Influence of free‐to‐total prostate specific antigen variability on the early diagnosis of prostate cancer: a comparative study of three immunoassays

Abstract

Objective To compare three immunoassays for total prostate specific antigen (tPSA), free PSA (fPSA), free-to-total PSA ratio (f/tPSA), and the f/tPSA thresholds for optimal differentiation between benign prostatic hypertrophy (BPH) and prostate cancer in 141 consecutive patients referred for prostatic disease. Patients and methods The study included 43 patients with prostate cancer and 98 with BPH, all confirmed histologically. PSA levels were assessed using the following assay kits just before histological analysis of the prostate; Hybritech (Tandem-R PSA, normal values, NV, <4 ng/mL, and Tandem-R free PSA), Cis Bio (PSA-RIACT, NV <2.5 ng/mL and fPSA-RIACT) and Immunocorp (PSA-IRMA, NV <4 ng/mL and Free PSA-IRMA). The results were assessed to determine the sensitivity, specificity and threshold values of the different assays to differentiate patients with BPH and cancer. Results The mean tPSA and f/tPSA ratio were statistically different in assays with different NVs. The mean fPSA values differed significantly between the Hybritech and Cis Bio, between the Hybritech and Immunocorp but not between the Cis Bio and Immunocorp assays. With receiver operator curve analysis, there were no statistically significant differences among the three immunoassays in f/tPSA (0.72 for Hybritech, 0.73 for Cis Bio and 0.64 for Immunocorp) or between the tPSA and fPSA curves for each manufacturer. With the sensitivity fixed at 90%, different f/tPSA thresholds were defined (0.22, 0.34 and 0.25 for Hybritech, Cis Bio and Immunocorp, respectively). The specificities (i.e. the percentage of unnecessary biopsies spared) were 22%, 21% and 31%, respectively (not significantly different). Conclusion Each immunoassay could be used to distinguish prostate cancer and BPH at different f/tPSA thresholds, with 21–30% of unnecessary biopsies spared. There was no difference in overall performance among the different assays. Further studies are needed to better define the exact use of the f/tPSA ratio in the routine diagnosis of prostate cancer.