Early cancer treatments help more than thought

People with advanced cancers who try experimental treatments are helped more than previously thought, according to the most comprehensive look at government-sponsored tests over a decade.

These are patients who haven't benefited from other therapies and have few options left. But testing new treatments on them has been criticized by some who feel the patients are given false hope since previous reviews showed they only worked in about 4 percent to 6 percent.

However, this latest and largest study found that about 11 percent were helped by experimental treatment, and in some cases as many as 27 percent were better off.

"The numbers aren't as bleak as they sometimes are portrayed," said one of the researchers, Christine Grady, of the National Institutes of Health's Department of Clinical Bioethics. "But the numbers don't tell the whole story."

She said their review covered tests on different types of drugs and vaccines, combinations of drugs and some that had already been approved by the Food and Drug Administration. Only 22 percent of the reviewed tests were for a single chemotherapy drug; that was the only kind included in the reviews that showed low overall response.

She said patients who want to volunteer for early drug testing should look at the details for the type of experiment they are considering.

"Phase I trials are not all alike," she said.

The first phase of drug testing in people is primarily designed to see if the treatment is safe and to find the right dose for future tests, though doctors also hope to get some idea of whether it works. These early tests are done in cancer patients who have failed other treatments.

Grady and her colleagues analyzed 460 tests with 11,935 participants, including some previously unpublished tests. They were done between 1991 and 2002 and were paid for by the NIH's National Cancer Institute. Their review did not include tests financed by drug makers or those done on children. The research appears in Thursday's New England Journal of Medicine.

Overall, about 3 percent saw their cancers disappear and about 8 percent had a substantial shrinkage of their tumors. Additionally, 34 percent saw some tumor shrinkage or saw their disease stabilized.

The rate of deaths blamed on the treatments stayed the same, about half a percent. Side effect information was only available for about a third of participants; about 14 percent had at least one serious side effect.