Abstract: “The paper presents an automatic approach to assessment of the stage of development of the kidney cancer on the basis of Fuhrman grades. The stage of advancement level of cancer is usually associated with 4 Fuhrman grades. Our approach to Fuhrman grade assessment is composed of few steps. The first one is extraction of the numerical features from the microscopic image of the histological slides of the biopsy of kidney by using mathematical morphology. The next step is the features selection providing descriptors of the best class separating abilities. The last one is application of the automatic classifiers and data mining techniques to assign the actually available samples to one of four classes.”

Keywords: features selection, neural networks, classification system, kidney cancer, data mining