2021/01/13 by Tomer Czyzewski, Nati Daniel, Czyzewski, Tomer +13
Immunology and Microbiology · Medicine · #Computer Vision and Pattern Recognition (cs.CV) #Eosinophilic Esophagitis #FOS: Biological sciences #FOS: Computer and information sciences #FOS: Electrical engineering #IL-33, ST2, and ILC Pathways #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Quantitative Methods (q-bio.QM) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2101.04989
openalex publication_date 2021/01/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Goal: Eosinophilic esophagitis (EoE) is an allergic inflammatory condition\ncharacterized by eosinophil accumulation in the esophageal mucosa. EoE\ndiagnosis includes a manual assessment of eosinophil levels in mucosal biopsies\n- a time-consuming, laborious task that is difficult to standardize. One of the\nmain challenges in automating this process, like many other biopsy-based\ndiagnostics, is detecting features that are small relative to the size of the\nbiopsy. Results: In this work, we utilized hematoxylin- and eosin-stained\nslides from esophageal biopsies from patients with active EoE and control\nsubjects to develop a platform based on a deep convolutional neural network\n(DCNN) that can classify esophageal biopsies with an accuracy of 85%,\nsensitivity of 82.5%, and specificity of 87%. Moreover, by combining several\ndownscaling and cropping strategies, we show that some of the features\ncontributing to the correct classification are global rather than specific,\nlocal features. Conclusions: We report the ability of artificial intelligence\nto identify EoE using computer vision analysis of esophageal biopsy slides.\nFurther, the DCNN features associated with EoE are based on not only local\neosinophils but also global histologic changes. Our approach can be used for\nother conditions that rely on biopsy-based histologic diagnostics.\n