Objective: The aim of this study was to test the ability of prostate cancer antigen-3 (PCA3) and Hansen's PCA3-based nomogram to predict prostate cancer (PCa) probability in a Norwegian cohort, with the goal of reducing unnecessary biopsies.
Material and methods: Altogether, 127 consecutive patients were recruited to this study at Haukeland University Hospital, Norway. Prostate-specific antigen (PSA), PCA3 score, digital rectal examination (DRE), prostate volume (Pvol) and age were determined. All patients had an extended 10-core biopsy. The performance of PCA3 score and Hansen's nomogram was tested.
Results: There were 124 evaluable patients. Among these, 59 patients had PCa on the initial biopsies. Mean PSA, PCA3 score and age were significantly higher and Pvol was significantly lower in patients with PCa. PCA3 scores of 35 and 21 led to a sensitivity of 71% and 81% and specificity of 72% and 55%, respectively. Hansen's nomogram gave an area under the curve (AUC) of 0.806. The intraclass correlation was 0.959 (Cronbach's alpha). Applied to this material, PCa would be missed in 15.2% of patients when applying the suggested threshold probability of 30%, among whom 66.7% had high-grade PCa. With a threshold probability of 20% only one patient had PCa and this was low grade.
Conclusions: Hansen's PCA3-based nomogram is valid for this cohort. A threshold probability of 20% seems more adequate than 30% for this less screened cohort. PCA3 score only affects the biopsy indication in some patients and is recommended only for this subset. The results need to be confirmed in a larger study.
Keywords: PCA3; initial biopsy; nomogram; prostate biopsy; prostate cancer; prostate cancer gene 3.