Nomogram-Based Prediction of the Risk of Diabetic Retinopathy: A Retrospective Study

Abstract

Objectives: This study is aimed at developing a risk nomogram of diabetic retinopathy (DR) in a Chinese population with type 2 diabetes mellitus (T2DM).

Methods: A questionnaire survey, biochemical indicator examination, and physical examination were performed on 4170 T2DM patients, and the collected data were used to evaluate the DR risk in T2DM patients. By operating R software, firstly, the least absolute shrinkage and selection operator (LASSO) regression analysis was used to optimize variable selection by running cyclic coordinate descent with 10 times K cross-validation. Secondly, multivariable logistic regression analysis was applied to build a predicting model introducing the predictors selected from the LASSO regression analysis. The nomogram was developed based on the selected variables visually. Thirdly, calibration plot, receiver operating characteristic (ROC) curve, and decision curve analysis were used to validate the model, and further assessment was running by external validation.

Results: Seven predictors were selected by LASSO from 19 variables, including age, course of disease, postprandial blood glucose (PBG), glycosylated haemoglobin A1c (HbA1c), uric creatinine (UCR), urinary microalbumin (UMA), and systolic blood pressure (SBP). The model built by these 7 predictors displayed medium prediction ability with the area under the ROC curve of 0.700 in the training set and 0.715 in the validation set. The decision curve analysis curve showed that the nomogram could be applied clinically if the risk threshold is between 21% and 57% and 21%-51% in external validation.

Conclusion: Introducing age, course of disease, PBG, HbA1c, UCR, UMA, and SBP, the risk nomogram is useful for prediction of DR risk in T2DM individuals.

Figure 1

The normal fundus characteristics of T2DM patients (a) and the abnormal fundus characteristics of DR patients (b).

Figure 2

Variable selection by LASSO binary logistic regression model. A coefficient profile plot was produced against the log(lambda) sequence (a). Seven variables with nonzero coefficients were selected by optimal lambda. By verifying the optimal parameter (lambda) in the LASSO model, the partial likelihood deviance (binomial deviance) curve was plotted versus log(lambda) and dotted vertical lines were drawn based on 1 standard error criteria (b).

Figure 3

Development of the risk nomogram (a) and the dynamic nomogram for an example (b). The DR risk nomogram was developed with the predictors including age, course of disease, PBG, HbA1c, UCR, UMA, and SBP.

Figure 4

ROC validation of the DR risk nomogram prediction. The y-axis meant the true-positive rate of the risk prediction. The x-axis meant the false-positive rate of the risk prediction. The blue line represented the performance of the nomogram. Figure 5(a) from the training set and Figure 5(b) from the validation set.

Figure 5

Calibration curves of the DR risk nomogram prediction. The y-axis meant the actual diagnosed DR. The x-axis meant the predicted risk of DR. The diagonal dotted line meant a perfect prediction by an ideal model. The solid line represented the performance of the training set (a) and validation set (b), which indicated that a closer fit to the diagonal dotted line represented a better prediction.

Figure 6

Decision curve analysis for the DR risk nomogram. The y-axis measured the net benefit. The thick solid line represented the assumption that all patients had no DR. The thin solid line represented the assumption that all patients had DR. The dotted line represented the risk nomogram. (a) From the training set and (b) from the validation set.