Automated classification of dynamic renal scintigraphy exams to determine the stage of chronic kidney disease: an investigation

Authors:
Publisher website
Google Scholar
Abstract
The glomerular filtration rate determination is the principal method to determine a patient’s chronic kidney disease stage. Currently, there are two main methods for its determination: blood and urine tests and renogram techniques. This paper aims to investigate the performance of machine learning techniques in the automated classification of renal scintigraphy exams to assist in determining the stage of chronic renal disease. To do so, we used the Database of Dynamic Renal Scintigraphy, containing 107 exams of dynamic renal scintigraphy, in which there are patients at various stages of chronic kidney disease. The images were segmented and, afterward, generated the renogram curves used as input to the classifiers. SVM (Support Vector Machine), KNN (K-Nearest Neighbors), Decision Trees, and Ensemble were used as classifiers. The method that showed the best results was the SVM using the original curves as input. However, all the methods generated satisfactory results.

This publication has 23 references indexed in Scilit: