Modelling oesophageal adenocarcinoma for precision oncology and immunotherapy

Abstract Oesophageal adenocarcinoma (OAC) remains a lethal disease with poor long-term survival and limited predictive biomarkers for therapy selection. Profound inter- and intra-tumour heterogeneity—often driven by chromosomal instability and large-scale genomic alteration—contributes to variable responses to chemotherapy, radiotherapy, targeted agents, and immune checkpoint inhibitors. Precision oncology aims to match treatment to the biology of an individual tumour, but in OAC this goal is constrained by the historical shortage of preclinical systems that reliably recapitulate patient-specific genotype, phenotype, and tumour microenvironment. Here, we critically review the spectrum of experimental and computational models available for OAC precision oncology, spanning conventional cell lines, co-culture systems, patient-derived organoids, and patient-derived xenografts including orthotopic and humanised variants, and ex vivo organotypic tissue slice platforms. We focus on how each model class performs against translationally relevant criteria—fidelity to the parent tumour, representation of stromal and immune compartments, scalability, time-to-result, and suitability for clinically aligned endpoints such as treatment response prediction and resistance evolution. We synthesise emerging evidence that immune-augmented organoid systems and organotypic cultures can support functional immunology readouts and short-horizon therapeutic testing, while PDX-based approaches provide organism-level pharmacology and evolutionary context at the expense of throughput and turnaround time. Finally, we discuss the accelerating role of in silico inference pipelines and computational histopathology in integrating multi-omic and imaging data to enable scalable prediction, while emphasising the need for rigorous benchmarking against patient outcomes and experimental models.