Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan
Abstract This cross-sectional study aimed to investigate the promoting and inhibitory factors of diabetic retinopathy (DR) according to diabetes mellitus (DM) stage using standardized evaluation of fundus images by artificial intelligence (AI). A total of 30,167 participants underwent blood and fund...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-11-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-47270-x |
| _version_ | 1797576776352792576 |
|---|---|
| author | Koji Komatsu Kei Sano Kota Fukai Ryo Nakagawa Takashi Nakagawa Masayuki Tatemichi Tadashi Nakano |
| author_facet | Koji Komatsu Kei Sano Kota Fukai Ryo Nakagawa Takashi Nakagawa Masayuki Tatemichi Tadashi Nakano |
| author_sort | Koji Komatsu |
| collection | DOAJ |
| description | Abstract This cross-sectional study aimed to investigate the promoting and inhibitory factors of diabetic retinopathy (DR) according to diabetes mellitus (DM) stage using standardized evaluation of fundus images by artificial intelligence (AI). A total of 30,167 participants underwent blood and fundus examinations at a health screening facility in Japan (2015–2016). Fundus photographs were screened by the AI software, RetCAD and DR scores (DRSs) were quantified. The presence of DR was determined by setting two cut-off values prioritizing sensitivity or specificity. DM was defined as four stages (no DM: DM0; advanced DM: DM3) based on treatment history and hemoglobin A1c (HbA1c) levels. Associated factors of DR were identified using logistic regression analysis. For cutoff values, multivariate analysis revealed age, sex, systolic blood pressure (SBP), smoking, urinary protein, and HbA1c level as positively associated with the risk of DR among all DM stages. In addition to glycemic control, SBP and Fibrosis-4 index might act as promoting factors for DR at all or an earlier DM stage. T-Bil, cholinesterase, and T-cho level might be protective factors at an advanced DM stage. |
| first_indexed | 2024-03-10T21:59:26Z |
| format | Article |
| id | doaj.art-7a8d9bb911a74f43929e4a985cf920d8 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-10T21:59:26Z |
| publishDate | 2023-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-7a8d9bb911a74f43929e4a985cf920d82023-11-19T13:01:06ZengNature PortfolioScientific Reports2045-23222023-11-011311810.1038/s41598-023-47270-xAssociated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in JapanKoji Komatsu0Kei Sano1Kota Fukai2Ryo Nakagawa3Takashi Nakagawa4Masayuki Tatemichi5Tadashi Nakano6Department of Ophthalmology, The Jikei University School of MedicineDepartment of Ophthalmology, The Jikei University School of MedicineDepartment of Preventive Medicine, School of Medicine, Tokai UniversityOmiya City ClinicOmiya City ClinicDepartment of Preventive Medicine, School of Medicine, Tokai UniversityDepartment of Ophthalmology, The Jikei University School of MedicineAbstract This cross-sectional study aimed to investigate the promoting and inhibitory factors of diabetic retinopathy (DR) according to diabetes mellitus (DM) stage using standardized evaluation of fundus images by artificial intelligence (AI). A total of 30,167 participants underwent blood and fundus examinations at a health screening facility in Japan (2015–2016). Fundus photographs were screened by the AI software, RetCAD and DR scores (DRSs) were quantified. The presence of DR was determined by setting two cut-off values prioritizing sensitivity or specificity. DM was defined as four stages (no DM: DM0; advanced DM: DM3) based on treatment history and hemoglobin A1c (HbA1c) levels. Associated factors of DR were identified using logistic regression analysis. For cutoff values, multivariate analysis revealed age, sex, systolic blood pressure (SBP), smoking, urinary protein, and HbA1c level as positively associated with the risk of DR among all DM stages. In addition to glycemic control, SBP and Fibrosis-4 index might act as promoting factors for DR at all or an earlier DM stage. T-Bil, cholinesterase, and T-cho level might be protective factors at an advanced DM stage.https://doi.org/10.1038/s41598-023-47270-x |
| spellingShingle | Koji Komatsu Kei Sano Kota Fukai Ryo Nakagawa Takashi Nakagawa Masayuki Tatemichi Tadashi Nakano Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan Scientific Reports |
| title | Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan |
| title_full | Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan |
| title_fullStr | Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan |
| title_full_unstemmed | Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan |
| title_short | Associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in Japan |
| title_sort | associated factors of diabetic retinopathy by artificial intelligence evaluation of fundus images in japan |
| url | https://doi.org/10.1038/s41598-023-47270-x |
| work_keys_str_mv | AT kojikomatsu associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT keisano associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT kotafukai associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT ryonakagawa associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT takashinakagawa associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT masayukitatemichi associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan AT tadashinakano associatedfactorsofdiabeticretinopathybyartificialintelligenceevaluationoffundusimagesinjapan |