CBC effectively stratifies the patients with different types of malignant tumors

Abstract

Early prediction of tumor process especially related to the detection of oncological diseases at early stage is a hot problem. We compared the complete blood count data for patients with various types of tumors; the patients with tumors of the hematopoietic system were excluded from the study. We used both linear (K-means) and non-linear (elastic maps) methods. No informative patterns have been found through the implementation of classic linear statistics methods. On the contrary, the elastic maps technique revealed the specific groups among the patients. Neither relation to the gender, nor age of patients has been found. However, the abundance of basophils, eosinophils, leukocytes and immunoglobulins exhibit specific pattern of the clusters occupancy. Thus, in our study we found that patients are stratified in accordance with quantitative changes in laboratory blood parameters, which is associated with the body’s response to stress caused by a malignant tumor.