Identification of pretreatment genomic biomarkers and mechanisms of acquired resistance to first-line osimertinib in advanced EGFR-mutant lung cancers.

Abstract

9100 Background: Despite its widespread use, few series explore genomic biomarkers that impact progression-free survival (PFS) for first line osimertinib use and how mechanisms of acquired resistance impact post-osimertinib progression survival. Methods: All pts treated with first line osimertinib at Memorial Sloan Kettering Cancer Center were identified; available pre-osimertinib and post-progression tumor samples underwent next-generation sequencing (NGS) using MSK-IMPACT. Real-world (rw)PFS was estimated with Kaplan-Meier methods from start of osimertinib to progression, defined using interpretation of imaging reports. Post-baseline factors were evaluated using time-dependent covariate Cox model. Results: We identified 331 pts, of which 67 had paired tumor samples and 28 had post-progression tumor samples that underwent MSK-IMPACT. Most pts with biopsies were women (68%), never smokers (83%) and did not have baseline brain metastases (56%). With median follow-up of 24 months (mo), median rwPFS was 14 mo (95% confidence interval (CI) 13 -17 mo, n = 331). EGFR driver alterations (n = 40 atypical, n = 108 L858R, n = 182 exon 19 deletion) were associated with distinct rwPFS (median 8 (95% CI 6-12), 14 (12-18) and 16 mo (13-21), respectively, p < 0.001 logrank test). Pts with concurrent pre-treatment TP53 (14 mo,12-17) or TP53/RB1 (12 mo, 9-15) alterations had shorter median rwPFS compared to pts without these alterations (20 mo,16-24, p < 0.001, logrank test). Fifty patients (53%) had an identified mechanism of resistance. Off target mechanisms (n = 9 MET amplification (amp), n = 3 HER2 amp, n = 3 PIK3CA mutations, n = 3 acquired fusions, n = 2 RB1 loss and n = 1 CCND1 amp, MYC amp, CDK4 amp, KRAS G12A, respectively) and histological transformation (n = 7 small cell, n = 5 squamous and n = 2 large cell neuroendocrine) were detected. On target acquired mechanisms were EGFR amplification (n = 7), and G724S (n = 2) and C797S (n = 3) mutations. Pts with an identified mechanism of resistance did not have improved post-progression survival (12 mo HR 1.6, p = 0.09), but receiving next line of therapy based on post-progression tumor biopsy results (including platinum-etoposide for transformation) did improve post-progression survival (12 mo HR 0.4, p = 0.01). Conclusions: Pts with atypical EGFR drivers or concurrent TP53+/- RB1 alterations have significantly shorter rwPFS on first line osimertinib, highlighting need for additional interventions for these patients. Given the high frequency of transformation and improvement in post-progression survival by tailoring next line of therapy to the identified mechanism, pts with EGFR-mutant lung cancer on first line osimertinib may benefit from tissue biopsies at progression. For pts without an identified resistance mechanism by NGS, additional methods of interrogating tumors at progression are needed.