北京生物医学工程

基于主成分分析的深度前馈神经网络的肾小球滤过率估算算法

Glomerular filtration rate estimation algorithm based on principal component analysis with deep feedforward neural network

作者：王露露杨震黄山张罡李飞詹曙

单位：大数据知识工程教育部重点实验室（合肥 230601）<br />合肥工业大学计算机与信息学院（合肥 230601）<br />安徽医科大学第二附属医院（合肥 230601）<br />通信作者：李飞。E-mail：lifei007@139.com

关键词：慢性肾脏病；肾小球滤过率；主成分分析；深度前馈神经网络;估算模型

分类号：R318

出版年·卷·期（页码）：2023·42·2（164-169）

摘要：

目的提出一种基于主成分分析（principal component analysis，PCA）的深度前馈神经网络（deep feedforward neural network, DFNN），建立一个适用于中国慢性肾脏病人群的肾小球滤过率估算模型，并探讨其在慢性肾脏病患者肾小球滤过率（glomerular filtration rate, GFR）估算中的应用。方法受试者为2019年5月—2021年1月就诊于安徽医科大学第二附属医院、年龄<18岁的肾功能不稳定，且排除服用甲氧苄啶或西咪替丁或接受透析的163例患者。本研究以肾动态显像测定GFR为标准，建立主成分分析的深度前馈神经网络模型，以此估算GFR，同时将估算GFR结果与传统CG方程和BP神经网络估算结果进行对比分析。结果通过PCA-DFNN-1神经网络训练出来的估算模型的15%符合率、30%符合率、50%符合率分别为为38.77%、55.1%、75.5%；曲线下面积ROC为0.845；Youden指数为0.58。结论提出的基于主成分分析的深度前馈神经网络模型有优于CG方程和BP神经网络模型的结果，可以用于估算GFR。

Objective A deep feedforward neural network based on principal component analysis was proposed to establish a glomerular filtration rate estimation model suitable for Chinese chronic kidney disease population, and to explore its use in the estimation of glomerular filtration rate (GFR) in chronic kidney disease patients. Methods The participants were 163 patients who were visited the The Second Hospital of Anhui Medical University from May 2019 to January 2021, and were under 18 years old with unstable renal function, and were excluded from taking trimethoprim or crimetidine or receiving dialysis. In this study, the GFR was determined by dynamic renal imaging as the standard, and a deep feedforward neural network model of principal component analysis was established to estimate GFR. Results The 15%, 30%, and 50% coincidence rates of the estimated models trained by the PCA-DFNN-1 neural network were 38.77%, 55.1%, and 75.5%; the area under the curve ROC was 0.845; the Youden index was 0.58. Conclusions The proposed deep feedforward neural network model based on principal component analysis has better results than CG equation and BP neural network model, and can be used to estimate GFR.

