Esophageal adenocarcinoma (EAC) is one of the deadliest cancers in the Western world, and its incidence has been rising for decades. Most patients are diagnosed only after the disease has silently advanced, contributing to a five-year survival rate of about 20%. A key opportunity to save lives is identifying which individuals with Barrett’s esophagus, a pre-cancerous condition, are truly on the path to cancer. However, only a small fraction of patients will develop malignancy, and current tools cannot reliably distinguish those at high risk. Existing methods rely on detecting dysplasia, a late and inconsistently assessed histological alteration, leaving clinicians without the early-warning system they need. This creates an urgent public health challenge: developing accurate, affordable, scalable biomarkers that can detect cancer risk long before EAC emerges.
This award will harness spatial transcriptomics, longitudinal genomic profiling, and generative AI–enabled digital pathology to define the earliest molecular and microenvironmental events that signal progression from Barrett’s esophagus to cancer. By mapping how genomic alterations, cell–cell interactions, and tissue architecture evolve over time, the team aims to uncover early signatures that precede conventional histologic changes. These multimodal data will inform a novel computational diagnostic that infers key genomic and spatial biomarkers directly from standard histology slides. Validated across deeply characterized international cohorts, this approach could deliver scalable, low-cost risk-stratification tools for routine pathology practice. Ultimately, these insights may transform early detection, enable personalized surveillance, and reveal new therapeutic targets to reduce the burden of this lethal cancer.