SAN ANTONIO (AP) ¿ Thousands of breast cancer patients each year could be spared chemotherapy or get gentler versions of it without harming their odds of beating the disease, new research suggests.

One study found that certain women did better ¿ were less likely to die or have a relapse ¿ if given a less harsh drug than Adriamycin, a mainstay of treatment for decades.

Another study found that a gene test can help predict whether some women need chemo at all ¿ even among those whose cancer has spread to their lymph nodes, which typically brings full treatment now.

The findings are sure to speed the growing trend away from chemo for many breast cancer patients and targetting it to a smaller group of women who truly need it, doctors said last week at the San Antonio Breast Cancer Symposium, where the studies were reported.

"We are backing off on chemotherapy and using chemotherapy more selectively" in certain women, said Dr. Eric Winer of the Dana-Farber Cancer Institute in Boston.

The gene test in particular "will start changing practice nearly immediately," said Dr. Peter Ravdin of the University of Texas M.D. Anderson Cancer Center in Houston. "The results are compelling that this test ... helps select patients who will most benefit from chemotherapy."

Breast cancer is the most common major cancer in American women. More than 178,000 new cases are expected this year. Most are helped to grow by estrogen, and hormone-blocking medicines like tamoxifen are used to treat those.

Chemo usually is added if the disease has spread to lymph nodes ¿ a situation faced by about 45,000 US women each year. Doctors know that chemo won't help most of these women, but they have had no good way to tell who can safely skip its cost and misery.

Here's where Oncotype DX, a test that measures the activity of 21 genes and gives a score to predict a woman's risk of recurrence, comes in. Doctors have used it for several years to guide treatment for certain women with early breast cancers, especially those that not spread.

The new study, led by Dr. Kathy Albain of Loyola University, looked at whether it accurately predicted chemo's benefit in 367 women whose hormone-driven cancer had spread to lymph nodes. A decade after these women were treated, those who had low scores on the gene test were found to have had no benefit from chemo. Conversely, chemo did a lot of good for those with high scores.

Because 40 percent of the women scored low, it means that as many as 18,000 women each year might safely skip chemo.

The National Cancer Institute and the test's maker, Genomic Health of Redwood City, California, sponsored the study. Albain, Winer and Ravdin have consulted or been paid speakers for the company in the past.

Dr. Kelly Marcom, a Duke University cancer expert with no ties to the company, said the test would give valuable information to guide treatment for more patients in the future. He has used it on about 50 women in the last year.

"I've had it cut both ways" ¿ ruling chemo in and out, Marcom said.

The test is expensive ¿ $3,400 ¿ though many insurers are paying for it because it can avoid even more costly chemo.

Albain plans to discuss using it with Andrea DeRosier, a 49-year-old health care administrator from suburban Chicago whose cancer has spread to a single lymph node.

When a surgeon said she likely would need chemo, "I remember thinking, 'Oh, that's terrible'," DeRosier said. "I want whatever protocol is going to keep me alive," but not futile treatment, she said.

Chemo's side effects are getting greater attention. One drug commonly used for early breast cancer ¿ doxorubicin, sold as Adriamycin and generic brands ¿ is known to cut the risk of having a recurrence or dying, but raises the risk of heart problems and even leukaemia.

Dr. Stephen Jones of Baylor-Sammons Cancer Center tested using Taxotere, a drug not linked to heart problems, in its place in more than 1,000 women with early breast cancer. After seven years, 87 percent of those given Taxotere survived, compared with 82 percent of those given Adriamycin. In addition, those given Taxotere were less likely to have had a recurrence.