Evaluation of Glomerular Filtration Rate in Chronic Kidney Disease by Radial Basis Function Neural Network.

OBJECTIVE: To develop a radial basis function (RBF) neural network and investigate its performance in the estimation of glomerular filtration rate (GFR) for patients with chronic kidney disease. METHODS: A total of 651 patients with chronic kidney disease were enrolled in this study. The GFR measured by (99m)Tc-DTPA renal dynamic imaging was used as the standard GFR. The RBF neural network model was established and the performance prediction GFR value was verified. It was found that the RBF neural network could better evaluate the GFR of patients with chronic kidney disease stage 2-5, which is superior to the Modification of Diet in Renal Disease equation. CONCLUSIONS: The RBF neural network evaluated GFR significantly for patients with chronic kidney disease stages 2-5, and it showed no difference with the (99m)Tc-DTPA renal dynamic imaging method, and it can be used for estimated GFR evaluation.