) or the L858R mutation who received first-line osimertinib monotherapy at 5 centres across Italy and the United States and had undergone baseline next-generation sequencing (NGS) that included EGFR AMP assessment were included in this analysis. EGFR AMP was defined as an EGFR copy number (CN) ≥6. The predominantly amplified allele was inferred by INCOMMON, a biostatistical classifier, as previously described. Results: Among 473 pts, 81 (17.1%) had EGFR AMP . Compared to pts with non-amplified EGFR ( EGFR Non-AMP , n = 392), pts with EGFR AMP more frequently had TP53 co-mutations (80% vs 55%, p &lt; 0.001) and baseline brain (51% vs 34%, p = 0.008), liver (26% vs 13%, p = 0.009), and bone metastasis (65% vs 51%, p = 0.03). When treated with osimertinib, pts with EGFR AMP achieved similar objective response rate (ORR) (88% vs 83%, p = 0.23), but shorter median progression-free survival (mPFS) (11.6 vs 19.0 months, HR 1.77, p &lt; 0.0001) and overall survival (mOS) (34.0 vs 40.1 months, HR 1.40; p = 0.040). In a multivariable Cox-regression model, EGFR AMP retained its association with worse PFS (HR 1.36, p = 0.04), but not OS. EGFR AMP correlated with shorter PFS in both TP53 co-mutated (n = 269) (11.7 vs 15.0 months, HR 1.43, p = 0.02) and TP53 wild-type (n = 181) (10.4 vs 21.9 months, HR 2.38, p &lt; 0.001) cases, despite similar ORR and mOS. Among pts with ex19del, EGFR AMP (n = 44) showed similar ORR compared to EGFR Non-AMP (n = 241), but shorter mPFS (14.2 vs 20.2 months, HR 2.01, p &lt; 0.001) and mOS (35.4 vs 44.9 months, HR 1.62, p = 0.04). In contrast, no difference was observed in the L858R-mutated subgroup between EGFR AMP (n = 39) and EGFR Non-AMP (n = 150) cases (ORR 86% vs 75%, p = 0.14; mPFS 11.4 vs 14.3 months, HR 1.47, p = 0.06; mOS 33.2 vs 36.0 months, HR 1.19, p = 0.5). Within EGFR AMP cases, increasing EGFR CN ( EGFR Non-AMP vs CN≥6 &lt; 15 vs CN 15-30 vs CN &gt; 30) correlated with a stepwise reduction of mPFS (19.0, 16.7, 10.3, and 8.7 months, respectively; log-rank p &lt; 0.001) and mOS (40.1, 37.5, 38.8, and 15.3 months, respectively; log-rank p = 0.02). Moreover, amplification of the mutant allele, as opposed to wild-type amplification, correlated with inferior mPFS (8.3 vs 12.1 months, HR 2.72, p = 0.002) and mOS (17.3 vs 39.7 months, HR 2.08, p = 0.046). Among pts with paired NGS before and after acquired osimertinib resistance (n = 113), those with baseline EGFR AMP more frequently showed acquired MET alterations (29% vs 12%, p = 0.04). Conclusions: EGFR AMP is associated with distinctive characteristics and worse outcomes to osimertinib among pts with stage IV EGFR -mutated NSCLC.