PSA kinetics did not enhance predictability of outcome in prostate cancer

PSA velocity and PSA doubling time were not superior to PSA alone in predicting patient outcome.

According to data from a recent analysis, PSA dynamics, specifically PSA velocity and doubling time, did not improve the predictability of patient outcomes among men undergoing radical prostatectomy. However, half of the pretreatment dynamics were associated with outcome in a univariate analysis.

“These improvements were small with wide CIs,” the researchers wrote of their findings. “We believe these improvements could have been produced by chance, given the large number of definitions tested.”

Using data gathered in their database, researchers from Memorial Sloan-Kettering Cancer Center identified 2,938 patients with two or more PSA values before radical prostatectomy. Three hundred eighty-four patients experienced biochemical recurrence and 63 experienced metastases. The median follow-up for patients without biochemical recurrence was 2.1 years. The purpose of the study was to assess whether pretreatmeant PSA velocity and doubling time predicted outcome.

The researchers used 11 definitions of PSA doubling time and seven definitions of PSA velocity; PSA velocity was analyzed as both a continuous variable and categorized as >2.0 ng/mL per year based on data from D’Amico et al.

PSA alone was associated with recurrence (P<.001) and metastases (P=.002). In a univariate analysis, 11 PSA definitions were associated with biochemical recurrence or metastasis, though a longer PSA doubling time predicted shorter survival in one definition. Two PSA doubling time and four PSA velocity definitions were associated with both biochemical recurrence and metastasis (P<.05).

Compared with PSA alone, one PSA doubling time and one PSA velocity definition had a higher predictive accuracy for biochemical recurrence; four PSA velocity definitions improved prediction of metastasis.

“However, the improvements in predictive accuracy were small, associated with wide CIs and markedly reduced if additional predictors of stage and grade were included alongside PSA,” the researchers wrote of the findings.

In addition, the researchers reported that the two PSA doubling time definitions and four PSA velocity definitions associated with biochemical recurrence and metastasis could have occurred by chance.

In an accompanying editorial, Anthony V. D’Amico, MD, PhD, chair, division of genitourinary radiation oncology, Brigham and Women’s Hospital and Dana-Farber Cancer Institute and professor of radiation oncology, Harvard Medical School, and Ming-Hui Chen, PhD, department of statistics, University of Connecticut, discussed the findings from O’Brien et al.

“This analysis is accurate and comes to a sound conclusion when considering all men who present with localized prostate cancer,” D’Amico told HemOnc Today. “However, the conclusion of their study can be misleading when one considers the subgroup of men with low-risk prostate cancer where PSA kinetic measures have been shown to be clinically important in determining outcomes such as cancer-specific and all-cause mortality.”

Though O’Brien and colleagues found that PSA kinetics did not add to the predictive accuracy of pretreatment PSA alone, D’Amico noted the importance of PSA kinetics.

“Given that in the United States the vast majority of people present with low-risk prostate cancer due to PSA screening, PSA kinetic measures maintain an important clinical role,” he said. – by Stacey L. Adams

D'Amico AV. J Clin Oncol. 2009;doi:10.1200/JCO2009.22.6068.

O'Brien MF. J Clin Oncol. 2009;doi:10.1200/JCO.2008.19.9794.