Fractal analysis and Gray level co-occurrence matrix method for evaluation of reperfusion injury in kidney medulla

Abstract

Fractal analysis and Gray level co-occurrence matrix method represent two novel mathematical algorithms commonly used in medical sciences as potential parts of computer-aided diagnostic systems. In this study, we tested the ability of these methods to discriminate the kidney medullar tissue suffering from reperfusion injury, from normal tissue. A total of 320 digital micrographs of Periodic acid-Schiff (PAS) – stained kidney medulla from 16 Wistar albino mice (20 per animal), were analyzed using National Institutes of Health ImageJ software (NIH, Bethesda, MD) and its plugins. 160 micrographs were obtained from the experimental group with induced reperfusion injury, and another 160 were obtained from the controls. For each micrograph we calculated the values of fractal dimension, lacunarity, as well as five GLCM features: angular second moment, entropy, inverse difference moment, GLCM contrast, and GLCM correlation. Discriminatory value of the parameters was tested using receiver operating characteristic (ROC) analysis, by measuring the area below ROC curve. The results indicate that certain features of GLCM algorithm have excellent discriminatory ability in evaluation of damaged kidney tissue. Fractal dimension and lacunarity as parameters of fractal analysis also had a relatively good discriminatory value in differentiation of injured from the normal tissue. Both methods have potentially promising application in future design of novel techniques applicable in cell physiology, histology and pathology.