Too Early in PSA Era to See Drop in Prostate Cancer Mortality

BOSTON—Men are far more likely to be diagnosed with prostate cancer today than 2 decades ago, and they have more aggressive treatment options. Is their chance of dying of the disease dropping as a result? It is too early in the era of prostate-specific antigen (PSA) testing to tell, keynote speaker Michael J. Barry, MD, said at the 42^nd Annual Meeting of the American Society for Therapeutic Radiology and Oncology .

The incidence of prostate cancer has increased substantially, and population-based mortality has dropped in the United States, said Dr. Barry, chief of the General Medicine Unit, Massachusetts General Hospital, and associate professor of medicine, Harvard Medical School.

Nonetheless, he cautioned that inconsistencies in everything from testing procedures to treatment choices make comparison of outcomes from observational studies impossible at this time. Only long-term controlled trials can provide conclusive answers, he said: "I think we’ll see a difference, but we need to prove it."

Dr. Barry cited unpublished results of an ongoing comparison of prostate cancer outcomes between SEER areas in Seattle and Connecticut. These two SEER areas had identical population-based prostate cancer mortality rates in the pre-PSA era.

Seattle adopted PSA detection and aggressive treatment much earlier and faster than Connecticut, according to Dr. Barry. His research group, therefore, developed a cohort study that followed 100,000 Medicare-aged men in each area from 1987 on.

From 1987 to 1990, men in Seattle were twice as likely to be diagnosed with prostate cancer and six times as likely to have a radical prostatectomy as men in Connecticut, he said. The Seattle group also received more radiation therapy.

With all that extra screening and treatment, Dr. Barry expected to see an earlier effect on prostate cancer mortality in Seattle than in Connecticut. "At least for 11 years, we haven’t seen it," he said. "The age-adjusted prostate cancer mortality, the rate ratio, for both areas is about 1.06—actually with Seattle being a little higher, but not significantly so—through 11 years of follow-up."

Austrian Study

Dr. Barry contrasted these results with a preliminary report from the state of Tyrol in Austria. An aggressive program of PSA screening and cancer treatment in Tyrol has been credited with a 42% reduction in population-based mortality in 1998 vs a base period of 1986 to 1990. The state has 65,000 men, ages 45 to 75. It began its PSA testing program in 1993, testing one third of the population the first year, and two thirds by 1996.

How could early detection produce such a large benefit if the lead time for prostate cancer diagnosis in the PSA era is 8 years? he asked. "I don’t fully understand these results, and I’m waiting for the paper to be published," he said.

Four randomized trials in Europe and the United States are collecting data on prostate cancer mortality, but Dr. Barry predicted that "because of the slow progression of cancers being diagnosed now, it will be years before we get results."

To dramatize the problems in comparing pre- and post-PSA outcomes, he presented a 10-year nonrandomized comparison of outcomes from the pre-PSA era, based on cases reported to SEER.

Cancer-specific survival for men with moderately differentiated cancer was about 87% for radical prostatectomy, 76% for radiation, and 77% for conservative therapy. "Do I think these figures represent a true comparison?" he asked. "Absolutely not. We can’t make the comparisons because of case selection."

He pointed to huge differences in overall survival rates in these patients—71% for radical prostatectomy, 48% for radiation, and 38% for conservative therapy. "The healthier patients are getting radical prostatectomy. The less healthy patients are getting radiation. The least healthy patients are getting conservative therapy," he said, citing differences in age, comorbidity, and other factors.

While praising several recent studies on treatment outcomes, Dr. Barry concluded that post-PSA-era patients have not been followed long enough to identify important predictors of outcomes. "It’s encouraging to see good outcomes over 5 or 6 years, but, remember, in the pre-PSA era, many of those patients wouldn’t have been diagnosed yet."