Eosinophilic Oesophagitis (EoE) is a chronic immune system disease characterised by inflammation of the oesophagus. It is primarily diagnosed and monitored by elevated levels of eosinophils in the oesophageal tissue.

However, EoE is also associated with significant yet less studied oesophageal mastocytosis. A recent study published in The Journal of Allergy and Clinical Immunology used machine learning to localise better and characterise oesophageal mast cells (MCs) to understand their role in the disease’s pathology.

The research team: Zhang S, Caldwell J, Rochman M, Collins M and Rothenberg M at Cincinnati Children's Hospital Rothenberg CURED Lab revealed new insights into the mechanisms driving EoE and potential new avenues for diagnosis and treatment.


Researchers stained oesophageal biopsy samples (EoE, control) for mast cells (MCs) using anti-tryptase and then imaged them via immunofluorescence. They then digitally assembled high-resolution whole-tissue images.

After that, they trained machine learning software to identify, enumerate, and characterise MCs, designated Mast Cell–Artificial Intelligence (MC-AI).

Photo: Mast Cells

Mast Cells: Laboratory of Allergic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health (LAD/NIAID/NIH)



MC-AI counted cells with high accuracy. During active EoE, the number of epithelial MCs increased while lamina propria (LP) MCs decreased. In both control subjects and EoE patients in remission, papillae had the highest MC density, negatively correlated with epithelial MC density. The MC density in the epithelium and papillae correlated with the severity of epithelial eosinophilic inflammation, basal zone hyperplasia, and LP fibrosis.

MC-AI detected increased MC degranulation in the epithelium, papillae, and LP in patients with EoE compared to control individuals. MCs in active EoE were located further from the basement membrane than in EoE remission and control individuals but were closer to the basement membrane than eosinophils in active EoE.

Infographic detailing the use of machine learning to identify and characterise mast cells in Eosinophilic Oesophagitis


Using MC-AI, researchers have identified a unique group of homeostatic oesophageal papillae MCs. During active EoE, this group decreases and undergoes degranulation, negatively correlates with epithelial mast cell levels, and significantly correlates with distinct histologic features. In general, MC-AI offers a way to comprehend the potential role of MCs in EoE and other disorders.

Read the full article

EndoSign EoE Study

EoE Genetics Research