The most common test for the detection of prostate cancer is based on the PSA level of blood. This inexpensive monitoring method can be repeated quite often compared to the empirical propagation velocity of the disease. The disadvantage of this test is that besides tumour other factors may also increase the PSA level (e.g. inflammation), therefore the number of false positive results is high [1].In this article we keep the series of one's measured PSA levels as a stochastic process and determine the coefficients of a quadratic equation, which suitable to estimate the expected value of PSA level at a given time. We give formulas for the speed and acceleration of the process, as well as counting the time when a measured value will doubles, or the time when the PSA level will exceed a critical value. The article shows both the closed version and the recursive version of this second-level polynomial [2].In the end we create the fundamental algorithms of a hypothetical IT service (called PRO-FILTER) modelling the mathematical results 1 .