1.
Automated Grading of Diabetic Retinopathy with Ultra-Widefield Fluorescein Angiography and Deep Learning.
Wang, X, Ji, Z, Ma, X, Zhang, Z, Yi, Z, Zheng, H, Fan, W, Chen, C
Journal of diabetes research. 2021;:2611250
Abstract
PURPOSE The objective of this study was to establish diagnostic technology to automatically grade the severity of diabetic retinopathy (DR) according to the ischemic index and leakage index with ultra-widefield fluorescein angiography (UWFA) and the Early Treatment Diabetic Retinopathy Study (ETDRS) 7-standard field (7-SF). METHODS This is a cross-sectional study. UWFA samples from 280 diabetic patients and 119 normal patients were used to train and test an artificial intelligence model to differentiate PDR and NPDR based on the ischemic index and leakage index with UWFA. A panel of retinal specialists determined the ground truth for our data set before experimentation. A confusion matrix as a metric was used to measure the precision of our algorithm, and a simple linear regression function was implemented to explore the discrimination of indexes on the DR grades. In addition, the model was tested with simulated 7-SF. RESULTS The model classification of DR in the original UWFA images achieved 88.50% accuracy and 73.68% accuracy in the simulated 7-SF images. A simple linear regression function demonstrated that there is a significant relationship between the ischemic index and leakage index and the severity of DR. These two thresholds were set to classify the grade of DR, which achieved 76.8% accuracy. CONCLUSIONS The optimization of the cycle generative adversarial network (CycleGAN) and convolutional neural network (CNN) model classifier achieved DR grading based on the ischemic index and leakage index with UWFA and simulated 7-SF and provided accurate inference results. The classification accuracy with UWFA is slightly higher than that of simulated 7-SF.
2.
Association between Normal Thyroid Hormones and Diabetic Retinopathy in Patients with Type 2 Diabetes.
Zou, J, Li, Z, Tian, F, Zhang, Y, Xu, C, Zhai, J, Shi, M, Wu, G, Zhang, Z, Yang, C, et al
BioMed research international. 2020;:8161797
Abstract
The relationship between normal thyroid function and type 2 diabetes mellitus (T2DM) has been a particular focus for concern. The present study determined the relationship between thyroid hormone levels and the prevalence of diabetic retinopathy (DR) in T2DM patients. A cross-sectional study (n = 633) was performed in Xi'an, Shaanxi Province, China. Subjects were evaluated for anthropometric measurements, thyroid function, and diabetic retinopathy. Logistic regression models were used to assess the relationships between thyroid hormones and DR. Of 633 patients, 243 (38.4%) patients suffered from DR. The prevalence of DR showed a significantly decreasing trend across the quartiles based on free triiodothyronine (FT3) (FT3 quartile 1 group [FT3-Q1] <4.35 pmol/L, FT3 quartile 2 group [FT3-Q2] 4.35-4.70 pmol/L, FT3 quartile 3 group [FT3-Q3] 4.70-5.08 pmol/L, and FT3 quartile 4 group [FT3-Q4] ≥5.08 pmol/L) (56.7%, 42.5%, 33.1%, 23.8%, P < 0.001). In comparison with all participants categorized in FT3-Q1, the multivariable adjusted odds ratios (95% confidence interval) of DR in FT3-Q2, FT3-Q3, and FT3-Q4 were 0.587 (0.340-1.012), 0.458 (0.258-0.813), and 0.368 (0.201-0.673), (P = 0.055, P = 0.008, P = 0.001), respectively. FT3 levels within the normal range are negatively associated with DR in euthyroid patients with type 2 diabetes. Further studies should be aimed at clarifying the relationship between thyroid hormones and T2DM.
3.
Diagnostic Performance of Retinopathy in the Detection of Diabetic Nephropathy in Type 2 Diabetes: A Systematic Review and Meta-Analysis of 45 Studies.
Jiang, S, Yu, T, Zhang, Z, Wang, Y, Fang, J, Yang, Y, Liu, L, Li, W
Ophthalmic research. 2019;(2):68-79
Abstract
AIMS: To conduct an evidence-based evaluation of diabetic retinopathy (DR) for the diagnosis of diabetic nephropathy (DN) in type 2 diabetics with kidney disease. METHODS We systematically searched PubMed, EMBASE, and the Cochrane Library from inception to June 27, 2018, including the reference lists of identified primary studies. A study was included if it (1) used DR as a diagnostic test for DN; and (2) used histological evaluation of renal tissues as the reference standard. RESULTS The analysis included 45 studies (4,561 patients). A bivariate analysis yielded a sensitivity of 0.67 (95% CI 0.61-0.74) and a specificity of 0.78 (95% CI 0.73-0.82). The summary receiver operating characteristic curve analysis provided an area under the curve (AUC) of 0.79 (95% CI 0.76-0.83). In a setting of 41% prevalence of DN, the probability of DN would be 68% if the test of DR was positive, and the probability of DN would be 23% if it was negative. In addition, although the mean specificity of proliferative DR for the detection of DN was 0.99 (95% CI 0.45-1.00), the mean sensitivity was 0.34 (95% CI 0.24-0.44), and the AUC was 0.58 (95% CI 0.53-0.62). CONCLUSIONS DR is helpful in diagnosing DN in persons with type 2 diabetes and kidney disease, but the severity of DR may not parallel the presence of DN.