Construction and clinical validation of a nomogram-based predictive model for diabetic retinopathy in type 2 diabetes

This study aimed to identify risk factors for diabetic retinopathy (DR) in patients with type 2 diabetes mellitus (T2DM) and construct a nomogram prediction model for DR.

T2DM patients (n = 520) who underwent funduscopic examinations from June 2020 to June 2022 were included. Of these patients, 220 had DR, yielding a disease rate of 40.38%. Patients were divided into a training set (n = 364) and a validation set (n = 156) at a 7:3 ratio. Feature variables were selected using LASSO regression, random forests, and decision trees. Venn diagrams identified common DR feature variables. The prediction model's validity was assessed using the C-index, decision curve analysis (DCA), receiver operating characteristic (ROC) curves, and calibration curves.

Factors influencing DR were age, Diabetic Peripheral Neuropathy (DPN), Hemoglobin A1C (HbA1C) levels, High-Density Lipoprotein (HDL) cholesterol, Low-Density Lipoprotein (LDL) cholesterol, Neutrophil-to-Lymphocyte Ratio (NLR), Triglycerides (TG), Blood Urea Nitrogen (BUN), and disease duration. Univariate analysis excluded LDL as being unrelated to DR. The DR prediction model, constructed using the remaining eight variables, showed internal validation metrics with a C-index of 0.937, area under the ROC curve (AUC) of 0.773, and DCA net benefit of 11%-95%. The external validation metrics demonstrated a C-index of 0.916, AUC of 0.735, and DCA net benefit of 17%-93%. Calibration curves indicated high consistency.

This study developed a nomogram prediction model to assess the risk of DR in patients with T2DM. The model demonstrated high precision through internal validation.