EACR25-0360

AREG or EREG Overexpression Identifies a Subset of Gastro-Esophageal Adenocarcinoma Patients Responsive to EGFR Targeting

D. Conticelli1, M. Volante2, F. Pietrantonio3, C. Orrù4, D. Kılıç4, A. Nottegar5, R. Petty6, S. Corso4, S. Giordano4, C. Migliore4
1Candiolo Cancer Institute - FPO, IRCCS, -, Candiolo, Italy
2University of Torino, Department of Oncology, Turin, Italy
3Fondazione IRCCS Istituto Nazionale dei Tumori, Medical Oncology Department, Milan, Italy
4Candiolo Cancer Institute - FPO, IRCCS, University of Torino, Department of Oncology, Candiolo, Italy
5Azienda Ospedaliera Universitaria Integrata, Department of Diagnostics and Public Health, Verona, Italy
6University of Dundee, Division of Molecular and Clinical Medicine, School of Medicine, Dundee, United Kingdom
Introduction:

The development of EGFR-targeting drugs for gastro-esophageal adenocarcinoma (GEA) treatment has been hampered by negative results from phase II/III clinical trials testing EGFR inhibitors in combination with chemotherapy. These failures could be explained by the lack of patient selection due to the absence of reliable predictive markers of response. However, preclinical and clinical findings indicate the effectiveness of EGFR blockade in GEA patients harboring EGFR gene amplification, promising a reassessment of EGFR as a therapeutic target in GEA within the context of a precise patient selection. Retrospective analyses also suggest that a subset of EGFR non-amplified GEA patients might benefit from EGFR targeting, highlighting the urgent need for predictive biomarkers of sensitivity.

Material and method:

We screened 27 GEA primary cancer cell lines and 10 Patient-Derived Xenografts for sensitivity to anti-EGFR drugs. The molecular characterization of these models was performed to uncover potential predictive biomarkers of response. Identified biomarkers were then validated by IHC and RNA in situ hybridization in samples from patients enrolled in the COG clinical trial.

Result and discussion:

Our analysis identified both in vitro and in vivo a subset of GEAs lacking EGFR gene amplification, but sensitive to EGFR inhibition. The molecular characterization of these models revealed overexpression of the EGFR ligands amphiregulin (AREG) or epiregulin (EREG) in the drug-sensitive group. To assess the translational significance of these findings, we conducted a post-hoc analysis on tumors from patients enrolled in the COG trial who were treated with the EGFR inhibitor gefitinib. Importantly, we detected a significant correlation between overall survival (OS) and AREG/EREG expression level (HR: 0,5827; CI:0,3445-0,9854, P=0,0439). Moreover, overexpression of EGFR ligands had no predictive power in the presence of KRAS mutations.

Conclusion:

We propose AREG or EREG overexpression as a new predictive marker to identify a subgroup of GEA patients who could benefit from EGFR inhibition.