Christina Curtis, PhD and Greg Charville, MD, PhD, Stanford University; William Grady, MD, University of Washington; Rebecca Fitzgerald, MD, University of Cambridge

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.

Peter K. Sorger, PhD, Sandro Santagata, MD, PhD, and David R. Walk, PhD, Harvard University; Ronny I. Drapkin, MD, PhD, University of Pennsylvania

High-grade serous ovarian cancer (HGSOC) is the most lethal gynecologic cancer and is now understood to arise primarily from serous tubal intraepithelial carcinomas (STICs) in the distal fallopian tubes, long before disease becomes clinically detectable. Although prophylactic removal of the fallopian tubes can prevent progression, this is not feasible for broad populations, creating an urgent need for molecular tools that can identify early STIC lesions and distinguish those with metastatic potential. Current diagnostic and risk-stratification methods lack the sensitivity and resolution needed to address this problem, leaving a significant gap in early intervention strategies.

This Early Detection Award will define the earliest steps of HGSOC development through single-cell and spatial characterization of incidental and cancer-associated STICs collected from a unique multi-institutional biobank. Using advanced tissue imaging and spatial transcriptomics, the investigators will map tumor intrinsic features and microenvironmental changes that accompany the transition from STIC to invasive cancer. These data will guide the discovery of secreted proteins enriched in STICs, with the most promising evaluated for use in highly sensitive, multiplexed plasma assays. Spatial and molecular features will also inform a tissue-based risk stratification framework to identify lesions likely to have already spread beyond the primary tumor site. This project is poised to deliver the first comprehensive molecular atlas of STIC biology, laying essential groundwork for a new clinical continuum that integrates early detection, personalized risk assessment, and informed decision-making around prophylactic surgery and surveillance. The long-term impact could be transformative for patients at risk of HGSOC, shifting care toward prevention and earlier intervention.