BOSTON (AP) _ Genetic researchers have developed what may be the most accurate method yet for answering a woman's scariest question about breast cancer: Will it spread and kill me?
By combining such a test with existing techniques, doctors could answer with greater certainty. That, in turn, could ease many women's fears and enable more of them to skip the grueling ordeal of chemotherapy.
The approach is still experimental, but researchers say a cancer genetic test _ perhaps one that is even more accurate than the newly developed method _ could begin coming into use within the next several years.
And researchers say the technique could, in theory, be adapted for use with other types of cancer, too.
``What this is all telling us is the fantastic promise of this field,'' said Dr. Carlos Caldas, a cancer specialist at the University of Cambridge in England. ``I think it's going to change completely the way we practice.''
He added: ``I think that the era of treating tumors just based on their size, the extent of where they have spread in the body, and how they look under the microscope _ these days are numbered.''
The test was developed at the Netherlands Cancer Institute, with help from Rosetta Inpharmatics in Kirkland, Wash. The findings were published in Thursday's New England Journal of Medicine.
About 200,000 U.S. women are diagnosed with breast cancer and 40,000 die yearly, according to the American Cancer Society. Even with tumor-removing surgery, about a quarter of breast cancers that show no sign of spreading will return elsewhere in the body within 10 years.
Standard prediction factors include the tumor's size, how it looks under a microscope, and the patient's age, since breast cancer in younger women tends to be more aggressive. However, such criteria give only a rough prediction. To be safe, doctors provide chemotherapy to the vast majority of these women, despite side effects such as nausea, fatigue and hair loss.
Genetic researchers have begun to study potentially powerful new methods to predict survival chances of patients with breast, lung, immune-system, prostate and other cancers. Even though all human cells, healthy or cancerous, carry virtually identical genes, they differ widely in which ones are switched on or off at a given time. A recent technology known as gene microarrays, or chips, can show this.
The Dutch researchers used a 1-by-3-inch chip to examine simultaneously which of 25,000 genes were expressed, or turned on, in tumor samples from 295 women all younger than 53. They discovered 70 genes that, when turned on or off in a certain pattern, predict either that the cancer will spread and kill the patient, or that it won't.
In the group genetically defined as high-risk, half developed cancer elsewhere and 45 percent died within 10 years. In the low-risk group, 85 percent stayed free of recurrences and 95 percent survived. The high-risk patients were five times more likely to see their cancer spread.
The genetic test would put fewer women, only 61 percent, into the high-risk category, compared with 90 percent under standard American methods, the researchers figured. High-risk cases are candidates for chemotherapy. That means the genetic test could reduce the number of women getting chemotherapy by a third.
The genetic test is also more accurate in predicting recurrences of cancer in women in the low-risk category. In 151 patients with no cancer spreading to their lymph nodes, only about 10 percent had cancer again within 10 years. About 20 percent did with current criteria.
The researcher who oversaw the work, molecular biologist Rene Bernards, said his group's genetic test might cost roughly $1,500. A round of chemotherapy would typically cost several thousand or more.
Outside experts said the research provides some of the strongest evidence yet that such a method can predict survival well.
``What I'm impressed about with this current work is the implication that these types of things can be done,'' said Dr. Anne Kallioniemi, a Finnish cancer geneticist at the University of Tampere. ``It's the best as we know of today _ not to say it will be the knowledge in 10 years.''
Dr. Mark Robson, a cancer expert at Memorial Sloan-Kettering Cancer Center in New York, added: ``Everybody's working hypothesis is that this type of study could be done with many other tumor types, and you would get similar results.''
There are some difficulties with the genetic approach, though. Such tests require special expertise. Also, tumors are normally preserved in a way that ruins them for such genetic testing.
Still, the Dutch group plans to begin using the technique on some breast cancer patients next year at the hospital of the Netherlands Cancer Institute. It also plans to join with researchers at Massachusetts General Hospital in Boston in further testing the results to satisfy U.S. government regulators.
Genetic researchers say tumors from other women, including older ones, must be studied before the best set of prediction genes can be identified with confidence.