A screening test for prostate cancer taken by millions of men every year is not terribly accurate, and even the best result does not ensure that a man is cancer-free, experts said yesterday.

Researchers found that the standard prostate-specific antigen, or PSA, test produces many false positives and false negatives -- meaning some men who think they do not have cancer actually do, while others may undergo uncomfortable biopsies only to learn there is no tumor after all.

"Patients have assumed, 'My PSA is below 4. It's normal. I have no risk,' " lead author Ian Thompson of the University of Texas Health Science Center at San Antonio said at a news conference.

Usually, a biopsy is not recommended unless the PSA level exceeds 4 nanograms per milliliter.

Some men with PSA levels of 1 had prostate cancer, the study showed. Others with higher PSA levels did not have prostate cancer.

If all men got biopsies when the PSA reached 1.1, more than 80 percent of all prostate cancers would be detected, Thompson said.

Men whose fathers or brothers had prostate cancer and black men are among those who have a higher risk, he said.

"PSA, like blood pressure, like cholesterol, like many other tests, shouldn't be considered to be 'normal' or 'abnormal' but should be considered as showing a range of risk," Thompson said.

For their study, published in the Journal of the American Medical Association, Thompson and colleagues examined more than 5,000 men who took part in a larger study that showed the drug finasteride, sold by Merck & Co. under the brand name Proscar, could help prevent prostate cancer in some men.

All the men got a placebo and were watched for seven years. All of them got biopsies, regardless of PSA level.

Nearly 22 percent of the 5,000 men were found to have prostate cancer.

Prostate cancer affects more than 200,000 men a year in the United States and will kill 29,000 in 2005, according to the American Cancer Society.