Elizabeth Smyth, medical oncologist, Cambridge University Hospital Cambridge, United Kingdom
Biography
www.esmo.org/about-esmo/biographies/elizabeth-lizzy-smyth
Summary of presentation
Gastroesophageal adenocarcinoma (GEA) remains a significant clinical challenge due to its aggressive nature and heterogeneous biological behaviour. The identification and application of tissue biomarkers are critical for guiding targeted treatment strategies and improving patient outcomes. Key biomarkers, such as mismatch repair (MMR) proteins, HER2, PD-L1, and claudin 18.2, have established roles in personalising therapeutic approaches.
MMR deficiency (MMRd or MSI using genomic testing) is a well-recognised predictive biomarker for response to immune checkpoint inhibitors. MSI-high tumours exhibit higher mutation loads, enhancing the potential efficacy of immunotherapy. Testing for MMR status is now a routine component of GEA management, in early and advanced disease . HER2 overexpression, assessed through immunohistochemistry (IHC) and confirmed by in situ hybridisation (ISH), has been the cornerstone of targeted therapy, guiding the use of trastuzumab and trastuzumab deruxtecan in HER2-positive GEA. However, variability in assay protocols and scoring systems necessitates rigorous validation and standardisation to ensure reproducibility and reliability of results. PD-L1 expression, measured by IHC with various assays and scoring methodologies, is another significant biomarker influencing the use of checkpoint inhibitors. The combined positive score (CPS) is widely employed, yet the heterogeneous expression of PD-L1 poses challenges in consistent evaluation and interpretation. Claudin 18.2, a tight junction protein, has recently emerged as a promising target. Its expression in a subset of GEA patients has led to the development of targeted therapies such as zolbetuximab. Standardisation of claudin 18.2 testing and understanding its variable expression are ongoing priorities.
Validation of biomarker assays is essential for clinical implementation. This includes ensuring assay sensitivity, specificity, and inter-laboratory reproducibility. Addressing tumour heterogeneity, both spatial and temporal, is crucial to improve sampling accuracy and treatment decision-making. Sampling approaches that consider primary and metastatic sites can provide a more comprehensive biomarker profile. Emerging biomarkers, including FGFR2b and methylthioadenosine phosphorylase (MTAP) loss, offer new avenues for targeted therapy. FGFR2b overexpression is associated with aggressive disease and targeted by specific inhibitors such as bemarituzumab, while MTAP loss could inform metabolic vulnerabilities and synergistic treatment combinations.
The incorporation of digital pathology and artificial intelligence (AI) enhances the potential for more precise biomarker evaluation. AI-assisted image analysis can standardise scoring, reduce observer bias, and facilitate the identification of complex biomarker patterns.
Integration of biomarkers into perioperative therapy remains an essential aspect of GEA management, to maximise patient survival. The strategic use of validated biomarkers can refine patient selection for perioperative protocols, ensuring that targeted and immune-based therapies are effectively applied.
In conclusion, the expanding landscape of tissue biomarkers in GEA holds the promise of more individualised treatment pathways. Continued efforts in assay validation, understanding heterogeneity, and the integration of digital tools will be pivotal in advancing biomarker-driven management in this complex disease.