Predicting diabetic retinopathy based on routine laboratory tests by machine learning algorithms

被引:0
|
作者
Wan, Xiaohua [1 ,2 ,3 ]
Zhang, Ruihuan [4 ]
Wang, Yanan [4 ]
Wei, Wei [5 ]
Song, Biao [4 ]
Zhang, Lin [5 ,6 ,7 ]
Hu, Yanwei [1 ,2 ]
机构
[1] Capital Med Univ, Beijing Chao Yang Hosp, Dept Clin Lab, Beijing, Peoples R China
[2] Beijing Ctr Clin Labs, Beijing, Peoples R China
[3] Capital Med Univ, Beijing Tongren Hosp, Dept Clin Lab, Beijing, Peoples R China
[4] Inner Mongolia Med Intelligent Diag Big Data Res I, Hohhot, Inner Mongolia, Peoples R China
[5] Capital Med Univ, Beijing Tongren Hosp, Dept Med Record, Beijing, Peoples R China
[6] Capital Med Univ, Beijing Tongren Hosp, Dept Endocrinol, Beijing, Peoples R China
[7] Beijing Diabet Res Inst, Beijing, Peoples R China
关键词
Type 2 diabetes mellitus; Diabetic retinopathy; Routine laboratory tests; Machine learning; XGBoost; Predictive model; CLASSIFICATION; COMPLICATIONS; DIAGNOSIS; SYSTEM; RISK;
D O I
10.1186/s40001-025-02442-5
中图分类号
R-3 [医学研究方法]; R3 [基础医学];
学科分类号
1001 ;
摘要
ObjectivesThis study aimed to identify risk factors for diabetic retinopathy (DR) and develop machine learning (ML)-based predictive models using routine laboratory data in patients with type 2 diabetes mellitus (T2DM).MethodsClinical data from 4259 T2DM inpatients at Beijing Tongren Hospital were analyzed, divided into a model construction data set (N = 3936) and an external validation data set (N = 323). Using 39 optimal variables, a prediction model was constructed using the eXtreme Gradient Boosting (XGBoost) algorithm and compared with four other algorithms: support vector machine (SVM), gradient boosting decision tree (GBDT), neural network (NN), and logistic regression (LR). The Shapley Additive exPlanation (SHAP) method was employed to interpret the XGBoost model. External validation was performed to assess model performance.ResultsDR was present in 47.69% (N = 1877) of T2DM patients in the model construction data set. Among the models tested, the XGBoost model performed best with an AUC of 0.831, accuracy of 0.757, sensitivity of 0.754, specificity of 0.759, and F1-score of 0.752. SHAP explained feature importance for XGBoost model and identified key risk factors for DR. External validation yielded an accuracy of 0.650 for the XGBoost model.ConclusionsThe XGBoost-based prediction model effectively assesses DR risk in T2DM patients using routine laboratory data, aiding clinicians in identifying high-risk individuals and guiding personalized management strategies, especially in medically underserved areas.
引用
收藏
页数:19
相关论文
共 50 条
  • [31] Personalizing routine lab tests with machine learning
    Alice Tang
    Tomiko Oskotsky
    Marina Sirota
    Nature Medicine, 2021, 27 : 1514 - 1515
  • [32] Personalizing routine lab tests with machine learning
    Tang, Alice
    Oskotsky, Tomiko
    Sirota, Marina
    NATURE MEDICINE, 2021, 27 (09) : 1514 - 1515
  • [33] Explainable machine learning model for predicting the risk of significant liver fibrosis in patients with diabetic retinopathy
    Zhu, Gangfeng
    Yang, Na
    Yi, Qiang
    Xu, Rui
    Zheng, Liangjian
    Zhu, Yunlong
    Li, Junyan
    Che, Jie
    Chen, Cixiang
    Lu, Zenghong
    Huang, Li
    Xiang, Yi
    Zheng, Tianlei
    BMC MEDICAL INFORMATICS AND DECISION MAKING, 2024, 24 (01)
  • [34] Fundus photograph-based deep learning algorithms in detecting diabetic retinopathy
    Rajiv Raman
    Sangeetha Srinivasan
    Sunny Virmani
    Sobha Sivaprasad
    Chetan Rao
    Ramachandran Rajalakshmi
    Eye, 2019, 33 : 97 - 109
  • [35] Fundus photograph-based deep learning algorithms in detecting diabetic retinopathy
    Raman, Rajiv
    Srinivasan, Sangeetha
    Virmani, Sunny
    Sivaprasad, Sobha
    Rao, Chetan
    Rajalakshmi, Ramachandran
    EYE, 2019, 33 (01) : 97 - 109
  • [36] Application of machine learning algorithms in predicting carotid artery plaques using routine health assessments
    Wei, Yuting
    Tao, Junlong
    Geng, Yifan
    Ning, Yi
    Li, Weixia
    Bi, Bo
    FRONTIERS IN CARDIOVASCULAR MEDICINE, 2024, 11
  • [37] Application of deep learning algorithms for diabetic retinopathy screening
    Poschkamp, Broder
    Stahl, Andreas
    ANNALS OF TRANSLATIONAL MEDICINE, 2022, 10 (24)
  • [38] Predicting psoriasis using routine laboratory tests with random forest
    Zhou, Jing
    Li, Yuzhen
    Guo, Xuan
    PLOS ONE, 2021, 16 (10):
  • [39] Risk Stratification of COVID-19 Using Routine Laboratory Tests: A Machine Learning Approach
    Mlambo, Farai
    Chironda, Cyril
    George, Jaya
    INFECTIOUS DISEASE REPORTS, 2022, 14 (06) : 900 - 931
  • [40] Machine Learning Approach Using Routine Immediate Postoperative Laboratory Values for Predicting Postoperative Mortality
    Cho, Jaehyeong
    Park, Jimyung
    Jeong, Eugene
    Shin, Jihye
    Ahn, Sangjeong
    Park, Min Geun
    Park, Rae Woong
    Park, Yongkeun
    JOURNAL OF PERSONALIZED MEDICINE, 2021, 11 (12):