SAN ANTONIO ? A first-of-its-kind genetic test soon will be available to help women with breast cancer make one of their most crucial decisions: whether to undergo the rigors of chemotherapy.

Genomic Health Inc., a Silicon Valley biotech company, said it had identified nearly two dozen genes that, taken together, could predict with a high degree of accuracy the likelihood that tumors would return in women whose breast cancer was caught at an early stage.

Currently, doctors predict the chances of a relapse in pretty much the same way they have been doing it for almost a century: by looking at the patient’s age, the size of the tumor, and the tumor’s aggressiveness.

If the chances of recurrence are seen as very low based on the gene test, a woman may opt not to endure the illness, hair loss and high cost of chemo. But if the odds of the cancer coming back are high, she may view chemo as the difference between life and death.

“For the women in that highest group, it makes their decision so much easier,” said Dr. Melody Cobleigh, a breast cancer researcher at Rush-Presbyterian-St. Lukes Medical Center in Chicago. “For women in the lowest group, they’ll still agonize.”

The biotechnology company’s research — done with the National Surgical Adjuvant Breast and Bowel Project — was outlined Thursday at a breast cancer conference in San Antonio.

Chief executive Randy Scott said his company planned to make the test available to newly diagnosed cancer patients in early 2004.

The genetic test was developed by analyzing tumor samples from nearly 700 women involved in a 1980s cancer study. Genomic Health used its findings to create a point system to express escalating chances of recurrence within 10 years.

The sample group included women from their 30s to their 70s. All had early-stage breast cancer that had not spread to other parts of the body, and all had been treated for at least five years with tamoxifen, a drug that slows or stops the growth of new breast cancer cells.

Genomic Health found that just over half the women in its sample group fell into the low-risk category. Their average recurrence rate was about 7 percent after surgery to remove the tumor, followed by tamoxifen.

At the other end of the scale, 27 percent of the sample were determined to be at high risk of recurrence. About three of every 10 of these patients developed new breast cancer within 10 years of treatment.

In the intermediate-risk group, the 10-year recurrence rate averaged about 14 percent.

The error rate in each of the three categories was plus or minus 2 to 3 percentage points, said Dr. Steven Shak, Genomic Health’s chief medical officer.

Scott said that by using a scoring system, his company’s test could give women a more individualized prognosis.

“What we’ll be able to provide is not just that you are low-risk, medium-risk or high-risk,” he said. “It’s that you have a score of X and that in our 668-patient clinical trial, that corresponded to a recurrence rate of Y.”

Dr. Daniel Hayes, who heads the breast oncology clinic at the University of Michigan, said too many breast cancer patients were needlessly undergoing chemo because the established ways of predicting recurrence did not provide enough information for them to say no.

Chemo “is a heart-wrenching, gut-wrenching, hand-wringing decision” that nearly half of his patients have to make, Hayes said. If the new genetic test proves accurate, he could tell a patient she might safely be able to avoid chemo, he said.

Dr. Matthew Ellis, a breast cancer specialist at Washington University in St. Louis, said he saw much promise in the scoring method, even though it did not precisely measure each woman’s individual risk of relapse.

“What we’re seeing here is version 1.0,” said Ellis, who is part of a National Cancer Institute group focused on finding the best way to predict recurrence. “Version 1.0 is a great step forward, but we should not overestimate what we have right now.”

Scott said his Redwood City, Calif., company would keep tinkering with the product.

“Our long-term vision is to take this technology and broaden it out,” he said. “We will continue to investigate not only in other areas of breast cancer, but in other cancers in general.”