PSA kinetics: Not a reliable signal to intervene

September 1, 2009
Mac Overmyer

PSA velocity and PSA doubling time, once considered promising tools for identifying the need for intervention in prostate cancer patients under active surveillance, now appear to have failed to meet that promise.

First author Ashley Ross, MD, PhD, a urology resident working with Patrick C. Walsh, MD, and colleagues, told Urology Times, "The theory is that PSA screening can lead to over-diagnosis and that active surveillance is a good way to reduce over-treatment in these men, but right now, we don't yet know the best way to follow them.

Hopkins urologists formerly recommended 12-core biopsies, but recent studies have shown that cancer can be under-sampled in the transition zone.

"In light of this, the men in our active surveillance program are now getting 14-core biopsies, to include two cores from that region," Dr. Ross said.

The researchers derived their conclusions from evaluations of 290 men in an active surveillance program. Entry criteria for the program are a PSA density of <0.15, Gleason score of ≤6 with no pattern >3, and two or fewer cores with cancer, neither of which can display more than 50% cancer involvement. These men were followed with twice-yearly PSA measurements and annual biopsies. Progression was defined as a Gleason score ≥7, more than two cores with cancer, or >50% cancer involvement of any core.

PSAV and PSADT were calculated from a patient's complete PSA history after diagnosis. At median overall follow-up of 3.6 years, 188 (65%) remained on active surveillance and 102 (35%) met progression criteria.

Receiver operating characteristic (ROC) curves were constructed and the data examined against a spread of PSAV and PSADT cut-points. Regardless of the cut-point used, "overall, PSAV and PSADT were not significantly associated with progression by biopsy or favorable pathology at the time of later prostatectomy, and no PSAV or PSADT cut-point had both high sensitivity and specificity," said Dr. Ross.

For instance, only 40% of men who progressed on biopsy had a PSAV >0.75 ng/mL/year, and 79% of men who did not progress had a PSAV <0.75 ng/mL/year. Using PSADT, 49.5% of those who progressed had a PSADT <3 years, and 51% of men who did not progress had a PSADT >3 years. Results were similar in men with a PSA of <4.0 ng/mL at diagnosis, where PSA kinetics might be thought to be more predictive.

Using these PSA kinetic values as cutoff points to indicate need for intervention "would mean that there would be intervention in 50% of patients not at risk of progressing and that 50% of patients requiring intervention might not get it. It would be like flipping a coin," Dr. Ross explained.

The study team also examined the association between PSA kinetics and final post-prostatectomy pathology in those patients who progressed on surveillance, then underwent surgery. Again, no cut-point could be found that would provide sufficiently high sensitivity and specificity (area under ROC curves of 0.56 and 0.51, respectively) to guide progression to treatment.

Although PSA velocity is important in the initial screening for prostate cancer and in predicting the prognosis prior to treatment, "for men with prostate cancer undergoing active surveillance, PSA kinetics were not sufficiently reliable to identify progression," the researchers concluded. "There was considerable overlap between the groups, and we were unable to identify a cut-point that could accurately distinguish progressors from non-progressors."

Dr. Ross said researchers are continuing to accumulate data. One avenue is to look for additional markers that might be useful in men who are on active surveillance.