Objectives: We review the results of 6000 patients with the clinical suspect of prostate cancer who underwent one or more prostate, biopsies, analyzing the role of PSA derived parameters in the probability of having prostate cancer in the TRUS biopsy.
Methods: We selected 6000 patients who under- went TRUS biopsy between 1994 and 2002. 861 of them underwent more than one is biopsy, adding up to a total of 7127 biopsies. For the study of PSA derived indexes we established ranges based on the 10th percentile for the first biopsy for all patients and also for those with PSA between 4 and 10 ng/ml. Several predictive models were determined by logistic regression of the variables related with presence/no presence of cancer.
Results: For first biopsies the ranges of PSAD established showed a diagnostic effectiveness below 8% with PSA densities lower than 0.11 ng/ml/cc. The free/total PSA ratio is less discriminant in the ranges obtained with a 13.7% incidence of prostate cancer for values above 0.24. In the case of second biopsies the group of patients with PSAD below 0.12 had only a 5.3% incidence, and only one patient with F/T PSA ratio higher than 0.24 had a prostate cancer (2.9%). All studied parameters but F/T PSA ratio showed statistical significance in the multivariant analysis.
Conclusions: Although the establishment of a cut point for PSAD diminishes sensitivity, prostate biopsy habits should be modified assuming the loss of tumors in patients with low PSAD and increasing the number of biopsies in patients with total PSA values below 4 ng/ml with higher densities.