In order to improve the diagnostic accuracy of the serum total and free prostate-specific antigen (PSA) in differential diagnosis between benign prostate hyperplasia (BPH) and prostate cancer, the serum total sialic acid (TSA) was measured and logistic regression (LR) models were built. Significantly higher serum PSA (p<0.001) concentrations were observed in patients with prostate cancer compared to control subjects, but no statistically significant differences were found in serum TSA concentrations between these groups. Serum PSA reliably discriminated patients with prostate cancer from control subjects, the area under the ROC curve (AUC) being 0.991 (0.010). When serum PSA was in the gray zone, from 4 to 10 microg/l, the diagnostic accuracy of PSA in discriminating patients with prostate cancer from BPH patients was very poor, AUC being 0.563 (0.132). However, using the same set of patients the LR model combining serum PSA, free to total PSA ratio and TSA values, as well as digital rectal examination results, had good diagnostic accuracy in discriminating the prostate cancer patients from patients with BPH, the area under the ROC curve being 0.895 (0.054). The present data suggest that the logistic regression model combining laboratory measurements and results of the clinical examination may be a useful adjunct in the differential diagnosis of benign and malignant prostate disease.