A sharp rise in the score on a test commonly used to screen for prostate cancer appears to offer a powerful way to identify men with the most dangerous tumors, researchers reported yesterday.

A study involving more than 1,000 men with prostate cancer found that those who had experienced a big jump in the results from their PSA tests in the previous year were at much greater risk of dying from the malignancy, the researchers found.

The findings may help solve one of the biggest dilemmas facing men with prostate cancer: whether to have it treated or to wait to see if it starts to cause problems, researchers said.

"This is something that is really urgently needed -- we need a way of judging the danger of a prostate cancer," said William J. Catalona, a prostate cancer specialist at Northwestern University who led the study. "This is the most powerful predictor we have."

The findings suggest the test should be used like a mammogram: Men should have a base-line test at age 35 or 40 and undergo the test annually by age 45 or 50 so doctors can monitor changes over time, Catalona and his colleagues said.

"If they measure it every year and it starts to take off on them, they have a very strong indication of how aggressive a cancer is," he said. "This is going to be a very valuable new tool."

More than 230,000 men receive prostate cancer diagnoses each year in the United States, making it the second leading type of cancer to strike males. Many prostate cancers are first detected during a PSA test, which measures a protein produced by prostate cells called the prostate-specific antigen.

But the discovery of prostate cancer often throws men into a quandary. Many prostate cancers never cause problems, and surgery to treat them frequently leaves men impotent or incontinent. At the same time, prostate cancer kills nearly 30,000 American men each year, making it the second leading cause of cancer death among men. Unless a man has a family history of the disease or his cancer is clearly advanced, he has no way of knowing how much risk he faces.

In the new study, Catalona and his colleagues analyzed data from 1,095 men who had undergone surgery for prostate cancer. Those whose PSA levels had risen rapidly in the year before the cancer was diagnosed were much more likely to die from their disease, the researchers reported in today's New England Journal of Medicine. Men whose PSA scores had jumped by more than two points were 10 times as likely to die as those whose scores increased by less than that, the researchers found.

"It's the most important risk factor," said Anthony D'Amico of the Brigham and Women's Hospital in Boston, the lead author of the study.

A rapidly rising "PSA velocity" predicted a bad outcome even if the score never exceeded what is currently considered a normal level, the researchers found.

The findings come after another study that found that many men had prostate cancer even though their PSA tests fell within the normal range. The new findings indicate the rate of change in the PSA level is more important than the result of a single reading.

Previous research had suggested that might be the case, but the new research is the most definitive to date and the first to show a clear link to survival.

D'Amico likened PSA tests to mammograms, which women typically start having at age 40.

"We hope physicians translate these results into taking a base-line PSA level on their patients at age 35, and then monitoring PSA level for men through their forties to determine risk and appropriate course of treatment if the cancer does present itself," D'Amico said.

Other experts agreed, although more research is probably needed to validate the findings.

"I think it makes sense," said Martin I. Resnick, a urologist at the Case Western Reserve School of Medicine, speaking on behalf of the American Urological Association. "The idea of a base-line PSA may be a good idea."

Currently, men are urged to start getting annual PSA tests at age 45 or 50. The exact age for a base-line test is likely to vary, with men who are at increased risk, such as those with a family history and African Americans, getting it earlier.

In addition, the findings provide a way to identify those who are the best candidates for immediate treatment.

"This is an important study," said Howard L. Parnes of the National Cancer Institute. "These data suggest that the rate of rise of PSA can be useful in helping men make the decision between watchful waiting and aggressive therapy for prostate cancer."

Parnes and other researchers warned, however, that this strategy would not necessarily guarantee that men who opt for watchful waiting will never need treatment. The test can only identify men who clearly should not wait, they said.

Parnes also questioned the necessity of having earlier base-line PSA tests, saying the current study showed the results in just the year before diagnosis were sufficient.

Parnes and the researchers agreed that the findings indicate more research is needed to evaluate chemotherapy and other therapies for prostate cancer, because nearly one-third of those who received treatment in the study died of the disease.