Histopathology-Based Spatial Profiling of Immune and Molecular Features Predicts Cancer Risk in Barrett's Esophagus Patients.

BEACON predicts cancer progression risk in Barrett’s esophagus by integrating histopathology-based DNA content abnormality and spatial immune ecology, stratifying DNA-abnormal patients into high- and low-risk groups linked to progression and survival.

• Why it matters: Barrett’s esophagus needs better cancer risk stratification to target high-risk patients and avoid over-treatment of low-risk individuals. The work addresses the gap in combining genomic and microenvironmental signals for prediction.
• What they did: BEACON uses DACOR, a multi-instance learning model trained on 777 BE biopsies with flow cytometry-based DNA content data from two institutions, to predict DNA content abnormality from routine histopathology. It also computes spatial immune ecology metrics from cell classification and tissue segmentation, then fits a logistic regression integrating molecular, immune ecological, and epithelial morphology features.
• The result: DNA content abnormality predicted by DACOR is associated with increased cancer progression, and abnormal regions show increased lymphoplasmacytic inflammation versus normal regions. Among DNA-abnormal patients, cancer progressors have increased plasma cell clustering adjacent to abnormal epithelium, and the integrated model stratifies patients into risk groups corresponding to progression and patient survival.