4/16/99
MEDIA CONTACT: Ruthann Richter, [email protected], or
M.A Malone, [email protected], (650) 723-6911
FOR COMMENT: Thomas Stamey, MD, (650) 725-5542
EMBARGOED FOR RELEASE: Tuesday, April 20, 1999, 3 p.m. CDT, to correspond with publication in the April 21 issue of the Journal of the American Medical Association
STANFORD RESEARCHERS DISCOVER KEY TO PREDICTING EFFECTIVENESS OF PROSTATE CANCER SURGERY
STANFORD -- One of the great dilemmas in treating men with prostate cancer is the problem of how to predict which patients are likely to respond well to surgery and which patients are destined to fail surgical treatment and would be better off seeking alternative options. In the past, doctors have considered several different tumor characteristics to help them make these critical treatment decisions.
In a new study, painstakingly conducted over five years, Stanford researchers say they have discovered which of these factors are the direct cause of failure to cure prostate cancer. They found that among nine variables, only two tumor characteristics are really critical - the size of the tumor and the presence of grade 4 cancer, which is a more aggressive pattern of disease, said Thomas Stamey, MD, professor of urology and chief author of the study. Other factors once thought to be important did not prove to be significant in predicting whether surgery will cure the patient or not, Stamey said. These included whether the tumor had penetrated the outer covering of the prostate and whether there were positive surgical margins, a pathologic measurement of residual cancer. The findings contradict traditional thinking about the disease and could change the way physicians manage prostate cancer patients in the future, he said.
"This paper challenges the conventional thinking about this cancer, giving us much more power to predict prognosis," Stamey said.
"The study defines who can be cured much better than we were able to do before," he added. "In terms of research, it's a quantum leap forward. If you know what part of the cancer drives the disease, you know where to focus your research."
The study is being published in the April 21 issue of the Journal of the American Medical Association.
Prostate cancer is the second most common cause of cancer deaths among men. It was responsible for more than 39,000 deaths in the United States in 1998, according to the American Cancer Society. Some 184,500 men were diagnosed with the disease that same year, according to cancer society figures.
In the Stanford study, the researchers tracked the progress of 379 cancer patients who had had their prostates surgically removed between August 1983 and July 1992 at Stanford University Medical Center. The researchers followed the men, whose median age was 65, for a minimum of five years in order to clearly define when surgery had been successful and when it had failed. To determine outcomes, they used a highly sensitive assay in their laboratory to measure prostate specific antigen (PSA), a protein in the blood commonly used as a marker of the disease. PSA levels in 38 percent of the patients showed there were residual cancer cells in their system - an indication, by this measure, that surgery had failed.
At the same time, the Stanford group examined the patients' diseased prostates in the laboratory, measuring nine tumor characteristics, such as tumor size, whether blood vessels in the prostate were invaded by the disease and whether the cancer extended outside the organ. They also ranked the tumors on a standard grading scale of 1 to 5, with 4 and 5 being the most aggressive. They discovered that for every 10 percent increase in grade 4, there was a 10 percent loss in the ability to cure the cancer. Eighty percent of all 379 men in the study had some grade 4 cancer.
The researchers then used three statistical models to compare the failure results with the nine tumor characteristics to determine how they might correlate. All three models pointed to only two key factors - tumor size and the presence of grade 4 cancer - as the primary determinants of surgical failure.
"It's telling us what's driving the cancer," Stamey said. "We now know what to look for, and only two factors out of the nine are actually driving this cancer with very significant probabilities. One is the percent of the tumor composed of grade 4-5 and the other is the size of the tumor. They are independent predictors of failure. And everything else paled in significance."
The results suggest that the current system for tumor grading, known as the Gleason scoring system, can be vastly improved by measuring only the percent of cancer consisting of grades 4 and 5, Stamey said. He noted that among the 20 percent of study participants who had no grade 4 or 5 cancer, the cure rate was 94 percent. This raises the question of whether these patients need to be treated at all, an issue that will have to be decided in a randomized clinical trial, he said.
In the meantime, "the finding that curability is lost with every 10 percent increase in Gleason grade 4 cancer will be a great help in deciding who should be treated," Stamey said.
He said other factors once considered important, such as how much cancer has penetrated the outer covering (capsule) of the organ, the presence of positive surgical margins and whether the cancer is totally confined or not, are not statistically significant as independent predictors of surgical failure.
"Thirty-three percent of the 234 men cured by surgery had lots of capsular penetration, and even in the patients where the cancer was totally confined within the capsule and who should have been cured, 15 percent ultimately failed. Only cancer volume and the Gleason grade 4/5 cancer were the major determinants of failure to cure prostate cancer," he said.
The study also provides new direction for research in prostate cancer, Stamey said. His goal now is to identify the genes expressed in grade 4 tumors, as these are the genes that should be responsible for disease progression, he said. Another goal of research, he said, is to develop more accurate pre-treatment estimates of cancer size.
Stamey's colleagues in the study were John McNeal, MD, professor of urology and pathology; Cheryl Yemoto, laboratory manager; Bronislava Sigal, PhD, a postdoctoral fellow in statistics; and Iain Johnstone, PhD, professor of statistics.
The study was conducted in part with a grant from the Lucas Foundation of Menlo Park, Calif.
###
RSS