このエントリーをはてなブックマークに追加
ID 60366
FullText URL
fulltext.pdf 1.94 MB
Author
Kitamura, Shinji Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences Kaken ID publons
Takahashi, Kensaku Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences
Sang, Yizhen Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences
Fukushima, Kazuhiko Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences
Tsuji, Kenji Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences ORCID Kaken ID researchmap
Wada, Jun Department of Nephrology, Rheumatology, Endocrinology and Metabolism, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences ORCID Kaken ID publons researchmap
Abstract
Artificial Intelligence (AI) imaging diagnosis is developing, making enormous steps forward in medical fields. Regarding diabetic nephropathy (DN), medical doctors diagnose them with clinical course, clinical laboratory data and renal pathology, mainly evaluate with light microscopy images rather than immunofluorescent images because there are no characteristic findings in immunofluorescent images for DN diagnosis. Here, we examined the possibility of whether AI could diagnose DN from immunofluorescent images. We collected renal immunofluorescent images from 885 renal biopsy patients in our hospital, and we created a dataset that contains six types of immunofluorescent images of IgG, IgA, IgM, C3, C1q and Fibrinogen for each patient. Using the dataset, 39 programs worked without errors (Area under the curve (AUC): 0.93). Five programs diagnosed DN completely with immunofluorescent images (AUC: 1.00). By analyzing with Local interpretable model-agnostic explanations (Lime), the AI focused on the peripheral lesion of DN glomeruli. On the other hand, the nephrologist diagnostic ratio (AUC: 0.75833) was slightly inferior to AI diagnosis. These findings suggest that DN could be diagnosed only by immunofluorescent images by deep learning. AI could diagnose DN and identify classified unknown parts with the immunofluorescent images that nephrologists usually do not use for DN diagnosis.
Keywords
immunofluorescent image
renal pathology
artificial intelligence
deep learning
diabetic nephropathy
Published Date
2020-07-09
Publication Title
Diagnostics
Volume
volume10
Issue
issue7
Publisher
MDPI
Start Page
466
ISSN
2075-4418
Content Type
Journal Article
language
英語
OAI-PMH Set
岡山大学
Copyright Holders
© 2020 by the authors.
File Version
publisher
PubMed ID
DOI
Web of Science KeyUT
Related Url
isVersionOf https://doi.org/10.3390/diagnostics10070466
License
http://creativecommons.org/licenses/by/4.0/