P76.41 Detection of Diverse EGFR c-terminal Truncations (C-trunc) and Sensitivity to Tyrosine Kinase Inhibitors (TKIs) in the Clinic

Abstract

EGFR kinase domain (KD) mutations are classic targetable drivers in non-small cell lung cancer (NSCLC) linked to approved TKIs. EGFR C-truncs are also found in human tumors and known to be activating in vitro. We aimed to describe the therapeutic implications of these C-trunc variants and hypothesized that they are targetable drivers in the clinic. The Foundation Medicine (FM) genomic database (FMGD) of patients having undergone hybrid capture-based comprehensive genomic profiling (CGP) of tumor tissue or circulating tumor DNA (ctDNA) was interrogated to assess the prevalence of EGFR C-truncs affecting exons 25-28 and co-occurring known oncogenic drivers. All exons of EGFR were baited for tissue and ctDNA whereas introns 24-27 were baited for in tissue only. A second database, the Flatiron Health (FH)-FM NSCLC Clinico-Genomic Database (CGDB), was explored to identify patients with advanced NSCLC (aNSCLC) whose tumors harbor EGFR C-truncs. CGDB is a nationwide (US-based) de-identified EHR-derived database, and includes patients in FMGD who also received care within the FH network. Clinical characteristics, treatment selections and response assessments were described for patients with aNSCLC diagnosed between 1/2011-9/2019. In the FMGD, EGFR C-truncs were detected in 0.27% (165/60,261) of aNSCLC cases, resulting from either rearrangements (70%) or mutations (nonsense, 14%; splice site, 8.9%; frameshift, 7.1%) affecting exons 25-28. EGFR C-truncs were detected in 0.28% and 0.21% of NSCLC tissue and ctDNA samples, respectively, with rearrangements with intronic breakpoints detected in 0.15% of tissue and 0.07% of ctDNA cases. Co-occurring EGFR KD mutations or other known drivers (primarily KRAS mutations and MET amplification) were detected with EGFR C-truncs in 29% and 11% of tissue and 38% and 19% of ctDNA cases, respectively. In the CGDB, EGFR C-truncs were present in 0.36% (25/6,687) of samples from aNSCLC patients. Of these, 14 had documented receipt of EGFR TKI(s) and 12/14 (5 mutations, 7 rearrangements) had evaluable outcomes, including 6 with co-occurring EGFR driver mutations (4 exon 19 deletion, 2 L858R), 1 with BRAF V600E, and 1 with KRAS G12C. Of those with co-EGFR drivers, all had partial (PR) or complete responses to erlotinib (n=1, 12.2 months), afatinib (n=1, 18.2 months) or erlotinib followed by osimertinib (n=4, 8.0-42.4 months total). Of 4 evaluable cases with no co-occurring known drivers, 3 had EGFR TKI post CGP including 2 with PRs (EGFR Q1174*, osimertinib 5 months, ongoing; and EGFR exon 25 rearrangement, erlotinib followed by osimertinib, total 28 months) and 1 with progressive disease (EGFR exon 27 rearrangement, erlotinib 3 months). The fourth evaluable patient with EGFR C-trunc alone received erlotinib followed by osimertinib for 25 months total with a PR, then CGP 3 months prior to osimertinib discontinuation revealed an EGFR exon 25 splice site mutation. EGFR C-truncs are detectable in NSCLC tissue and plasma using broad hybrid-capture based genomic profiling and may occur alone or with other EGFR activating mutations. Consistent with our hypothesis, preliminary real-world clinical data suggest that patients with these alterations may derive benefit from approved EGFR TKIs and additional studies are warranted.