K Fischer1, H Loertzer, P Fornara. 1. University Clinic for Urology, Martin-Luther-University Halle-Wittenberg, Halle, Germany. kersten.fischer@medizin.uni-halle.de
Abstract
UNLABELLED: The advent of complexed PSA (cPSA) raised great expectations concerning the role of this parameter for improving the early detection of prostate cancer. MATERIALS AND METHODS: Total PSA (tPSA), free PSA (fPSA) and cPSA were evaluated from the serum of 178 of our clinic's patients (74 patients with prostate carcinoma, 104 patients with benign prostate illness) prior to prostate histology. ROC curves were calculated for all of these parameters as well as for the ratios fit-PSA, c/t-PSA and f/c-PSA. RESULTS: The ROC analysis for the whole examined PSA area and PSA levels of 4 to 10 ng/ml showed a statistically significant difference between the AUCs of the ratios on the one hand and the cPSA and tPSA parameters on the other hand. However, there was no difference between these parameters in PSA levels of up to 6 ng/ml. In the comparison of specificities at PSA levels of 4 to 10 ng/ml, the best results were achieved for the c/tPSA ratio. Neither in the PSA level area between 4 and 10 nglml, nor in the whole examined PSA area, could a difference between the cPSA and tPSA parameters be detected. CONCLUSION: Firm conclusions regarding low PSA concentrations cannot be drawn because of the small number of cases included in our study. However, 5 out of 13 patients with prostate carcinoma, whose tPSA values were still in the employed method's reference area, would have been identified as carcinoma-suspicious and brought to further diagnosis by determining the cPSA value with a recommended cut-off of 2.5 ng/ml.
UNLABELLED: The advent of complexed PSA (cPSA) raised great expectations concerning the role of this parameter for improving the early detection of prostate cancer. MATERIALS AND METHODS: Total PSA (tPSA), free PSA (fPSA) and cPSA were evaluated from the serum of 178 of our clinic's patients (74 patients with prostate carcinoma, 104 patients with benign prostate illness) prior to prostate histology. ROC curves were calculated for all of these parameters as well as for the ratios fit-PSA, c/t-PSA and f/c-PSA. RESULTS: The ROC analysis for the whole examined PSA area and PSA levels of 4 to 10 ng/ml showed a statistically significant difference between the AUCs of the ratios on the one hand and the cPSA and tPSA parameters on the other hand. However, there was no difference between these parameters in PSA levels of up to 6 ng/ml. In the comparison of specificities at PSA levels of 4 to 10 ng/ml, the best results were achieved for the c/tPSA ratio. Neither in the PSA level area between 4 and 10 nglml, nor in the whole examined PSA area, could a difference between the cPSA and tPSA parameters be detected. CONCLUSION: Firm conclusions regarding low PSA concentrations cannot be drawn because of the small number of cases included in our study. However, 5 out of 13 patients with prostate carcinoma, whose tPSA values were still in the employed method's reference area, would have been identified as carcinoma-suspicious and brought to further diagnosis by determining the cPSA value with a recommended cut-off of 2.5 ng/ml.