Test predicts the recurrence of breast cancer

Calif. company analyzes key genes in tumors

December 05, 2003|By KNIGHT RIDDER/TRIBUNE

REDWOOD CITY, Calif. - An experimental test for breast cancer from a Redwood City company was able to predict with unprecedented precision which patients, among hundreds in a study, had a higher risk that the cancer would return after surgery, researchers announced yesterday.

The test, from Genomic Health, could prove helpful to many patients with early breast cancer who must make difficult decisions about further treatment. Those with a low risk of cancer recurrence might decide to forgo chemotherapy, the powerful chemical treatments that often cause severe nausea, hair loss and other problems. Those with a high risk would have an extra reason to go ahead.

At least one cancer specialist applauded the results of the first large-scale study of the new tumor test and said she wished it were immediately available to help plan treatment for her patients.

"It's the most important diagnostic in breast cancer for risk prediction I think I've ever seen," said Dr. Melody Cobleigh, director of the comprehensive breast cancer center at Rush Medical College in Chicago.

But others were more skeptical, calling for further studies that could show whether or not the tumor assay, which is bound to be costly, actually saves lives.

"Strictly speaking, this still doesn't answer the complete question," said Dr. Judith Luce, director of oncology services at San Francisco General Hospital. She said she wanted to see evidence that decisions based on the test "show patients do better." Such a study could take five years or more to complete.

Meanwhile, Genomic Health isn't waiting. The small, privately held company is gearing up to offer the test, which it calls OncotypeDX, early next year. It is setting up its own laboratory to test samples and has hired a small sales force to market the service to leading cancer treatment centers and then others.

The company must also persuade government programs and private insurance companies to pay for the testing.

"Decisions should not be made in the dark," said Dr. Steven Shak, the company's chief medical officer.

The test developed at Genomic Health looks at 21 genes thought to play some role in the growth of cancerous cells. The company developed techniques for testing the activity of those genes in tumor samples that had been preserved for a decade or more.

In preliminary studies, the Genomic Health team calculated a numerical risk score, from zero to 100, for a tumor, depending on the amount of activity of each of the genes.

The higher the score, the higher the risk that the tumor would come back after surgery. Or so the company scientists hoped.

To prove that it worked, the scientists retrieved samples of tumors that had been removed from 668 women who were part of another study begun 20 years ago to test the effectiveness of the hormone-blocking drug tamoxifen in patients whose cancer had not yet spread to their lymph nodes.

The researchers wanted to see if the experimental test from Genomic Health could accurately predict which of the 668 women would have a recurrence. They came up with risk scores for each of the tumor specimens and then reviewed the medical histories of the patients, who had been followed for a decade or more.

They found that over 10 years the cancer came back in 30 percent of patients with high scores on the Genomic Health test. But the cancer returned in only 7 percent of patients with low scores.

The results were released yesterday at the annual San Antonio Breast Cancer Symposium, which is widely attended by cancer specialists.

Conventional wisdom is that the larger the tumor and the younger the patient, the more likely that the tumor will return. But, at least in the group of 668 patients, the new test appears to be far better at predicting the likelihood of recurrence than either tumor size or patient age.

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.