参考文献：

[1] 李丽琴. 基于人工智能的慢性肾病分级预警模型[J]. 时代报告（下半月）, 2012(5): 330-331.
[2] 徐静. 径向基函数神经网络评估慢性肾脏病肾小球滤过率[D]. 大连: 大连医科大学, 2016.
Xu J. Evaluate glomerular filtration rate by the radial basis function neural network in patients with chronic kidney disease [D]. Dalian: Dalian Medical University, 2016.
[3] Zahran A, EI-Husseini A, Shoker A. Can cystatin C replace creatinine to estimate glomerular filtration rate? A literature review[J]. American Journal of Nephrology, 2007, 27(2): 197-205.
[4] Filler G, B?Kenkamp A, Hofmann W, et al. Cystatin C as a marker of GFR--history, indications, and future research[J]. Clinical Biochemistry, 2005, 38(1): 1-8.
[5] 陈有维, 万福俊, 何乐愚, 等. ?BP 神经网络用于评估肾小球滤过率的研究[J]. 临床医学, 2008, 28(5): 107-109, 128.?
Chen YW, Wan FJ, He LY, et al. Study of BP neural network in predicting renal glomerular filtration rate[J].Clinical Medicine, 2008, 28(5): 107-109, 128.
[6] 张雨浓，刘迅，何良宇，等. 应用WASD神经网络估算肾小球滤过率的研究[J]. 中国科技信息，2014(8): 212-216.
Zhang YN, Liu X, He LY, et al. Research on application of WASD neural network to estimating glomerular filtration rate[J]. China Science and Technology Information,2014(8): 212-216.?
[7] 张雨浓，何良宇，刘迅，等. 应用RBF激励WASD神经网络估算GFR[J]. 计算技术与自动化，2016, 35(1): 22-26.
Zhang YN, He LY, Liu X, et al. Application of RBF-activated ?WASD neuronet in estimating GFR[J]. Computing ? Technology and Automation, 2016, 35(1): 22-26.
[8] 邹海英，李智，杨帆. 基于特征选择的自适应模糊神经网络在肾小球滤过率中的应用[J]. 软件导刊，2018, 17(6): 153-156.
[9] 李宁山，刘迅，吴效明，等. 人工神经网络在肾小球滤过率估算中的应用[J]. 第三军医大学学报，2012, 34(5): 409-411.
Li NS, Liu X, Wu XM, et al. Estimating glomerular ?filtration ?rate with artificial neural network: a model establishment[J]. Acta Academiae Medicinae Militaris Tertiae, 2012, 34(5): 409-411.
[10] Li NS, Huang H, Qian HZ, et al. Improving accuracy of estimating glomerular filtration rate using artificial neural network: model development and validation[J]. ?Journal of Translational Medicine, 2020, 18: 120.
[11] Liu X, Pei XH, Li NS, et al. Improved glomerular filtration rate estimation by an artificial neural network[J]. PLoS One, 2013, 8(3): e58242.
[12] Liu X, Chen YR, Li NS, et al. Estimation of glomerular filtration rate by a radial basis function neural network in patients with type-2 diabetes mellitus[J]. BMC Neuphrology, 2013, 14: 181.
[13] Xu J, Guo BY, Liu CY. Evaluation of glomerular filtration rate in chronic kidney disease by radial basis function neural network[J]. Transplantation Proceedings, 2020, 52(3) : 748-753.
[14] 梁哲.改进的RBF神经网络在肾小球滤过率估算中的应用[D]. 大连: 大连理工大学, 2016.
Liang Z. Application of improved RBF neural networks on glomerular filtration rate estimation[D]. Dalian: Dalian University of Technology, 2016.
[15] 杨万元，王胜男，赵卫红，等. 肾小球滤过率估算模型研究[J].生物医学工程学杂志, 2013, 30(5): 963-967.?
Yang WY, Wang SN, Zhao WH, et al. Research of the glomerular ?filtration ?rate ?estimation ?model[J]. Journal of Biomedical Engineering, 2013, 30(5): 963-967.
[16] 高峰，吴晓东，周科平. 基于主成分分析和PSO-ELM算法的排土场稳定性预测模型[J]. 黄金科学技术，2021, 29(5): 658-668.
Gao F, Wu XD, Zhou KP. Prediction model of soil dump stability based on principal component analysis and PSO-ELM algorithm[J]. Gold Science and Technology, 2021, 29(5): 658-668.
[17] 敬微微，韩倩，吴昊，等. 主成分分析和反向传播神经网络模型在血液透析机预防维护中的应用[J]. 中国医学装备, 2020, 17(7): 137-140.
Jing WW, Han Q, Wu H, et al. Application of PCA and BP neural network model in the preventive maintenance of hemodialysis machine[J]. China Medical Equipment, 2020, 17(7): 137-140.
[18] 吴定安，钟建伟，王新磊，等. 主成分分析和长短期记忆网络的电力负荷预测[J]. 物联网技术, 2021, 11(8): 47-51.
[19] 刘斌，李立欣，李静. 一种改进的基于深度前馈神经网络的极化码BP译码算法[J]. 移动通信，2019, 43(4): 8-14.
Liu B, Li LX, Li J. An improved polar BP decoding algorithm based on deep feedforward neural network[J]. Mobile Communications, 2019, 43(4): 8-14.
[20] 杜方洲. 基于深度前馈神经网络的TRMM降水产品降尺度研究[D]. 南京: 南京信息工程大学，2020.
Du FZ. Downscaling of TRMM precipitation products based on deep feedforward neural network[D]. ?Nanjing: Nanjing University of Information Science and Technology, 2020.
[21] Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine[J]. Nephron, 1976, 16(1): 31-41

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录