Improving precision of glomerular filtration rate estimating model by ensemble learning

Journal of Translational Medicine

Research

Tanqi Lou¹  Shaomin Li¹  Cailian Cheng¹  Caixia Wang¹  Yuqiu Ye¹  Xun Liu²  Yongmei Fu³  Linsheng Lv⁴  Ningshan Li⁵ 

[1] Division of Nephrology, Department of Internal Medicine, The Third Affiliated Hospital of Sun Yat-sen University, 510630, Guangzhou, China;Division of Nephrology, Department of Internal Medicine, The Third Affiliated Hospital of Sun Yat-sen University, 510630, Guangzhou, China;Division of Nephrology, The 3rd Affiliated Hospital of Sun Yat-sen University, Yuedong Hospital, 514700, Meizhou, China;Emergency Department, The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China;Operating Room, The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China;SJTU-YALE Joint Center for Biostatistics, Shanghai JiaoTong University, Shanghai, China;

关键词: Chronic kidney disease;    Glomerular filtration rate;    Ensemble learning;    Prediction;    Precision;   

DOI  :  10.1186/s12967-017-1337-y

 received in 2017-05-29, accepted in 2017-11-01,  发布年份 2017

来源: Springer

【 摘 要 】

BackgroundAccurate assessment of kidney function is clinically important, but estimates of glomerular filtration rate (GFR) by regression are imprecise.MethodsWe hypothesized that ensemble learning could improve precision. A total of 1419 participants were enrolled, with 1002 in the development dataset and 417 in the external validation dataset. GFR was independently estimated from age, sex and serum creatinine using an artificial neural network (ANN), support vector machine (SVM), regression, and ensemble learning. GFR was measured by 99mTc-DTPA renal dynamic imaging calibrated with dual plasma sample 99mTc-DTPA GFR.ResultsMean measured GFRs were 70.0 ml/min/1.73 m2 in the developmental and 53.4 ml/min/1.73 m2 in the external validation cohorts. In the external validation cohort, precision was better in the ensemble model of the ANN, SVM and regression equation (IQR = 13.5 ml/min/1.73 m2) than in the new regression model (IQR = 14.0 ml/min/1.73 m2, P < 0.001). The precision of ensemble learning was the best of the three models, but the models had similar bias and accuracy. The median difference ranged from 2.3 to 3.7 ml/min/1.73 m2, 30% accuracy ranged from 73.1 to 76.0%, and P was > 0.05 for all comparisons of the new regression equation and the other new models.ConclusionsAn ensemble learning model including three variables, the average ANN, SVM, and regression equation values, was more precise than the new regression model. A more complex ensemble learning strategy may further improve GFR estimates.

【 授权许可】

CC BY   
© The Author(s) 2017

【 图 表 】

12936_2017_2051_Article_IEq85.gif

Fig. 1