Purpose: Since the implementation of widespread serum total prostate specific antigen based screening, the risk of prostate cancer over diagnosis has become a concern. We evaluated the amount of possible over and under diagnosis of prostate cancer in an asymptomatic screening population with a total prostate specific antigen of 2.0 to 3.9 (lower range) and 4.0 to 10.0 ng/ml (higher range).
Materials and methods: A total of 680 patients with prostate cancer were included. Possible over diagnosis was defined as Gleason score less than 7, pathological stage pT2a and negative surgical margins. Under diagnosis was defined as pathological stage pT3 or greater, or positive surgical margins. Furthermore, insignificant tumors according to the Epstein criteria were evaluated in a small subset of patients for whom cancer volume information was available.
Results: In the lower and higher total prostate specific antigen ranges there was an over diagnosis rate of 19.7% and 16.5%, and an under diagnosis rate of 18.9%* and 36.7%, respectively (p<0.05). In the prostate specific antigen range of 2.0 to 10.0 ng/ml combined the rates of over and under diagnosis were 17.6% and 30.3%, respectively. In addition, 8.7% of tumors with total prostate specific antigen 2.0 to 10.0 ng/ml met the Epstein criteria for insignificance.
Conclusions: These data show that the reported estimates of over diagnosis in the low total prostate specific antigen group are exaggerated in a screening population. Using our criteria prostate cancer under diagnosis occurs more frequently than over diagnosis in the total prostate specific antigen range of 4.0 to 10 ng/ml.