V.N. Grigorenko, O. V. Shcherbina, V.S.Sakalo, N.D.Sosnin, R.O.Danylets
SUMMARY
The article is devoted new approaches to the problem of early diagnostics of prostate cancer and differential diagnostics with benign prostate hyperplasia. For the construction of prognostics tests the non-parametric method of binary logistic regression, allowing to probe dependence of dichotomy variables (there are two states in our case: a prostate cancer or benign prostate hyperplasia) on independent variables most informing from which are percent of free and total PSA and closeness of PSA. 177 patients are inspected, the high informing offered approach is rotined in differential diagnostics of a prostate cancer or benign prostate hyperplasia.