Response: Re: An Empirical Evaluation of Guidelines on Prostate-Specific Antigen Velocity in Prostate Cancer Detection

Extract

Loeb et al. make three criticisms of our article. First, they claim that the age distribution of our sample limits the applicability of the findings, specifically, that the mean age was too high for prostate-specific antigen (PSA) screening. Accordingly, we repeated our analyses restricting the dataset to patients aged less than 65 years at biopsy. Although as Loeb et al. point out, this constitutes only 20% of our cohort, it is nonetheless a large sample (n = 1181). The results were very similar to the main analysis. For example, the area under the curve for log PSA was 0.698, which increased only to 0.699 for a model including both log PSA and PSA velocity. Similarly, for men with PSA less than 4 ng/mL and negative digital rectal exam (DRE), the PSA cut point of 2.5 ng/mL would lead to a better ratio of cancers detected and biopsies performed (27 cancers, 105 biopsies) than the PSA velocity cut point of 0.35 ng/mL/year (29 cancers, 123 biopsies).