70-Gene Signature Identifies Breast Cancer Subgroup Unlikely to Benefit From Adjuvant Chemo
Martine Piccart, MD, PhD
A 70-gene signature (MammaPrint) demonstrated a high level of accuracy at identifying a large subset of women with clinically high-risk early stage breast cancer for whom adjuvant chemotherapy was unlikely to produce benefit, according to findings from a phase III trial presented at the AACR Annual Meeting 2016.
In the study known as MINDACT, patients deemed high risk clinically but low risk by the gene signature had a similar 5-year rate of distant metastasis-free survival (DMFS) whether randomized to adjuvant chemotherapy or not, said Martine Piccart, MD, PhD. “The important message here is among the clinically high-risk patients, the clinical use of MammaPrint is associated with almost a halving of the use of chemotherapy,” she said
The rate of 5-year DFMS in women who were clinically high risk/genomically low risk and randomized to no chemotherapy was the primary statistical test for MINDACT. In this group, in which 48% of the women had positive nodes, the 5-year DMFS was 94.7% (95% CI, 92.5%-96.2%), which passed the bar for significance of 92%.
Overtreatment with adjuvant therapies, especially chemotherapy, is common for patients with breast cancer, in an attempt to eradicate micrometastases. In this setting, overtreatment is considered adjuvant chemotherapy that is associated with a small survival benefit, in the range of 2%, while exposing the patient to long-term risks such as secondary cancers, secondary leukemia, and congestive heart failure, said Piccart, Head, Department of Medicine, Jules Bordet Institute in Brussels, Belgium, and Co-Founder and Chair of the Breast International Group.
The MINDACT trial was opened in 2007 and originally included women with negative nodes, but was amended in 2009 to enroll women with one to three positive nodes. “Essentially, we got very confident that the genomic assay would outperform the clinical criteria by reducing the prescription of adjuvant chemotherapy without impairing patient outcome,” said Piccart. As such, MINDACT is the only clinical trial pitting tumor biology against tumor anatomy with a few biological features added, she said.
MINDACT enrolled 6693 patients with early breast cancer from 112 centers in nine European countries who had their risk of tumor recurrence following surgery assessed in two ways: through use of MammaPrint, performed on frozen tumor tissue, and also via Adjuvant! Online. Overall, 2745 women were categorized as low risk using both methods, 1806 were categorized as having high risk of recurrence by both methods, 592 were categorized as high risk of recurrence by MammaPrint and low risk of recurrence by Adjuvant! Online, and 1550 were categorized as low risk of recurrence by MammaPrint and high risk of recurrence by Adjuvant! Online.
Patients characterized by both assessments as low risk were spared adjuvant chemotherapy while chemotherapy was advised for those characterized as high risk by both methods. Those with discordant results were randomized to adjuvant chemotherapy or no adjuvant chemotherapy.
Altogether, 88% of the patients enrolled had hormone receptor (HR)-positive tumors and 10% had biologically aggressive HER2-positive disease. After a median follow-up of 5 years, 3% of the study population died and 5.4% experienced either distant metastases or death.
Five-year DMFS was 97.6% among the women who were low risk by both assessment methods. In contrast, DMFS was 90.6% among the women who were high risk by both methods and received adjuvant chemotherapy. “The low-risk patients were mostly node-negative [with] small tumors with hormone receptors,” said Piccart. The high-risk patients had larger tumors, were node-positive in 25% of the cases, and triple-negative in about one third of cases, she added.
The “discordant” groups, meaning low by one measure and high by another, had a rate of DMFS in between that of the “concordant” groups (low-low and high-high). The 5-year DMFS rates were 94.8% and 95.1% in the patients who were clinically low risk/genomically high risk and clinically high risk/genomically low risk, respectively.
The trial was not powered to detect a clinical benefit to chemotherapy in discordant risk groups, she explained, but on intent-to-treat analysis, a 22% relative reduction in the risk of 5-year DMFS with chemotherapy was observed “which would translate into a very small absolute benefit that would not justify the risks of chemotherapy,” she said.
Among the entire MINDACT population, using a clinical strategy to assign chemotherapy would result in 50% of patients receiving chemotherapy, compared with only 36% using the genomic strategy, for an absolute reduction of 14% in chemotherapy prescription, Piccart noted.
“MINDACT provides level 1A evidence of the clinical utility of MammaPrint,” she concluded. This is the highest level of evidence shown to date for a risk assessment tool for determining whether chemotherapy should be used.
Because nearly all ER-negative or HER2-postive tumors would be expected to rate as both clinically and genomically high risk, MammaPrint and the MINDACT results bear mostly on the heterogeneity within ER-positive, HER2-negative cancers, commented Harold Burstein, MD, PhD, clinical oncologist at Dana-Farber Cancer Institute, Boston.
The study “confirms the primary hypothesis that integration of genomic signature permits identification of a cohort of ER-positive tumors with good prognosis with endocrine therapy alone, regardless of larger T stage and N1 status,” he said, advocating that most ER-positive, HER2-negative stage 1 and 2 cancers, including N1, warrant tumor genomic profiling for optimal adjuvant decision-making.
Piccart M, Rutgers E, van’t Veer L, et al. Primary analysis of the EORTC 10041/ BIG 3-04 MINDACT study: a prospective, randomized study evaluating the clinical utility of the 70-gene signature (MammaPrint) combined with common clinical-pathological criteria for selection of patients for adjuvant chemotherapy in breast cancer with 0 to 3 positive nodes. Presented at: AACR Annual Meeting 2016, New Orleans; April 16-20, 2016. Abstract CT-039.
<<< View more from the 2016 AACR Annual Meeting