Hopkins study confirms: PSA kinetics not useful risk indicator for early prostate cancer

May 26, 2010

Once PSA kinetics were thought to be a good way to predict which patients with early prostate cancer were at risk of progression. Now they're not. Even those who had the most hope for these biomarkers have evidence of their unreliability for this purpose.

How badly we need a way to know which early prostate cancers are truly cause for concern! Let us now pause to lament the loss of another promising biomarker for progression. A study from the Brady Urological Institute at Johns Hopkins, by some of the very authors who once promoted the velocity of changes in prostate-specific antigen (PSA) as a prognostic biomarker for prostate cancer, now undermines its validity for prostate cancer patients on active surveillance.

Among 290 such men, they report in the Journal of Clinical Oncology, neither PSA velocity (PSAV) nor PSA doubling time (PSADT) had suitable sensitivity or specificity to be used as a marker of progression. Nor was either significantly correlated with adverse pathology in those prostates that were removed.

Beginning in the 1990s, PSAV studies from the Brady Institute became a bellwether of controversy about the actual usefulness of PSA measures in general. After studies of men from the Baltimore Longitudinal Study of Aging showed an association between PSA velocity and the 25-year risk of death from prostate cancer, Hopkins urologist H. Ballentine Carter, MD, and others urged routine measure of changes in PSA as the most important way to predict the risk of progression to cancer and ultimate death for men whose PSA levels were too low to trigger biopsy.

"When you have a test that powerful," said Carter at the time, "you should use it."

Physicians did use it, and the question quickly arose whether too many men were being tested (and frightened) too early. The American Urological Association reportedly considered updating guidelines to incorporate PSAV screening, but its "best practices" document ultimately reviewed how to measure PSAV without outright advocating it. In 2007, a study appeared in the Journal of the National Cancer Instituteshowing that neither PSAV nor baseline PSA successfully predicted the lethality of prostate cancer, even 10 years after baseline.

Is the new study from the Brady Institute a U-turn, I asked Carter, or a return from a detour? "I interpret the findings as meaning that in men with prostate cancer being monitored, changes in PSA cannot always tell us when a patient has disease that is progressing and needs treatment," he responded. "Annual biopsies are necessary to monitor the situation."

The new report from Hopkins stands as something of a footnote to last year's systematic review in JCO from authors at Memorial Sloan-Kettering Cancer, which included 87 studies of PSAV and PSADT. Although these measures have been "widely advocated" as useful prognostic measures in early prostate cancer, the authors wrote, the review provides "no justification for the use of PSA dynamics in clinical decision making."

If nothing else, the Hopkins publication highlights the pressing need for good biomarkers to monitor early prostate cancer among men who choose active surveillance. "Modestly accurate models" involving patient age and baseline PSA have been developed, observe urologists Scott Eggener, MD, and Michael Large, MD, of the University of Chicago Department of Surgery. But these have yet to be validated in large studies.

"The way I look at it," says Eggener, "PSA increases should be repeated to confirm the value and even if it remains high should serve as a 'smoke alarm' leading to consideration of surveillance prostate biopsies. It does not necessarily mean there is a 'fire' (cancer progression)."