MammaPrint Predicts Early Outcomes in Breast Cancer

Roxanne Nelson

February 08, 2013

Editor's note: The headline of this article has been changed. The previous headline, "Confirmed: MammaPrint Predicts Treatment in Breast Cancer," is not supported by data, an outside expert approached for comment, Christopher Benz, MD, told Medscape Medical News.

The first prospective trial of the 70-gene-signature MammaPrint (Agendia) breast cancer test shows that it can help determine which breast cancer patients can forgo adjuvant systemic treatment.

The results of the Microarray Prognostics in Breast Cancer (RASTER) study were published online January 31 in the International Journal of Cancer.

Of the women deemed to be low risk by the MammaPrint test, 85% chose not to undergo adjuvant chemotherapy, and 97% were disease-free after 5 years. Of the women deemed to be high risk, 91% underwent chemotherapy were disease-free at 5 years.

It might be prudent to incorporate MammaPrint results with the traditional clinical parameters, but only in patients older than 45 years for whom the value of adjuvant chemotherapy is unclear, said lead author Sabine C. Linn, MD, PhD, from the Department of Medical Oncology at the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital in Amsterdam.

The RASTER study is the first to prospectively evaluate the performance of MammaPrint. Decisions about adjuvant therapy were made on the basis of the 2004 Dutch Institute of Healthcare Improvement (CBO) guidelines, MammaPrint results, and preferences of clinicians and their patients.

Five-year distant-recurrence-free-interval probabilities estimated with MammaPrint and another clinical common tool, Adjuvant! Online (AOL) — used to assess the risk of developing recurrent disease and/or dying within 10 years — were compared.

Points to Ponder

This study is an important step forward, according to Christopher Benz, MD, who is director of the cancer and developmental therapeutics program at the Buck Institute for Research on Aging in Novato, California. He was not involved in the study.

"The study has largely accomplished its primary aims of demonstrating the feasibility of implementing a multigene signature test in a community-based setting and showing that its implementation can have a significant impact on medical decision-making," he told Medscape Medical News.

In more than 25% of the study population, MammaPrint gave a risk estimate (low risk) that was discordant with the AOL risk estimate (high risk). However, the researchers assumed that clinicians equate a high-risk estimate from AOL with a less than 90% chance of recurrence in 10 years, "a definition that is not necessarily used in common practice," Dr. Benz explained.

On the basis of that definition, 124 of 427 patients in this study were classified as high risk by AOL but low risk by MammaPrint, he continued. Despite this discordance, more than 75% of Dutch clinicians chose not to recommend adjuvant chemotherapy.

"As apparent validation of the MammaPrint prediction, the authors show that over 98% of these patients were still disease-free 5 years later," Dr. Benz noted. Although this makes gene-signature tests like MammaPrint appealing to oncologists "it is hazardous to draw conclusions from breast cancer recurrence rates after only 5 years of follow-up," he said.

In the RASTER study, all breast cancers were node-negative and 80% were estrogen-receptor (ER)-positive. "The natural history of such breast tumors is that most clinical relapses do not occur until 5 or even 10 years after diagnosis — with or without the use of adjuvant therapy," he added.

MammaPrint vs Oncotype

MammaPrint and Oncotype DX (Genomic Health) are the 2 major gene-profiling tests for breast cancer currently available, and there are notable differences between the 2. Oncotype DX measures the expression of 21 genes with polymerase chain reaction, and can use fixed tissue. MammaPrint applies microarray technology to a 70-gene signature, and can use both fresh and fixed tissue.

In addition, MammaPrint can be used in patients with any ER status, whereas Oncotype DX can only be used in ER-positive patients. "MammaPrint is also useful in HER2-positive disease, but Oncotype is not," said Dr. Linn. "There is currently uncertainty about the magnitude of additional information generated by Oncotype DX, since a British group showed that Oncotype is similar to immunohistochemical testing for ER, PGR, HER2, and Ki67," she explained.

The prognostic value of MammaPrint and Oncotype DX overlap considerably, and both are often discordant with the AOL risk estimator, Dr. Benz noted. However, he pointed out that Oncotype DX specifically addresses whether an ER-positive breast cancer patient will likely benefit from adjuvant chemotherapy in addition to adjuvant hormonal therapy (tamoxifen), whereas MammaPrint does not address this "very practical and more specific question."

"Clinical oncologists across the United States are eagerly awaiting the results of the randomized MINDACT trial, which will definitively address the accuracy of this 70-gene signature (relative to AOL), regardless of tumor ER status, in predicting which breast cancer patients can safely avoid adjuvant chemotherapy," he explained. Until then, the widespread use of the 70-gene MammaPrint signature is unlikely, he noted.

Study Details

The primary goal of RASTER was to evaluate MammaPrint in a community-based setting and to study its clinical impact on decision making for adjuvant therapy.

The definitions of high risk and low risk used by Dr. Linn and colleagues were in accordance with CBO guidelines. Patients of low clinical risk were older than 35 years with a grade 1 tumor of 30 mm or smaller, a grade 2 tumor of 20 mm or smaller, or a grade 3 tumor of 10 mm or smaller; or they were younger than 36 years with a grade 1 tumor of 10 mm or smaller.

All other patients were considered to be at high risk.

Adjuvant endocrine treatment in the CBO guidelines is only advised in clinically high-risk patients with hormone-receptor-positive tumors in combination with chemotherapy.

The median follow-up was 61.6 months. In the MammaPrint group, 15% (33 of 219) of the low-risk patients received adjuvant chemotherapy, as did 81% (69 of 208) of the high-risk ones.

The 5-year distant-recurrence-free-interval probabilities in the low-risk and high-risk MammaPrint groups were 97.0% and 91.7%, respectively. The probabilities in the low-risk (n = 132) and high-risk (n = 295) AOL groups were 96.7% and 93.4%, respectively.

The researchers observed that the MammaPrint and AOL risk estimations were discordant in 38% of the cases (161 of 427 patients). Most of these (29%) were low-risk MammaPrint and high-risk AOL; 9% were high-risk MammaPrint and low-risk AOL.

Of the patients who were categorized as low risk on both MammaPrint and AOL, 93% did not receive any adjuvant therapy (not chemotherapy or endocrine therapy). In addition, 56% of patients who were low-risk MammaPrint and high-risk AOL did not receive any adjuvant therapy.

The RASTER study was financially supported the Dutch Health Care Insurance Board. Coauthor L.J. van't Veer, PhD, from the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital and the Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, reports being a shareholder and being employed by Agendia, which markets MammaPrint, and is named as an inventor on the patent. Coauthor M.J. van de Vijver, MD, PhD, from the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, is named as an inventor on the MammaPrint patent. Coauthor W.H. van Harten, MD, PhD, from the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, reports being a nonremunerated, nonstakeholding member of the supervisory board of Agendia. Coauthor M. Knauer, from the Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital and the Sisters of Charity Hospital and Cancer Center in Linz, Austria, reports receiving unrestricted educational grants from Agendia.

Int J Cancer. Published online January 31, 2013. Abstract

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....