ABSTRACT
INTRODUCTION: Numerous prostate cancer predictive tools have been developed to help with decision-making in men needing prostate biopsy. However, they have been modelled and validated almost exclusively in Caucasian cohorts, hence limiting their use in other population groups. The aim of this study was to assess the validity of the ERSPC risk calculator in a South African cohort. METHODS: Patients who have had a transrectal ultrasound (TRUS)-guided prostate biopsy at Groote Schuur Hospital from January 2008 to August 2017 were reviewed. Predictor variables were entered into the ERSPC risk calculator and results were compared with prostate biopsy pathology results. Predictive accuracy of the ERSPC risk calculator for these patients was derived using receiver operator characteristics (ROC) Area under the curve and is expressed as a percentage. RESULTS: 516 prostate biopsy sessions in 475 different men were analysed. The predictive accuracy of the ERSPC risk calculator was better than a PSA/DRE strategy for the presence of cancer-0.738 (95% CI 0.695-0.781) vs 0.686 (95% CI 0.639-0.732), and for significant PCa-0.833 (95% CI 0.789-0.876) vs 0.793 (95% CI 0.741-0.846). This translated into 50 less biopsies when compared to a PSA > 4/abnormal DRE strategy. Use of the ERSPC RC would have missed eight non-significant cancers [Significant cancer being defined as having a tumour stage T2b (> 1/2 lobe involved with prostate cancer) and/or a Gleason Score equal to or greater than 7]. CONCLUSION: Our results confirm the validity of the ERSPC RC in a South African cohort. Application of this calculator to the wider South African population would allow better selection of patients for prostate biopsy and spare a significant number its adverse consequences.
