|
Conference publicationsAbstractsXXIII conferenceApplication of the discriminant analysis for diagnosis of the tumor of the brainblack_drive@smtp.ru 1 pp. (accepted)Application of the discriminant analysis for preoperative diagnosis of a glial tumor of a brain is described. At all stages, for processing of statistical data, the tools of the Statistica program were applied. At the first stage on the training selection, values of coefficients in discriminant functions are calculated. These are functions diagnose the patient to one of three clusters. Clusters were marked: N – absence of a tumor; D – glial tumor of the I-II stage of an anaplasia; Z – glial tumor of the III-IV stage of an anaplasia. As parameters of discriminant functions, parameters of the immune status of the patient were chosen. The training selection for clusters of D (31 patients) and Z (30 patients) was based on comparison of values of parameters of the immune status to results of the histologic analysis of substance of a brain. Data for N cluster (41 "patients") were generated in intervals of norm on the casual law. At the second stage, on the basis of the received discriminant functions, it was carried out diagnosing of the patients passing conservative treatment. In the Statistica program diagnosing was carried out: on a posteriori probabilities. Results of katamnestichesky supervision over the patients passing conservative treatment confirmed operability of statistical model on the basis of the discriminant analysis.
|
