Clinicopathologic correlates of claudin 18.2 expression in esophagogastric cancer at multiple expression levels

Background: Claudin-18 isoform 2 (CLDN18.2) is a validated therapeutic target in gastric cancer (GC) and gastroesophageal junction cancer (GEJ). The approved antibody zolbetuximab requires higher expression levels (≥75% of tumor cells with 2-3+ membranous staining), yet broader thresholds are increasingly explored. Data on clinicopathologic correlates and biomarker overlap at lower cut-offs remain limited. Patients and methods: We retrospectively analyzed patients with esophageal cancer (EC), GEJ, or GC who underwent CLDN18.2 immunohistochemistry at a single high-volume center (2023-2025). CLDN18.2 positivity was defined using three thresholds: (i) stringent (75% threshold: ≥75% of tumor cells with 2-3+ staining); (ii) broader (H-score 25; ≥1+ intensity in ≥25% of cells); and (iii) the lowest (H-score 10; ≥1+ intensity in ≥10% of cells). Associations between CLDN18.2 and clinicopathologic variables, including histology and standard biomarkers [human epidermal growth factor receptor 2 (HER2), programmed death-ligand 1 (PD-L1), and mismatch repair (MMR)], were assessed. Results: In total, 149 patients were included: 43 EC (28.9%), 46 GEJ (30.9%), and 60 GC (40.3%). CLDN18.2 positivity rates were 48.3%, 71.1%, and 73.8% at the 75%, H-score 25, and H-score 10 thresholds, respectively. CLDN18.2 positivity significantly correlated with tumor location, highest in GC and followed by GEJ and EC across thresholds. At the lowest cut-off, positivity was 85.0% in GC, 69.6% in GEJ, and 62.8% in EC (P = 0.030). Histological subtypes were also associated across thresholds: poorly cohesive carcinoma (PCC) demonstrated significantly higher positivity than non-PCC. No significant associations were found with HER2, PD-L1 combined positive score, or MMR. Triple positivity for CLDN18.2, HER2, and PD-L1 was rare (≤5.5%). Conclusion: CLDN18.2 expression in upper gastrointestinal cancers was significantly associated with gastric origin and PCC histology, independent of HER2 and PD-L1 status. Broadening the threshold for positivity identified nearly three-quarters of patients as CLDN18.2 positive, underscoring the impact of selection cut points on trial eligibility. These findings highlight the potential to expand patient access to CLDN18.2-targeted therapies using lower thresholds.