A comparative performance analysis of total prostate-specific antigen, percentage free prostate-specific antigen, prostate-specific antigen velocity and urinary prostate cancer gene 3 in the first, second and third repeat prostate biopsy
Auprich, M.; Augustin, H.; Budäus, L.; Kluth, L.; Mannweiler, S.; Shariat, S.F.; Fisch, M.; Graefen, M.; Pummer, K.; Chun, F.K.-H.
Bju International 109(11): 1627-1635
Study Type - Diagnosis (exploratory cohort) Level of Evidence 2b What's known on the subject? and What does the study add? Risk factor assessment in the repeat biopsy setting is affected by a decreasing diagnostic accuracy of each single risk factor (e.g. DRE, tPSA, %fPSA, complexed PSA, PSA density or PSAV] with increasing number of prostate biopsy sessions. PCA3 shows impressive diagnostic performance in the initial and early repeat biopsy settings. In a head-to-head comparison we demonstrate the concept that the number of previous repeat biopsy session strongly influences performance characteristics of biopsy risk factors, including PCA3. While the novel diagnostic marker would have avoided a considerable number of unnecessary biopsies in the first repeat biopsy scenario, its effects dissipated at second and ≥ third repeat biopsies. To compare the performance characteristics of prostate cancer risk factors such as total prostate-specific antigen (tPSA), percentage free PSA (%fPSA), PSA velocity (PSAV) and urinary prostate cancer gene 3 (PCA3) at first, second and ≥ third repeat biopsy session. Patients (n= 127) aged ≤70 years, with suspicious digital rectal examination (DRE) and/or persistently elevated age-specific total PSA levels (2.5-6.5 ng/mL) and/or suspicious prior histology (atypical small acinar proliferations [ASAPs]≥ two cores affected by high-grade prostatic intra-epithelial neoplasia [HGPIN]) undergoing either a first, second, or ≥ third repeat biopsy were investigated using a 12- or 24-core biopsy scheme. PSAV (≥ three values collected over ≥12 months) was calculated using the log-slope method. PCA3 scores were assessed using the Progensa assay®. After stratification according to the number of previous biopsies (first, second and ≥ third), calculation of specificity, positive and negative predictive values (PPV, NPV) and the proportion of avoided unnecessary repeat biopsies (PAB) compared with tPSA at fixed sensitivity thresholds (75, 85 and 95%) were performed. Finally, accuracy estimates (area under the curve [AUC]) were quantified for each repeat biopsy scenario. At repeat biopsy, overall prostate cancer (PCa) detection was 34.6%. At first repeat biopsy, PCA3 predicted PCa best (AUC = 0.80) and would have avoided 72.2% of repeat biopsies (75% sensitivity) compared with tPSA. At second repeat biopsy, %fPSA demonstrated the highest accuracy (AUC = 0.82) and would have avoided 66.7% of repeat biopsies (75% sensitivity) compared with tPSA. At ≥ third repeat biopsy, again %fPSA demonstrated the highest accuracy (AUC = 0.70) and would have avoided 45.0% of repeat biopsies (75% sensitivity) compared with tPSA. The main limitation of our study resides in its small sample size. The findings of the present study promote the concept that the number of previous repeat biopsy sessions strongly influences the performance characteristics of biopsy risk factors. Total PSA was no significant risk factor in the entire analysis. By contrast, %fPSA performed best at second and ≥ third repeat biopsy. PSAV's diagnostic potential was reserved to patients at second and ≥ third repeat biopsy. Finally, PCA3 demonstrated the highest diagnostic accuracy and potential to reduce unnecessary biopsies at first repeat biopsy. However, this advantage dissipated at second and ≥ third repeat biopsy.