Related Experiment Video
Updated: Nov 4, 2025
06:55
Diagnosis of Neoplasia in Barrett’s Esophagus using Vital-dye Enhanced Fluorescence Imaging
Published on: May 11, 2014
12.2K
Advancing Eosinophilic Esophagitis Diagnosis and Phenotype Assessment with Deep Learning Computer Vision
William Adorno1, Alexis Catalano2,3, Lubaina Ehsan3
1Dept. of Engineering Systems and Environment, University of Virginia, Charlottesville, VA, U.S.A.
Biomedical Engineering Systems and Technologies, International Joint Conference, BIOSTEC ... Revised Selected Papers. BIOSTEC (Conference)
|May 28, 2021
Summary
This study introduces an automated deep learning method to quantify eosinophils for diagnosing Eosinophilic Esophagitis (EoE). This approach aids in assessing disease severity and progression, potentially guiding treatment plans.
Area of Science:
- Gastroenterology
- Computational Pathology
- Artificial Intelligence
Background:
- Eosinophilic Esophagitis (EoE) is a growing inflammatory esophageal condition.
- Current diagnosis relies on manual pathological review of eosinophil counts in biopsies, which is time-consuming and subjective.
- Assessing EoE severity and progression presents diagnostic challenges.
Purpose of the Study:
- To develop and validate an automated deep learning approach for quantifying eosinophils in esophageal biopsies for EoE diagnosis.
- To correlate eosinophil statistics with clinical and treatment phenotypes to guide patient management.
- To explore deep learning for identifying novel diagnostic features beyond eosinophil counts.
Main Methods:
- Utilized a U-Net deep image segmentation model for automated eosinophil quantification.
- Developed a post-processing system to generate EoE diagnostic and severity statistics.
- Applied a deep image classification model to identify additional diagnostic features.
- Compared automated statistics with patient metadata including clinical and treatment phenotypes.
Main Results:
- The deep learning model successfully quantified eosinophils, enabling automated EoE diagnosis.
- Generated statistics provided insights into disease severity and progression.
- Explored potential linkages between eosinophil counts, clinical phenotypes, and treatment responses.
- Identified novel image-based features for EoE diagnosis.
Conclusions:
- Deep learning offers a robust, automated solution for EoE diagnosis and disease monitoring.
- Automated eosinophil quantification can streamline pathological review and improve diagnostic accuracy.
- This approach has the potential to personalize treatment strategies for EoE patients.
