Sidebar

Seeking answers from complexed PSA

Back to Cover Story

April 2000

The value of the assay for prostate-specific antigen has been increased by adding measurement of percent free PSA, but its specificity needs to be improved.

Many derivatives of PSA have been tried, including age-related reference ranges, PSA density, and PSA velocity. "In our hands, none of these derivative tests have been found to be very useful," says Michael Brawer, MD, director, Northwest Prostate Institute, Northwest Hospital, Seattle.

Dr. Brawer has taken a different approach—measuring complexed PSA (cPSA), the fraction of circulating PSA bound to serum protease inhibitors. In men with prostate cancer, a higher proportion of PSA occurs in the complexed form, mostly bound to alpha-1 antichymotrypsin. Measurement of percent free PSA (%fPSA) gives an indirect approximation of the fraction of cPSA. But Dr. Brawer sees problems with %fPSA assays. For one, different assays give different cutoffs for triggering biopsies. Also, determining %fPSA requires doing two assays—one for total PSA (tPSA) and one for free PSA. This effectively doubles the cost of testing, Dr. Brawer says.

"What we really want to know is the amount of complexed PSA," he continues. "It has been very difficult to develop a method to measure that directly." Dr. Brawer used such a method (now in development by Bayer Diagnostics), based on an antibody that binds fPSA and renders it nonreactive, to study, retrospectively, archived serum from 300 patients. "Complexed PSA was better than free/total PSA for specificity," he says. "Now, with one measurement-the complexed form of PSA—we get all the information that we get with free/total but at half the cost of PSA testing and with no quotient bias or manufacturer’s assay variability."

But Dr. Brawer’s assertion is not generally accepted. A group led by Klaus Jung, MD, professor, Department of Urology, Charite Hospital of Humboldt University, Berlin, generated contradictory data. Subjects had either prostate cancer or benign prostatic disease and tPSA concentrations between 2 and 10 ng/mL.

"We compared all the possibilities," Dr. Jung told CAP TODAY. "We measured total PSA, cPSA, and the ratios fPSA/tPSA and cPSA/tPSA. We found that there is no possibility to make a good differentiation [between prostate cancer and benign prostatic disease] if cPSA alone is measured. But we can use the ratio of cPSA to tPSA, which is equivalent to the established ratio fPSA to tPSA." When cutoffs for the three parameters were chosen to yield a sensitivity of 90 percent, specificities were: cPSA alone, 25 percent; cPSA/tPSA, 57 percent; fPSA/tPSA, 55 percent (Clin Chem. 2000;46(1):55-62).

Dr. Jung also cites a study recently published in Prostate (2000;42:181-185) by Filella, et al, that arrives at the same conclusion.

Moreover, he points out, Dr. Brawer’s study did not verify that fPSA/tPSA was better than tPSA, "so the overall result is suspect."

Dr. Brawer responds that %fPSA was not better than total PSA in his study due to an artifact—fPSA degrades with storage. Better results were achieved in a more recent study, he says. And he counters that Dr. Jung’s paper used a subset analysis of 40 patients with cancer and 40 without who were selected to have the same average total PSA. "I’m convinced that when there is no difference in total PSA in your population, complexed PSA isn’t going to work," Dr. Brawer says.

"My hunch is that complexed PSA is going to be substantially equivalent to percent free PSA in specificity enhancement but at different prices and with no assay bias," he states.

Daniel Chan, PhD, professor of pathology, urology, oncology, and radiology, Johns Hopkins University, and director of the Clinical Chemistry Division, Johns Hopkins Hospital, calls Dr. Brawer’s conclusion "premature." "Measuring cPSA alone is not generally accepted by leading urologists," he notes. "In the studies that I did, percent free PSA and percent complexed PSA gave about the same performance. If I just use complexed PSA without a percentage, its performance is not as good."

An ongoing prospective multicenter study with the Bayer cPSA assay may shed more light on this issue.

-William Check, PhD