AUTHOR=Deng Xiaowen , Li Zijing , Zeng Peng , Wang Jing , Liang Jiaqi , Lan Yuqing 

TITLE=A Diagnostic Model for Screening Diabetic Retinopathy Using the Hand-Held Electroretinogram Device RETeval

JOURNAL=Frontiers in Endocrinology

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2021.632457

DOI=10.3389/fendo.2021.632457

ISSN=1664-2392

ABSTRACT=Purpose:To construct a proper diabetic retinopathy (DR) screening model using RETeval.
Methods:This was a cross-sectional study. Two hundred and thirty-two diabetic patients and seventy controls were recruited. The DR risk assessment protocol was performed to obtain subjects' DR risk score using RETeval. Afterwards, the receiver operating characteristic(ROC) curve was used to determine the best cut-off of DR while the random forest and decision tree models were constructed.
Results:With the increment of DR severity, DR score gradually increased. When DR score was used to diagnose DR, the ROC curve had an area under the curve of 0.881 (95% confidence interval: 0.836-0.927, P < 0.001) with a best cut-off value of 22.95, a sensitivity of 74.3%, a specificity of 90.6% and a Youden index of 0.648. The top four risk factors selected by the random forest were used to construct the decision tree to diagnose DR with a sensitivity of 87.01% in train set and 100% in test set, a specificity of 86.75% in train set and 86.05% in test set, a classification error rate of 13.125% in train set and 8.33% in test set.
Conclusions: The DR risk assessment protocol combined with the decision tree model was innovatively used to evaluate the risk of DR, improving the sensitivity, which would be more suitable for DR screening.