The paper deals with the possibility analysis of developing a cancer diseases classification model based on layered artificial neural networks. We studied the data obtained by the IMMUNOSIGNATURE technology. Immunosignaturing requires a drop of test subject blood that contains cells of immune system to be applied to a biochip. The biochip is divided into sectors with unique amino acid sequences. Depending on the intensity of the immune response manifestation in specific cells, a specific profile (signature) is formed for a certain individual. The complexity of immunosignature data analysis lies in external stochastic influence on the obtained results. In the paper, aspects of correlation analysis data correction to get rid of random artifacts are considered. Data normalization procedure to minimize the impact of systemic distortions due to the influence of external factors, such as temperature, humidity, substance concentration, etc., is discussed. Classification is performed by the backpropagation artificial neural network.