Study of cancer trials finds significant safety improvement

Risk of dying from experimental cancer treatment drops by 90 percent over 12 years

The chance that patients participating in early-stage cancer research studies will die from the experimental treatments has dropped dramatically over the past decade, according to a study from the Massachusetts General Hospital (MGH) Cancer Center and the Massachusetts Institute of Technology (MIT). In an analysis of more than 200 Phase 1 research trials from 1991 through 2002, the researchers found that treatment-related deaths decreased by 90 percent during the study period. The report appears in the November 3 Journal of the American Medical Association.

“We undertook this study because there has been so much concern in recent years about the safety of clinical trials. We wanted to see if the increased attention to patient safety had made a difference and if the addition of targeted therapies, which tend to be less toxic, had also helped,” says Thomas Roberts Jr., MD, of the MGH Cancer Center, the paper’s lead author. “Now we can tell patients with cancer that, compared with 10 years ago, they can expect a higher level of safety when they enroll in early-stage clinical trials.”

Representing the first time potential new drugs are tested in humans, phase 1 trials have a goal of assuring drug safety and determining the best dosage. While most of these early-stage studies enroll healthy volunteers, cancer studies have several important differences. Phase 1 cancer studies are usually restricted to patients with cancer who have exhausted established therapeutic options. While identifying toxic effects is the primary goal of most Phase 1 studies, cancer trials also have a secondary goal of evaluating anti-tumor effects.

Since there had been no comprehensive analysis of Phase 1 trials since the mid-1980s, the researchers compiled a database of Phase 1 trial results reported at the annual meeting of the American Society for Clinical Oncology from 1991 through 2002. In order to insure that the studies were comparable, the investigators focused on published studies of single agents that had not yet received FDA approval and excluded those involving radiation therapy or treatment of leukemia or lymphoma. This strategy narrowed the study group down to 213 trials, enrolling aproximately 6,500 patients. For these studies, the researchers analyzed how often participants died from drug toxicity, cancer-related deaths, other toxic treatment effects, and whether or not the treatment caused the tumor to shrink, as measured by CT scanning.

The most significant change during the study period was the more than 90 percent drop in the risk of drug-related deaths – from a risk of about 1 percent in the first four years of the study to .06 percent in the last four years. The chance that the tested drug would have a measurable anti-tumor effect also dropped during the study period but by only 50 percent, suggesting a possible improvement in the overall risk/benefit ratio.

The researchers cite several possible reasons for the improvements in safety, including increased attention to patient safety regulations, the use of less-toxic targeted therapies, and improvements in supportive care, such as new treatments for chemotherapy-induced anemia and neutropenia. While they were surprised and concerned about the reduced chance of a therapeutic benefit, Roberts explains several potential underlying reasons. “We have gotten more systematic in the determination of response rates, so the later results may be more accurate,” he says. “In addition, some of the newer agents like angiogenesis inhibitors, could be stopping cancer progression without actually shrinking the tumor. We may need to find new ways to measure treatment success.”

Roberts continues, “Many investigators feel frustrated about the regulatory hurdles they have to go through to initiate and conduct clinical trials. There will always be a balance between optimizing patient safety and conducting research efficiently. We need to be aware of that balance and to find ways to monitor patient safety in real time.” Roberts is an instructor in Medicine at Harvard Medical School and visiting scientist at MIT.

Roberts’ co-authors are Bernardo Goulart, MD, Bruce Chabner, MD, and Jeffrey Clark, MD of the MGH Cancer Center; Elkan Halpern, PhD, and G. Scott Gazelle, MD, MPH, PhD, of the MGH Institute of Technology Assessment; and Sarah Stallings, PhD, Stan Finkelstein, MD, and Lee Squitieri of Massachusetts Institute of Technology. The study was supported by grants from the Alfred P. Sloan Foundation and the National Cancer Institute.