Effect of intra‐observer variation in prostate volume measurement on prostate‐specific antigen density calculations among prostate cancer active surveillance participants

Abstract

What's known on the subject? and What does the study add? The Epstein criteria, which utilize prostate specific antigen density (PSAD) benchmarks, are recognized to be a reasonable method of selecting men for active surveillance of prostate cancer. Transrectal ultrasonography, however, may not be a sufficiently precise method of measuring prostate volume for the determination of PSAD. This study shows that despite impressive intra-observer variability in transrectal ultrasonography guided prostate volume measurements, this variability typically does not affect the PSAD to an extent by which qualification for active surveillance would be altered. • To determine intra-observer variability in transrectal ultrasonography (TRUS) guided prostate volume measurements in the Johns Hopkins active surveillance group and to establish whether or not this variability could affect prostate-specific antigen density (PSAD) estimates in this cohort. • In all, 253 patients with a combined total of 1111 prostate biopsies underwent TRUS-guided prostate volume measurements performed by the same physician at least three times over the course of their care. • Coefficients of variation (CV) were calculated for each set of measurements performed on each patient by the same physician, and average CVs were determined for each physician and for physicians overall. The CVs were correlated with the average of each patient's measured prostate volumes to look for any trend. • Finally, measured prostate volumes were used with each patient's initial prostate-specific antigen (PSA) value to calculate PSAD to reveal whether or not the degree of variability found in these measurements would have led to PSADs that would have otherwise precluded qualification for active surveillance. • The average CV for all sets of prostate volume data was 0.168. Average CVs for each physician ranged from 0.136 to 0.234. • However, actual CVs ranged anywhere from 0.013 to 0.549. The CVs were found to have no correlation with prostate volumes (Pearson correlation coefficient: 0.04). • In 95% of cases, variability in TRUS-guided prostate volume measurement did not affect PSAD sufficiently to elicit a value greater than 0.15. • Even among individuals who are highly experienced in TRUS-guided prostate volume measurement, significant intra-observer variation exists. However, this variability is not enough to affect one's eligibility for prostate cancer active surveillance when PSAD criteria are used. • The TRUS-guided prostate volume measurements remain a reliable method of assessing PSAD in patients with prostate cancer.