Complete evaluation of resistance mechanisms to first-line osimertinib requires tissue biopsy.

e21154 Background: Osimertinib (osi) is the preferred 1st line (1L) therapy for patients (pts) with EGFR-mutant NSCLC, but acquired resistance is universal. Identification of mechanisms of resistance (MoR) to 1L osi can inform subsequent treatment strategies. Here, we report MoRs identified at our institution from tissue and/or liquid biopsy (bx) upon 1L osi progression. Methods: We included pts seen at MGH with advanced EGFR+ NSCLC who had tissue and/or liquid bx (for ctDNA analysis) obtained at progression on 1L osi. Molecular testing was performed using local next-generation sequencing (NGS) and, when available, MET/ EGFR FISH. We retrospectively collected clinical characteristics, tissue histology and NGS results. Results: 54 pts (38 Female; median age 61 yrs) had tissue (n = 35) and/or liquid (n = 32) bx upon progression on 1L osi between November 2016-January 2022. 13/54 pts had both tissue and ctDNA testing. Primary EGFR mutations included del19 (32), L858R (16) and other (6). Median time from osi start to 1st resistance bx was 12.3 mos (range, 1-37). MoRs are summarized in the table. Among 35 pts with tissue bx, histologic transformations (trans) were the most common MoR, seen in 23% pts, including both small cell histology (tSCLC; 5/35), squamous histology (tSqCC; 2/35) and 1 pt with both SCLC + SqCC features in one bx. Among 5 pts with tSCLC and pre-trans NGS, only 2 had baseline TP53/RB1 loss. MET amp was seen in 20% tissue (including one with tSqCC) and secondary EGFR mutations in 9%; 43% of tissue bx had no identified MoR. Among 32 pts with ctDNA NGS, secondary EGFR mutations (16%) were the most prevalent MoR, with MET amp seen in 13%. 65% pts had no MoR by ctDNA. 6/13 pts with both tissue and ctDNA had concordant findings (MET amp-2, EGFR L718Q – 1; No MoR - 3). As expected, the 2 pts with tSCLC had no MoR in ctDNA. In 2 cases, MET amp was seen in tissue but not ctDNA; the converse was true in 1 pt. Conclusions: Histologic transformations were the most common MoR to first-line osimertinib accounting for 23% tissue biopsies and, in 3/5 pts, occurred in the absence of the known risk factor of baseline TP53/RB1 loss. In this small cohort, MET amp was more commonly seen in tissue than in ctDNA, and 63% of liquid biopsies showed no detectable MoR. Our findings suggest that tissue biopsy is critical for complete evaluation of MoRs in pts progressing on first line osimertinib, and that tissue and ctDNA may be complementary.[Table: see text]