Neural network model outperforms conventional equations in LDL cholesterol estimation: A comparative study of 188,887 Chinese individuals with focus on hypertriglyceridemia

神经网络模型在低密度脂蛋白胆固醇估算方面优于传统方程:一项针对188,887名中国受试者(重点关注高甘油三酯血症)的比较研究

阅读:1

Abstract

BACKGROUND: Accurate estimation of low-density lipoprotein cholesterol (LDL-C) is crucial for atherosclerotic cardiovascular disease (ASCVD) risk management. Extensive international validation studies have demonstrated that traditional formulas (Friedewald, Martin/Hopkins, Sampson) often yield significant errors under conditions of extreme hypertriglyceridemia. This study aimed to assess the performance of these conventional formulas in Chinese populations and develop a novel neural network-based LDL-C estimation model [LDL-C(NN)]. METHODS: In this retrospective study, we analyzed 188,887 lipid profiles-including total cholesterol, triglycerides, high-density lipoprotein cholesterol, and directly measured LDL-C-from Peking University Shenzhen Hospital using Mindray (outpatients, n = 83,731) and Beckman (inpatients, n = 105,156) systems. The test results from the two detection systems are non-overlapping. We used stratified random sampling based on TG levels to select 30,000 profiles from each of the two systems as the training dataset (60,000 profiles in total). Within this training dataset, 70 % of profiles were used for parameter learning, 15 % were used for early-stopping validation, and 15 % were used for post-training testing. The remaining profiles constituted the independent test set for the final performance evaluation (Mindray: n = 53,731; Beckman: n = 75,156). We then compared the performance of LDL-C(NN) with the Friedewald, Martin/Hopkins, and Sampson formulas using correlation coefficient (r), root mean square error (RMSE), Concordance Correlation Coefficient (CCC) and clinical risk stratification accuracy. RESULTS: Compared with directly measured LDL-C, LDL-C(NN) demonstrated higher correlation and lower RMSE than other traditional LDL-C equations in the Mindray system (r = 0.9778, RMSE = 0.1762 mmol/L; vs Friedewald quation: r = 0.8894, RMSE = 0.4783 mmol/L; vs Martin/Hopkins quation: r = 0.9658, RMSE = 0.2463 mmol/L; vs Sampson quation: r = 0.9548, RMSE = 0.2934 mmol/L, particularly patients with high triglycerides (TG levels, 9.03-13.56 mmol/L, neural network Model: CCC = 0.8750, vs Friedewald quation: CCC = 0.3320; vs Martin/Hopkins quation: CCC = 0.7278; vs Sampson quation: CCC = 0.4176). Beckman database shows the same performance. The clinical classification accuracy for LDL-C(NN) reached 87.5 % (Mindray) and 83.4 % (Beckman), surpassing that of other traditional LDL-C equations (66.6-78.7 %). CONCLUSIONS: By overcoming the linear assumptions of conventional equations, the neural network-based model significantly improves LDL-C estimation in hypertriglyceridemia (especially≥9.03 mmol/L) and complex lipid profiles, thereby expanding the applicability of traditional formulas, while demonstrating robust performance across multiple analytical systems.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。