Actionable biomarker co-expression in G/GEJ cancer

In this video, Dr Filip Van Herpe speaks with Dr Rutika Mehta to discuss a real-world tumour profiling study presented at ESMO GI 2026, evaluating actionable biomarker co-expression in patients with gastric, gastro-oesophageal junction and oesophageal adenocarcinoma.¹ The retrospective analysis used a US molecular-linked claims database and included more than 2,500 patients who had undergone CLDN18.2 testing between 2021 and 2025. The aim was to better understand overlap between CLDN18.2, HER2, PD-L1 and MMR/MSI status in a treatment landscape where multiple biomarker-directed therapies are becoming available. 

Overall, 2,697 patients with evaluable CLDN18.2 results were included. Using the zolbetuximab-related definition of CLDN18.2 positivity, defined as ≥75% of tumour cells with ≥2+ staining intensity, 41% of tumours were CLDN18.2-positive. HER2 positivity was observed in 15% of patients, PD-L1 CPS ≥10 in 15%, and dMMR/MSI-H status in 7%. Biomarker overlap was clinically relevant: approximately 4.5% of tumours were both CLDN18.2-positive and HER2-positive, and 5.0% were both CLDN18.2-positive and PD-L1 CPS ≥10. 

Dr Mehta highlights that this overlap creates an important knowledge gap in daily practice. For example, when a tumour is both CLDN18.2-positive and HER2-positive, current evidence generally supports prioritising HER2-directed therapy, as zolbetuximab was not studied in HER2-positive patients. However, the optimal sequencing of therapies, or the potential value of future combination approaches, remains unknown.

The study has limitations, including its retrospective and descriptive design, the use of US physician-ordered molecular profiling, and the lack of prognostic or outcomes data. Nevertheless, the large cohort provides useful insight into the frequency of biomarker co-expression and supports comprehensive biomarker testing at diagnosis to guide treatment prioritisation and future trial design.