Automatic discovery of clinically interpretable imaging biomarkers for Mycobacterium tuberculosis supersusceptibility using deep learning.
Algorithms
Animals
Biomarkers
Computational Biology
/ methods
Deep Learning
Disease Models, Animal
Disease Susceptibility
Female
Humans
Image Processing, Computer-Assisted
Immunohistochemistry
/ methods
Machine Learning
Male
Molecular Imaging
/ methods
Mycobacterium tuberculosis
Prognosis
Reproducibility of Results
Tuberculosis
/ diagnosis
Biomarkers
Diversity outbred mice
Granuloma
Lung
Machine learning
Mice
Multiple instance learning
Tuberculosis
Journal
EBioMedicine
ISSN: 2352-3964
Titre abrégé: EBioMedicine
Pays: Netherlands
ID NLM: 101647039
Informations de publication
Date de publication:
Dec 2020
Dec 2020
Historique:
received:
20
05
2020
revised:
09
10
2020
accepted:
12
10
2020
pubmed:
10
11
2020
medline:
25
8
2021
entrez:
9
11
2020
Statut:
ppublish
Résumé
Identifying which individuals will develop tuberculosis (TB) remains an unresolved problem due to few animal models and computational approaches that effectively address its heterogeneity. To meet these shortcomings, we show that Diversity Outbred (DO) mice reflect human-like genetic diversity and develop human-like lung granulomas when infected with Mycobacterium tuberculosis (M.tb) . Following M.tb infection, a "supersusceptible" phenotype develops in approximately one-third of DO mice characterized by rapid morbidity and mortality within 8 weeks. These supersusceptible DO mice develop lung granulomas patterns akin to humans. This led us to utilize deep learning to identify supersusceptibility from hematoxylin & eosin (H&E) lung tissue sections utilizing only clinical outcomes (supersusceptible or not-supersusceptible) as labels. The proposed machine learning model diagnosed supersusceptibility with high accuracy (91.50 ± 4.68%) compared to two expert pathologists using H&E stained lung sections (94.95% and 94.58%). Two non-experts used the imaging biomarker to diagnose supersusceptibility with high accuracy (88.25% and 87.95%) and agreement (96.00%). A board-certified veterinary pathologist (GB) examined the imaging biomarker and determined the model was making diagnostic decisions using a form of granuloma necrosis (karyorrhectic and pyknotic nuclear debris). This was corroborated by one other board-certified veterinary pathologist. Finally, the imaging biomarker was quantified, providing a novel means to convert visual patterns within granulomas to data suitable for statistical analyses. Overall, our results have translatable implication to improve our understanding of TB and also to the broader field of computational pathology in which clinical outcomes alone can drive automatic identification of interpretable imaging biomarkers, knowledge discovery, and validation of existing clinical biomarkers. National Institutes of Health and American Lung Association.
Sections du résumé
BACKGROUND
BACKGROUND
Identifying which individuals will develop tuberculosis (TB) remains an unresolved problem due to few animal models and computational approaches that effectively address its heterogeneity. To meet these shortcomings, we show that Diversity Outbred (DO) mice reflect human-like genetic diversity and develop human-like lung granulomas when infected with Mycobacterium tuberculosis (M.tb) .
METHODS
METHODS
Following M.tb infection, a "supersusceptible" phenotype develops in approximately one-third of DO mice characterized by rapid morbidity and mortality within 8 weeks. These supersusceptible DO mice develop lung granulomas patterns akin to humans. This led us to utilize deep learning to identify supersusceptibility from hematoxylin & eosin (H&E) lung tissue sections utilizing only clinical outcomes (supersusceptible or not-supersusceptible) as labels.
FINDINGS
RESULTS
The proposed machine learning model diagnosed supersusceptibility with high accuracy (91.50 ± 4.68%) compared to two expert pathologists using H&E stained lung sections (94.95% and 94.58%). Two non-experts used the imaging biomarker to diagnose supersusceptibility with high accuracy (88.25% and 87.95%) and agreement (96.00%). A board-certified veterinary pathologist (GB) examined the imaging biomarker and determined the model was making diagnostic decisions using a form of granuloma necrosis (karyorrhectic and pyknotic nuclear debris). This was corroborated by one other board-certified veterinary pathologist. Finally, the imaging biomarker was quantified, providing a novel means to convert visual patterns within granulomas to data suitable for statistical analyses.
IMPLICATIONS
CONCLUSIONS
Overall, our results have translatable implication to improve our understanding of TB and also to the broader field of computational pathology in which clinical outcomes alone can drive automatic identification of interpretable imaging biomarkers, knowledge discovery, and validation of existing clinical biomarkers.
FUNDING
BACKGROUND
National Institutes of Health and American Lung Association.
Identifiants
pubmed: 33166789
pii: S2352-3964(20)30470-9
doi: 10.1016/j.ebiom.2020.103094
pmc: PMC7658666
pii:
doi:
Substances chimiques
Biomarkers
0
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
103094Subventions
Organisme : NCI NIH HHS
ID : P30 CA016058
Pays : United States
Organisme : NHLBI NIH HHS
ID : R01 HL145411
Pays : United States
Organisme : NIAID NIH HHS
ID : R21 AI115038
Pays : United States
Informations de copyright
Copyright © 2020 The Authors. Published by Elsevier B.V. All rights reserved.
Déclaration de conflit d'intérêts
Declaration of Competing Interest Mr. Tavolara has nothing to disclose. Dr. Niazi has nothing to disclose. Ms. Ginese has nothing to disclose. Dr. Piedra-Mora has nothing to disclose. Dr. Gatti has nothing to disclose. Dr. Beamer has nothing to disclose. Dr. Gurcan has nothing to disclose.