A Metabolism-Based Interpretable Machine Learning Prediction Model for Diabetic Retinopathy Risk: A Cross-Sectional Study in Chinese Patients with Type 2 Diabetes

Guo-Wei Zong,Wan-Ying Wang,Jun Zheng,Wei Zhang,Wei-Ming Luo,Zhong-Ze Fang,Qiang Zhang

DOI: https://doi.org/10.1155/2023/3990035

IF: 4.0608

2023-05-17

Journal of Diabetes Research

Abstract:The burden of diabetic retinopathy (DR) is increasing, and the sensitive biomarkers of the disease were not enough. Studies have found that the metabolic profile, such as amino acid (AA) and acylcarnitine (AcylCN), in the early stages of DR patients might have changed, indicating the potential of metabolites to become new biomarkers. We are amid to construct a metabolite-based prediction model for DR risk. This study was conducted on type 2 diabetes (T2D) patients with or without DR. Logistic regression and extreme gradient boosting (XGBoost) prediction models were constructed using the traditional clinical features and the screening features, respectively. Assessing the predictive power of the models in terms of both discrimination and calibration, the optimal model was interpreted using the Shapley Additive exPlanations (SHAP) to quantify the effect of features on prediction. Finally, the XGBoost model incorporating AA and AcylCN variables had the best comprehensive evaluation ( ROCAUC = 0.82 , PRAUC = 0.44 , Brier score = 0.09 ). C18 : 1OH lower than 0.04 μmol/L, C18 : 1 lower than 0.70 μmol/L, threonine higher than 27.0 μmol/L, and tyrosine lower than 36.0 μmol/L were associated with an increased risk of developing DR. Phenylalanine higher than 52.0 μmol/L was associated with a decreased risk of developing DR. In conclusion, our study mainly used AAs and AcylCNs to construct an interpretable XGBoost model to predict the risk of developing DR in T2D patients which is beneficial in identifying high-risk groups and preventing or delaying the onset of DR. In addition, our study proposed possible risk cut-off values for DR of C18 : 1OH, C18 : 1, threonine, tyrosine, and phenylalanine.

endocrinology & metabolism,medicine, research & experimental

What problem does this paper attempt to address?

This paper aims to solve the problem of risk prediction for diabetic retinopathy (DR). Specifically, the research objective is to construct an interpretable machine - learning prediction model based on metabolites for predicting the risk of DR in patients with type 2 diabetes (T2D). Through this model, researchers hope to identify high - risk groups and prevent or delay the occurrence of DR. ### Research Background Diabetic retinopathy (DR) is a common microvascular complication of diabetes and one of the leading causes of blindness among working - age people. In China, the prevalence of DR is relatively high, and the early symptoms are not obvious, while in the late stage it is irreversible and the treatment effect is limited. Therefore, finding new biomarkers and developing effective prediction models are of great significance for early identification of high - risk groups. ### Research Methods 1. **Data Collection**: The study included 1,032 patients with type 2 diabetes, among which 162 had DR and 870 did not have DR. The clinical characteristics, amino acid (AA) and acyl - carnitine (AcylCN) metabolite profile data of the patients were collected. 2. **Feature Selection**: LASSO regression was used for feature screening, and finally 15 features were selected, including 7 amino acids (alanine, citrulline, glutamic acid, ornithine, phenylalanine, threonine, tyrosine) and 3 acyl - carnitines (C18:1, C18:1OH, C18:2), as well as the patient's age, systolic blood pressure, total cholesterol and the course of type 2 diabetes. 3. **Model Construction and Validation**: Logistic regression (LR) and extreme gradient boosting (XGBoost) algorithms were used respectively to construct prediction models. The performance of the models was evaluated by the area under the receiver operating characteristic curve (ROC) (AUC), the area under the precision - recall curve (PR) (PRAUC) and the Brier score. 4. **Model Explanation**: The Shapley Additive exPlanations (SHAP) method was used to explain the model and quantify the influence of each feature on the prediction result. ### Main Findings - **Best Model**: The XGBoost model combined with AA and AcylCN variables had the highest comprehensive evaluation in the test set (ROCAUC = 0.82, PRAUC = 0.44, Brier score = 0.09). - **Important Features**: C18:1OH lower than 0.04 μmol/L, C18:1 lower than 0.70 μmol/L, threonine higher than 27.0 μmol/L, tyrosine lower than 36.0 μmol/L are associated with an increased risk of DR. Phenylalanine higher than 52.0 μmol/L is associated with a reduced risk of DR. ### Conclusion The study successfully constructed an interpretable machine - learning model based on metabolites for predicting the risk of DR in patients with type 2 diabetes. This model not only helps to identify high - risk groups, but also proposes possible DR risk thresholds, providing a scientific basis for early prevention and intervention.

A Metabolism-Based Interpretable Machine Learning Prediction Model for Diabetic Retinopathy Risk: A Cross-Sectional Study in Chinese Patients with Type 2 Diabetes

Ethanolamine As a Biomarker and Biomarker-Based Therapy for Diabetic Retinopathy in Glucose-Well-Controlled Diabetic Patients

A risk prediction model for type 2 diabetes mellitus complicated with retinopathy based on machine learning and its application in health management

Interpretable Machine Learning-Derived Nomogram Model for Early Detection of Diabetic Retinopathy in Type 2 Diabetes Mellitus: a Widely Targeted Metabolomics Study

An enhanced machine learning algorithm for type 2 diabetes prognosis with a detailed examination of Key correlates

Hybrid Explainable Artificial Intelligence Models for Targeted Metabolomics Analysis of Diabetic Retinopathy

Interpretable machine learning identifies metabolites associated with glomerular filtration rate in type 2 diabetes patients

Development and External Validation of Machine Learning Models for Diabetic Microvascular Complications: Cross-Sectional Study With Metabolites

Explainable Machine Learning-Based Prediction Model for Diabetic Nephropathy

Predicting diabetic retinopathy and identifying interpretable biomedical features using machine learning algorithms

Predicting the risk of subclinical atherosclerosis based on interpretable machine models in a Chinese T2DM population

Metabolite-assisted models improve risk prediction of coronary heart disease in patients with diabetes

Explainable Artificial Intelligence Paves the Way in Precision Diagnostics and Biomarker Discovery for the Subclass of Diabetic Retinopathy in Type 2 Diabetics

Ten metabolites-based algorithm predicts the future development of type 2 diabetes in Chinese

Novel risk score model for non-proliferative diabetic retinopathy based on untargeted metabolomics of venous blood

Prediction and analysis of risk factors for diabetic retinopathy based on machine learning and interpretable models

Plasma Metabolomics Identifies Key Metabolites and Improves Prediction of Diabetic Retinopathy: Development and Validation across Multinational Cohorts

A new predictive model for the concurrent risk of diabetic retinopathy in type 2 diabetes patients and the effect of metformin on amino acids

Development and external validation of a predictive model for type 2 diabetic retinopathy

Metabolomics-based Multidimensional Network Biomarkers for Diabetic Retinopathy Identification in Patients with Type 2 Diabetes Mellitus

[Development and application of a prediction model for incidence of diabetic retinopathy in newly diagnosed type 2 diabetic patients based on regional health data platform]