FDA gives its approval to new gene test to predict breast-cancer relapse

WASHINGTON -- Women with early stage breast cancer may soon get another gene test to help predict whether they'll relapse in five or 10 years, information that could influence how aggressively they fight the initial tumor.

The MammaPrint test isn't the nation's first such predictor for breast cancer -- a competitor has sold here since 2004 -- but on Tuesday it became the first to win formal approval from the Food and Drug Administration.

The test is far from perfect, warned FDA's Dr. Steven Gutman.

Indeed, the MammaPrint is much better at predicting who isn't likely to relapse than who is, Gutman said.

He cited studies suggesting that when the MammaPrint predicts a woman is at high risk of cancer returning in five years, it will be right just a quarter of the time.

That compares with 95 percent accuracy if the woman is told she's at low risk of a relapse.

"This information has to be used very carefully by physicians," Gutman said. "This is a complex test."

MammaPrint already sells in Europe, but it's unclear how soon sales here would begin.

More than 178,000 U.S. women will be diagnosed with breast cancer this year.

More than 100,000 of them will have early stage tumors with a remarkably good prognosis: They're small, they haven't yet spread to the lymph nodes, and they're sensitive to hormones.

The vast majority of those women would survive with surgery, radiation and hormone treatment. Yet guidelines today recommend chemotherapy as well for most to catch the few -- two or three out of every 100 patients -- who need more aggressive help to try to prevent a recurrence years later.

Gene testing can help doctors and patients winnow out who's most likely to need the chemo from who might gamble on skipping it. For about two years, a growing number of breast cancer specialists have begun doing just that, mostly using a $3,000 gene test called Oncotype DX.

But until now, the FDA hasn't required formal approval of that testing, allowing sales instead as doctor-ordered services that laboratories provide, Gutman said.

Tuesday's approval doesn't mean the MammaPrint works better than what's already used, just that its maker sought formal FDA review and approval. The agency now is debating whether existing gene-predictor tests need similar oversight, Gutman said.

How does MammaPrint work? It analyzes the level of activity of 70 genes in a sample of the woman's surgically removed tumor, and provides a numerical ranking. Women ranked as high-risk have about twice the likelihood of a relapse in the next five to 10 years as women who are ranked low-risk, Gutman said.

The FDA has no information on whether the test results actually helped improve the survival of high-risk women, Gutman said.