ORLANDO, FL—Two factors have been found to predict the time to first bone metastasis in men with prostate cancer undergoing anti-androgen treatment—baseline prostate-specific antigen (PSA) levels above 10 ng/mL and a rapidly rising PSA.
A study of the natural history of prostate cancer presented here at the 2005 Multidisciplinary Prostate Cancer Symposium found the two variables to be independently more predictive of metastasis than Gleason score, a history of bilateral orchiectomy, regional lymph node metastases at diagnosis, or prior prostatectomy.
PSA baseline and velocity may also predict overall survival and metastasis-free survival, said first author Matthew R. Smith, MD, PhD, of Massachusetts General Hospital Cancer Center and Assistant Professor of Medicine at Harvard Medical School.
“High baseline PSA and the rate of PSA rise appear to provide a reliable way to help physicians and patients estimate risk and the need for additional treatment.”
Dr. Smith and his colleagues investigated the relationship between predictive factors and prostate cancer progression in 201 men with nonmetastatic disease with rising PSA levels despite hormonal therapy. The mean age of the men was 73, and a median of approximately seven years had passed between diagnosis and study entry.
A total of 29% of patients had Gleason scores of 8 to 10, 17% had node-positive disease, and approximately 33% had had a prostatectomy. The median PSA at study entry was 13.8 ng/ml. The median PSA doubling time was 9.7 months, which Dr. Smith said was approximately the same as observed in large studies of rising PSA after prostatectomy and prior to hormone therapy.
“That was somewhat of a surprise, since we would have thought that the PSA doubling time would be much faster after failing hormone therapy, but it turned out not to be the case in this study,” Dr. Smith said.
The median bone metastasis-free survival was 30 months, but median time to first bone metastases and overall survival were not reached.
“At two years only one third of the patients had developed bone metastases, he said. “I say ‘only’ one third because at the time the study was designed, a panel of experts felt it would be about 80 percent at two years. This really tells us that this disease state, although previously felt to be very aggressive, turned out to be not quite as aggressive as appreciated.”
Breaking the cohort into tertiles, Dr. Smith said that at study entry one third of patients had PSA levels of less than 7.7 ng/mL, and one third had the highest levels at greater than 24 ng/mL.
“And metastasis-free survival separated out quite strikingly by baseline PSA tertile as well,” he said. “For example, at two years, only about 25 percent of men with low PSAs had met the primary study outcome of metastasis or death, versus 77 percent of patients with high PSAs.”
PSA doubling time also sorted out neatly by tertiles, he said. Among slow PSA risers, with doubling times longer than 18 months, 80% remained alive and metastasis free at two years, compared with only about 25% alive and metastasis free in the group with doubling times shorter than six months.
“From this body of data we conclude that castrate, nonmetastatic prostate cancer has a relatively indolent natural history, that PSA and PSA velocity independently predict time to first bone metastasis, bone metastasis-free survival, and overall survival,” Dr. Smith said.
New Patient Category
The situation of having progressive castrate nonmetastatic prostate cancer was virtually unknown 10 years ago, Dr. Smith noted, but such patients now represent an increasingly large population because of a large lead-time bias with prostate cancer screening, routine use of PSA surveillance to identify recurrences prior to onset of metastases, and standard use of hormone therapy at the time of biochemical failure.
“More than half a million American men per year are treated with hormone therapy, and the vast majority do not have metastatic disease,” he said. “Hormone therapy works, but the disease inevitably progresses.”
Remarkably little was known about the natural history of that disease state, he said, but a poor prognosis was inferred from castrate metastatic disease. In two recent large randomized trials of docetaxel, for example, the median survival was 16 to 18 months.
“It was inferred that a similar poor outcome would be seen in men with nonmetastatic progressive hormone-refractory disease,” he said.
Original Trial Aborted
The men in Dr. Smith's study had been enrolled in the placebo arm of an aborted, randomized, controlled trial, Zometa 704, designed to test the ability of zoledronic acid to prevent metastasis. Subjects had rising PSA levels during androgen deprivation but negative bone scans.
The trial was originally designed to include 1,000 patients randomized to standard care plus zoledronic acid or to standard care plus placebo.
Although the study was aborted because the event rate was lower than expected, the men had undergone frequent bone scans to ensure that time to metastasis would be unbiased to PSA.
“In routine clinical practice what drives testing is often a rising PSA,” Dr. Smith explained. “In this case, since the assessments were done every four months, we felt there would be less PSA bias towards that clinical outcome.”
Perhaps Not Indolent
A member of the symposium program committee, in a press conference highlighting several papers presented at the meeting, disagreed with Dr. Smith's use of the term indolent.
“The study tells us, surprisingly, that people who we thought would do very, very poorly—that is, develop metastases in a relatively short period of time—actually have a better prognosis than we had previously anticipated,” said Philip Kantoff, MD, Professor of Medicine at Harvard Medical School and Director of the Lank Center for Genitourinary Oncology in Boston.
“Dr. Smith uses the word ‘indolent’—I don't know if I would go as far as to say that, but their median time to development of metastases is a lot longer, a more protracted natural history for these individuals than we anticipated.”
Dr. Kantoff also noted that individuals who enter clinical trials often do better than those who don't, even the placebo-group patients. In this study the placebo arm was the study arm, and it is possible that a good natural history in some cases was due to patient selection for entry into the clinical trial, he said. “But I don't think that explains the whole story here. I think the larger story is that individuals with nonmetastatic castrate prostate cancer do better than we anticipated.”
Rising PSA May Help Limit Size of Trials
The finding that baseline PSA levels and a rapidly rising PSA independently predict time to metastasis and overall survival in men with recurrent prostate cancer may help researchers design smaller yet clinically significant trials for these patients, noted Matthew R. Smith, MD, PhD, of Harvard Medical School.
By limiting enrollment to men with those risk factors, a randomized placebo-controlled trial could have as few as 1,000 subjects and still produce meaningful results on metastasis prevention, he said.
Prostate cancer is a relatively slow growing disease, and few other disease states allow researchers to study the prevention of metastasis when disease recurs. But that natural history has also meant that the studies need very large cohorts in order to reach a high enough rate of events.
Dr. Smith's Zometa 704 study tested the ability of the bisphosphonate zoledronic acid to delay or prevent metastases in men with negative bone scans but rising PSA levels during androgen deprivation. Men were randomly assigned to standard care plus zoledronic acid or to standard care plus placebo.
But the trial was stopped early, in September 2002, because there were far fewer events than would be necessary to see any difference, even if there had been a beneficial effect, Dr. Smith said at a news conference.
“One of the main obstacles to taking on this problem has been identifying the right clinical opportunity. This disease state—castrate, nonmetastatic prostate cancer—was felt to be the right place to look at agents to delay or prevent metastases, but the low event rate made it a daunting prospect. In an all-comers trial with rising PSA despite castration, the studies would have to be prohibitively large.”
But if a study were to include only those patients with higher baseline PSAs or faster PSA doubling times, the event rates might become quite considerable, he said.
“With PSA and PSA kinetics you could do clinical trials with as few as 1,000 patients. Hopefully, one of the larger unmet needs, prevention of metastases, becomes an operationally possible clinical trial in this particular setting.”
Dr. Smith said that he and colleagues have proposed just such a clinical trial. “Having thought a lot about this, I don't think there's another disease state in which you can ask the question.”