Association of ERBB2 extrachromosomal DNA (ecDNA) with first-line HER2-targeted treatment outcomes in advanced esophagogastric adenocarcinoma (EGC).

4041 Background: ecDNA-enabled amplification of oncogenes is associated with increased expression and poor prognosis and is believed to contribute to treatment resistance. However, the prognostic implications of ecDNA amplification on specific targeted therapy outcomes remains unknown. We investigated whether ERBB2ecDNA positivity (+) in EGC is associated with survival on HER2-targeted agents. Methods: Demographics of patients with metastatic or recurrent EGC treated with first-line HER2-directed therapy and chemotherapy at Memorial Sloan Kettering between 2012 and 2021 and who had undergone MSK-IMPACT testing on their tumors were reviewed via the electronic medical record. ERBB2 amplification and ecDNA status were determined from samples collected within 90 days prior to treatment initiation via the ECHO pipeline, a proprietary algorithm provided by Boundless Bio that uses NGS sequence data to detect the presence of ecDNA-amplified oncogenes, by analyzing sequence data for features common to ecDNA but not chromosomal amplifications. Copy number fold-change (FC) was determined via the previously published MSK FACETS pipeline. Progression-free survival (PFS) and overall survival (OS) were calculated from treatment initiation to event or last follow-up and compared via Cox proportional hazards test. Two-sample comparison of non-parametric data was performed via Wilcoxon rank test. Results: 162 patients received first-line HER2 inhibition, all with concurrent platinum-based chemotherapy, and in 35% of patients also in combination with immune-checkpoint blockade. Median age at treatment was 59.9 years. Most patients were male (75%) and white (80%), and the majority had esophageal/GE junction tumors (70%). 64% of patients were HER2 IHC 3 prior to treatment, 19% were IHC 2, 12% had unknown IHC status, and the remaining 5% of patients were IHC 0 or 1 at MSK and treated on the basis of outside IHC positivity and/or genomic ERBB2 amplification. ERBB2 ecDNA+ (n = 77) was associated with higher ERBB2 copy number fold-change (median FC 13.5) and with prolonged PFS and OS compared to ERBB2 non-amplified/ecDNA- (median FC 1, p < 0.0001; PFS HR 2.60, 95% CI 1.70 - 3.98, p < 0.0001; OS HR 3.42, 95% CI 2.18 - 5.36, p < 0.0001) and ERBB2 amplified/ecDNA- patients (median FC 2.7, p < 0.0001; PFS HR 1.89, 95% CI 1.19 - 3.00, p = 0.007; OS HR 1.83, 95% CI 1.10 - 3.04, p = 0.02). ecDNA+ in EGFR, FGFR2, KRAS, and/or MET(n = 17) was associated with inferior PFS (HR 2.22, 95% CI 1.30 - 3.79, p = 0.004) and OS (HR 2.41, 95% CI 1.36 - 4.28, p = 0.003) compared to ecDNA- in all four oncogenes. Conclusions: Despite untargeted oncogenic ecDNA+ amplifications portending an inferior prognosis, we demonstrate that ERBB2 ecDNA+ may be prognostic for improved survival on HER2-targeted therapy in advanced EGC. ecDNA+ in untargeted oncogenes may represent a mechanism of treatment resistance.