Tamotsu Sagawa, Hiroyuki Nagashima, Koshi Fujikawa
Tissue immunohistochemistry remains the reference method for treatment selection; circulating tumor DNA cannot currently reproduce membranous protein intensity or spatial distribution, and liquid-biopsy approaches for CLDN18.2 reassessment remain investigational.
PURPOSE: Claudin 18 isoform 2 (CLDN18.2) has emerged as a clinically actionable target in gastric and gastroesophageal junction adenocarcinoma. Zolbetuximab plus chemotherapy established the first validated CLDN18.2-directed strategy, but also highlighted clinically important limitations: incomplete primary sensitivity, spatial and temporal heterogeneity of antigen expression, gastrointestinal toxicity, uncertain biomarker persistence after treatment, and the absence of an evidence-based sequence after progression. These limitations have accelerated development of antibody-drug conjugates (ADCs), chimeric antigen receptor (CAR) T-cell therapy, bispecific antibodies, T-cell engagers, and rational combinations with chemotherapy or immune checkpoint blockade.
METHODS: This review summarizes current evidence through July 2026, examines plausible mechanisms of resistance and biomarker evolution, and proposes a cautious, modality-specific framework for future clinical development.
RESULTS: Updated clinical data support proof of concept across multiple modalities. IBI343, an exatecan-based ADC, has shown antitumor activity with predominantly hematologic toxicity; vedotin-based ADCs demonstrate a different linker-payload profile with additional concern for microtubule-related cumulative toxicity. Randomized phase II data with satricabtagene autoleucel have established progression-free survival benefit over treatment of physician's choice in previously treated disease, although lymphodepletion-related cytopenias and cytokine-release syndrome require specialized infrastructure. Early studies of givastomig and IBI389 further support immune-engaging approaches.
CONCLUSION: Tissue immunohistochemistry remains the reference method for treatment selection; circulating tumor DNA cannot currently reproduce membranous protein intensity or spatial distribution, and liquid-biopsy approaches for CLDN18.2 reassessment remain investigational.