Machine learning-based prediction of LDL cholesterol: performance evaluation and validation

• Published in: PeerJ (San Francisco, CA) Vol. 13; p. e19248
• Main Authors: Meng, Jing-Bi, An, Zai-Jian, Jiang, Chun-Shan
• Format: Journal Article
• Language: English
• Published: United States PeerJ. Ltd 09.04.2025; PeerJ, Inc; PeerJ Inc
• Subjects: Accuracy; Adult; Aged; Algorithms; Biochemistry; Cardiovascular diseases; Cholesterol; Cholesterol, LDL - blood; Correlation coefficients; Data mining; Data Mining and Machine Learning; Data Science; Datasets; Decision trees; Estimates; Female; High density lipoprotein; Humans; Hypertriglyceridemia; Hypertriglyceridemia - blood; Learning algorithms; Lipids; Lipoproteins; Low density lipoprotein; Low density lipoproteins; Low-density lipoprotein cholesterol; Machine Learning; Machinie learning; Male; Medical laboratories; Methods; Middle Aged; Multilayer perceptrons; Performance evaluation; Phenols; Quality control; Regression analysis; Risk assessment; Risk Assessment - methods; Triglyceride; Triglycerides; Triglycerides - blood; China; Machinie learning; Lipids; Triglyceride; Low-density lipoprotein cholesterol
• ISSN: 2167-8359; 2167-8359; 2376-5992
• DOI: 10.7717/peerj.19248

This study aimed to validate and optimize a machine learning algorithm for accurately predicting low-density lipoprotein cholesterol (LDL-C) levels, addressing limitations of traditional formulas, particularly in hypertriglyceridemia. Various machine learning models-linear regression, K-nearest neighbors (KNN), decision tree, random forest, eXtreme Gradient Boosting (XGB), and multilayer perceptron (MLP) regressor-were compared to conventional formulas (Friedewald, Martin, and Sampson) using lipid profiles from 120,174 subjects (2020-2023). Predictive performance was evaluated using R-squared ( ), mean squared error (MSE), and Pearson correlation coefficient (PCC) against measured LDL-C values. Machine learning models outperformed traditional methods, with Random Forest and XGB achieving the highest accuracy (  = 0.94, MSE = 89.25) on the internal dataset. Among the traditional formulas, the Sampson method performed best but showed reduced accuracy in high triglyceride (TG) groups (TG > 300 mg/dL). Machine learning models maintained high predictive power across all TG levels. Machine learning models offer more accurate LDL-C estimates, especially in high TG contexts where traditional formulas are less reliable. These models could enhance cardiovascular risk assessment by providing more precise LDL-C estimates, potentially leading to more informed treatment decisions and improved patient outcomes.