Prediction of hypertension, hyperglycemia and dyslipidemia from retinal fundus photographs via deep learning: A cross-sectional study of chronic diseases in central China

• Published in: PloS one Vol. 15; no. 5; p. e0233166
• Main Authors: Zhang, Li, Yuan, Mengya, An, Zhen, Zhao, Xiangmei, Wu, Hui, Li, Haibin, Wang, Ya, Sun, Beibei, Li, Huijun, Ding, Shibin, Zeng, Xiang, Chao, Ling, Li, Pan, Wu, Weidong
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 14.05.2020; Public Library of Science (PLoS)
• Subjects: Algorithms; Alzheimer's disease; Anthropometry; Artificial intelligence; Artificial neural networks; Biochemistry; Blood pressure; Blood tests; Body mass index; Cardiovascular diseases; Chronic diseases; Chronic illnesses; Classification; Complications; Complications and side effects; Computer aided medical diagnosis; Computer and Information Sciences; Cross-sectional studies; Datasets; Deep learning; Diabetes mellitus; Diabetic retinopathy; Diseases; Dyslipidemia; Dyslipidemias; Erythrocytes; Glycosylated hemoglobin; Health risks; Heart diseases; Hematocrit; Hemoglobin; Hemoglobins; Hyperglycemia; Hypertension; Image quality; Machine learning; Mathematical models; Medical examination; Medical imaging; Medical imaging equipment; Medical research; Medicine and Health Sciences; Metabolic disorders; Methods; Neural networks; Parameters; People and Places; Photography; Population; Public health; Retina; Retinal images; Risk analysis; Risk factors; Rural areas; China
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0233166

Retinal fundus photography provides a non-invasive approach for identifying early microcirculatory alterations of chronic diseases prior to the onset of overt clinical complications. Here, we developed neural network models to predict hypertension, hyperglycemia, dyslipidemia, and a range of risk factors from retinal fundus images obtained from a cross-sectional study of chronic diseases in rural areas of Xinxiang County, Henan, in central China. 1222 high-quality retinal images and over 50 measurements of anthropometry and biochemical parameters were generated from 625 subjects. The models in this study achieved an area under the ROC curve (AUC) of 0.880 in predicting hyperglycemia, of 0.766 in predicting hypertension, and of 0.703 in predicting dyslipidemia. In addition, these models can predict with AUC>0.7 several blood test erythrocyte parameters, including hematocrit (HCT), mean corpuscular hemoglobin concentration (MCHC), and a cluster of cardiovascular disease (CVD) risk factors. Taken together, deep learning approaches are feasible for predicting hypertension, dyslipidemia, diabetes, and risks of other chronic diseases.