Background: There is a large amount of confusion in interpreting prostate specific antigen (PSA) values for prostate cancer. More precise risk assessments for prostate cancer detection are needed for men faced with an abnormal PSA.
Methods: We studied a sample of 2,637 men who underwent a prostate biopsy for an abnormal digital rectal exam (DRE) or PSA. Using factors including age, ethnicity, family history of prostate cancer, previous negative biopsy, presence of voiding symptoms, prostate volume, DRE and PSA, we constructed nomograms to predict the probability of prostate cancer at biopsy.
Results: Of the 2,637 men, 1,282 men (48.6%) had prostate cancer detected. Age, ethnicity, family history of prostate cancer, a previous negative biopsy, prostate volume, DRE and PSA were all significant predictors of prostate cancer. Nomograms were constructed based on these factors to predict the risk of prostate cancer and of aggressive prostate cancer (defined as a Gleason Score 7 or more). The positive predictive value varied from 5% to 95% based on the nomograms. The nomograms were validated using bootstrapping methods and the expected and observed proportions were found to be highly concordant.
Conclusions: For men with an abnormal PSA or DRE, the risk for prostate cancer can be accurately estimated using a nomogram based on age, ethnicity, family history of prostate cancer, previous negative biopsy, presence of voiding symptoms, prostate volume, DRE and PSA. This tool will aid physicians and patients in determining the need for prostate biopsy.