Assessment of Machine Learning Detection of Environmental Enteropathy and Celiac Disease in Children

Abstract

Key Points Question Can deep learning image analysis distinguish pathological vs healthy features in duodenal tissue? Findings In this diagnostic study, a deep learning convolutional neural network was trained on 3118 images from duodenal biopsies of patients with environmental enteropathy, celiac disease, and no disease. The convolutional neural network achieved 93.4% case-detection accuracy, with a false-negative rate of 2.4%, and automatically learned microlevel features in duodenal tissue, such as alterations in secretory cell populations. Meaning A machine learning–based feature detection platform distinguished pathological from healthy tissue in gastrointestinal biopsy images.

A novel histological index for evaluation of environmental enteric dysfunction identifies geographic-specific features of enteropathy among children with suboptimal growth

“…Our data demonstrate similarities and differences between the cohorts that warrant comment. While EED and GSE cohorts share histopathologic similarities, as suggested by prior studies [26,40,[57][58][59][60], our index showed differences in the main histologic features that are classically associated with GSE, including intraepithelial lymphocytes. We also had the Environmental enteric dysfunction histology index opportunity to compare a GSE case from Pakistan (excluded from this analysis) whose total histologic score was somewhat less severe than the median scores of the St. Louis GSE cohort, although overall showed a similar pattern of tissue injury (detailed comparison of scores to St. Louis GSE and Pakistani cohorts in S4 Table).…”

“…We also had the Environmental enteric dysfunction histology index opportunity to compare a GSE case from Pakistan (excluded from this analysis) whose total histologic score was somewhat less severe than the median scores of the St. Louis GSE cohort, although overall showed a similar pattern of tissue injury (detailed comparison of scores to St. Louis GSE and Pakistani cohorts in S4 Table). A single case does not permit generalizations about the overlap between EED and GSE pathology, but raises the possibility that GSE disease may be distinguishable from EED by histopathology, as has been suggested in previous literature as well [26,40,57,59,60]. This may be most important in settings where both enteropathies co-exist.…”

Machine Learning Supports Automated Digital Image Scoring of Stool Consistency in Diapers

“…The use of ML for image recognition has been explored in a number of specialized areas in medicine including gastroenterology ( 21 ), for example, for distinguishing pathological versus healthy features in duodenal tissue ( 22 ). Other applications range from classification of skin lesions ( 23 ) to glaucoma diagnosis based on fundus images ( 24 ).…”

Error in Funding/Support

“…In the Original Investigation titled “Assessment of Machine Learning Detection of Environmental Enteropathy and Celiac Disease in Children,” 1 published June 14, 2019, there was an error in the Funding/Support section of the Article Information. It should have read, “This work was supported by the University of Virginia Center for Engineering in Medicine Grant awarded to Drs Syed and Brown, grant OPP1066203 from the Bill and Melinda Gates Foundation awarded to Dr Ali, grant OPP1066118 from the Bill and Melinda Gates Foundation awarded to Dr Kelly, and the University of Virginia THRIV Scholar Career Development Award awarded to Dr Syed.” This article has been corrected.…”

“…It should have read, “This work was supported by the University of Virginia Center for Engineering in Medicine Grant awarded to Drs Syed and Brown, grant OPP1066203 from the Bill and Melinda Gates Foundation awarded to Dr Ali, grant OPP1066118 from the Bill and Melinda Gates Foundation awarded to Dr Kelly, and the University of Virginia THRIV Scholar Career Development Award awarded to Dr Syed.” This article has been corrected. 1…”