Analysis of structural features of multidimensional data for peripheral blood in chronic lymphocytic leukemia and non-Hodgkin В-lymphoma compared with the norm

Abstract

Background. Development of modern methods for diagnosis of various forms of lymphoproliferative diseases is still relevant. Evaluation of multidimensional relationships in the immunity-tumor growth system creates an opportunity to differentiate these diseases based on blood tests, without using materials from other organs and tissues for diagnosis.

Aim. To introduce methods of the multidimensional analysis of the composition of lymphocyte and tumor cell subpopulations in peripheral blood to identify the immuno-structural features of non-Hodgkin B-cell lymphomas (NHL), and chronic lymphocytic leukemia (CLL) by comparison with the norm.

Methods. Data analysis was carried out using an artificial neural network - self-organizing Kohonen maps (SOM (self organased map). This ANN allows one to reveal the structure in multidimensional data by projecting multidimensional images into a reduced-dimensional space (2- or 3-dimensional).

Results. To detect the differences in multidimensional images of CLL and NHL from the norm, artificial neural networks were used. The state of immunity and tumor cells was compared in patients with B-cell NHL (352 patients) and CLL (315 patients) and in healthy people (184 people) based on data for the peripheral blood. The structures reflecting the differences between «healthy people – NHL patients» and «healthy people - CLL patients» were identified. The nature of distributions and the values of immunity and tumor growth indicators in a multidimensional space distinguish between CLL-normal state and NHL- normal state.

Conclusion. The differences between NHL and CLL images detected by the multivariate analysis provide a basis for creating an algorithm for automated diagnosis of NHL and CLL.