Fine-tuning the standard screening test for prostate cancer can help doctors reduce false alarms by up to 40 percent and perhaps find early-stage tumors that might otherwise be missed, researchers say.

About three-quarters of men with prostate cancer have elevated levels of a protein called prostate-specific antigen - PSA. Testing blood for PSA has been a standard screening method for older men since about 1990.

However, other conditions, such as non-cancerous prostate enlargement, can also make PSA levels shoot up, so doctors must perform biopsies to make sure cancer is truly the cause. For every biopsy that detects prostate cancer, two or three others are false alarms.

"The economic and psychological cost of this is huge," said Dr. Peter H. Gann of Northwestern University.

One way to try to reduce this problem is to make the PSA more precise. In the bloodstream, PSA can be either free-floating or attached to other proteins. Doctors have noticed that men with cancer often have lower levels of free PSA than do those whose elevated PSA has other causes.

Free PSA tests have been available for several years, but doctors are still unsure how to interpret the results. To help guide this, researchers conducted the first look-ahead study of the potential of free PSA testing to pinpoint cancer.

Doctors used blood samples stored in ysicians Health Study at Brigham and Women's Hospital in Boston. The landmark study established the benefits of taking aspirin to prevent heart attacks.

During 12 years of follow up, 430 men in the study were diagnosed with prostate cancer. By doing PSA tests on the stored samples, the researchers could predict the test's power to detect men's prostate cancer.

Gann and colleagues compared the cancer victims' PSA readings with those of 1,642 men the same age who stayed free of prostate cancer. He presented the analysis Sunday at a meeting in San Francisco of the American Association for Cancer Research.

If the total PSA level is over 10 nanograms per milliliter of blood, there is about a 50-50 chance the man has cancer. However, the risk of the cancer is much less certain if the PSA is between 4 and 10.

Gann said that looking for the percentage of free PSA should help doctors avoid negative biopsies in men whose PSA levels are in this gray area. He calculated how much cancer would be found if doctors limited biopsies to those whose free PSA fell below a specific cutoff.

Limiting biopsies to men whose free PSA is below 20 percent of the total, the number of unnecessary biopsies falls by 42 percent, although there is a slight increase in the number of missed cancers.

However, Gann said that by factoring free PSA into the equation, it might be worthwhile to do biopsies on men whose total PSA levels are above 3, rather than the current standard of 4. He said this would actually detect 10 percent more cancer and trigger 13 percent fewer needless biopsies than simply doing biopsies on everyone with a total PSA over 4.

Presumably, tumors in men whose PSAs are between 3 and 4 are likely to be small and still curable using surgery or radiation.

Dr. Ronald Morton Jr. of Baylor College of Medicine said the study helps establish a cutoff for free PSA levels that "lets us distinguish between clinically important and clinically non-important cancers."