Comparison of methods for calculating prostate specific antigen velocity.

PURPOSE: Prostate specific antigen velocity is frequently calculated using regression analysis of multiple prostate specific antigen measurements during an interval of 18 to 24 months. It has been reported that the prostate specific antigen velocity in the year before prostate cancer diagnosis is associated with the cancer specific mortality rate following radical prostatectomy and radiation therapy. There are limited data comparing alternate methods of calculating prostate specific antigen velocity. In this study we compared simple arithmetic prostate specific antigen velocity calculations to the more complicated regression analysis using prostate specific antigen measurements from varying intervals to determine whether the methods were interchangeable.
MATERIALS AND METHODS: From 2003 to 2005 a total of 540 men underwent radical retropubic prostatectomy for localized prostate cancer. Preoperative prostate specific antigen velocity was calculated using arithmetic and linear regression methods during 12 and 18-month intervals. The correlation, mean and median prostate specific antigen velocity were compared among methods.
RESULTS: Prostate specific antigen velocity calculations using simple arithmetic and linear regression had a strong correlation when restricted to prostate specific antigen values from 12 months before diagnosis (r = 0.92). When prostate specific antigen measurements beyond 1 year were included in the calculation, prostate specific antigen velocity was significantly lower than in the 12 months before cancer diagnosis.
CONCLUSIONS: Using a simple arithmetic prostate specific antigen velocity calculation yields results comparable to those of the more complicated regression analysis when restricted to prostate specific antigen values from within the year before diagnosis in patients with clinically localized prostate cancer. In this manner arithmetic prostate specific antigen velocity calculations may be used in daily clinical practice to help predict the outcome following definitive therapy. Prostate specific antigen velocity is significantly lower when calculated during a longer interval before prostate cancer diagnosis.