News

PAM50 assay aids prediction of metastasis in early node-positive breast cancer

Major finding: Among patients with one positive node, those falling into the ROR low-risk group (41% of the one-node subset) and those having the luminal A intrinsic subtype (71% of the one-node subset) had 10-year risks of metastases that were less than 9%.

Data source: A combined analysis of 543 postmenopausal patients with node-positive, hormone receptor–positive, early-stage breast cancer who received adjuvant endocrine therapy in the ABCSG-8 and ATAC trials.

Disclosures: Dr. Gnant disclosed ties to several companies including Sanofi-Aventis, Novartis, and Roche, as well as NanoString Technologies, the maker of the PAM50 assay.

View on the News

Confirmatory analysis would be welcomed

The ATAC population was used in this study for training and validation of PAM50, which could lead to overestimation of the test’s prognostic value. That said, the data looked pretty impressive.

Is the assay ready for prime time in node-positive patients? In my mind, probably not quite yet. The present analysis involved the testing of multiple comparisons, and in the end, even with 543 specimens it was limited by a relatively small number of events.

PAM50 classification approaches are probably useful, but which approaches are best – the ROR score, the intrinsic subtype, or the ROR group – remains uncertain, and a confirmatory analysis would certainly be welcomed.

Dr. Eric P. Winer is with the Dana-Farber Cancer Institute in Boston. He was the invited discussant of the study. Dr. Winer disclosed ties to AstraZeneca, Merrimack, Pfizer, and other companies.


 

AT THE ASCO ANNUAL MEETING 2013

CHICAGO – The multigene PAM50 assay identifies postmenopausal women with node-positive early breast cancer who have a low risk of distant metastases and may be able to skip adjuvant chemotherapy.

The 543 women studied had one to three positive nodes and low- or intermediate-grade tumors; all received adjuvant endocrine therapy but no chemotherapy in a pair of randomized trials: the ABCSG-8 (Austrian Breast Cancer Study Group 8) trial and the ATAC (Arimidex, Tamoxifen, Alone or in Combination) trial.

Results showed that three PAM50 parameters – the risk of recurrence (ROR) score, ROR risk group, and intrinsic subtype – significantly added to clinicopathologic factors in predicting the 10-year risk of metastases.

Dr. Michael Gnant

"A significant proportion of early breast cancer patients with only one positive node has a limited long-term risk of metastases, as well as at least some patients with two or three positive nodes," first author Dr. Michael Gnant commented in presenting the findings at the annual meeting of the American Society of Clinical Oncology.

In particular, among patients with only one positive node, those falling into the ROR low-risk group (41% of the one-node subset) and those having the luminal A intrinsic subtype (71% of the one-node subset) had 10-year risks of metastases that were less than 9%.

"Based on their low risk of metastases, these patients could – and I would personally add should be spared the side effects of adjuvant chemotherapy despite having node-positive early breast cancer," maintained Dr. Gnant, who is a professor of surgery at the Medical University of Vienna, Austria.

"In the adjuvant treatment of hormone receptor–positive breast cancer, we every day face the dilemma of undertreatment versus overtreatment. Assessing an individual patient’s risk of metastasis as accurately as possible is an important goal in early breast cancer, and it directly impacts on treatment decisions," he added. "Still, risk assessment is mainly based on clinicopathological variables, of which nodal status is considered to be most important. As a consequence, most node-positive patients with hormone receptor–positive breast cancer globally still receive adjuvant chemotherapy."

In the study, the researchers assessed the added prognostic value derived from NanoString Technologies’ PAM50, a 50-gene assay designed to identify intrinsic subtypes of breast cancer. The subtypes are combined with a proliferation score and tumor size to yield the ROR score, and scores are used to classify patients into ROR-based risk groups.

The clinical usefulness of the PAM50 assay and other similar molecular assays, such as Genomic Health’s Oncotype DX and Agendia’s MammaPrint, has not been defined in node-positive patients, he noted.

The node-positive patients studied were drawn from the ABCSG-8 and ATAC trials. The former trial assigned women either to tamoxifen for 5 years or to 2 years of tamoxifen followed by 3 years of the aromatase inhibitor anastrozole (Arimidex); the latter assigned patients to 5 years of adjuvant anastrozole alone, tamoxifen alone, or the (subsequently discontinued) combination.

For each patient, the investigators analyzed tumor tissue to determine the PAM50 ROR score, ROR risk group, and intrinsic subtype. In addition, they derived the clinical treatment score, which is an overall integrated measure of clinicopathologic factors.

The median follow-up was 11 years, according to Dr. Gnant. As the ROR score increased from 0 to 100, so did the 10-year risk of metastases.

When added to the clinical treatment score alone, the ROR score significantly improved on the prediction of metastasis (chi-square for the difference in likelihood ratios, 32.45; P less than .0001).

The ROR risk groups also stratified patients well. The risk of metastases was about 30% in the high-risk group, 14% in the intermediate-risk group, and 8% in the low-risk group.

When added to the clinical treatment score, the luminal subtype significantly improved on prediction of metastasis (chi-square for the difference in likelihood ratios, 20.48; P less than .0001).

The estimated 10-year risk of metastasis was 11% in the luminal A group and 31% in the luminal B group.

In a combined analysis, the risk of metastases was less than 10% in several subsets of patients: those with one or two positive nodes in the ROR low-risk group (6.6% and 7.2% risks, respectively) and those with one positive node having the luminal A subtype (8.4% risk).

Finally, each factor – ROR score, ROR risk group, and intrinsic subtype – significantly improved on the prediction of risk above and beyond clinicopathologic factors in patients with one positive node, two positive nodes, and two or three positive nodes individually.

One session attendee asked, "Was there any opportunity to look at the interaction [of PAM50 parameters] with therapy?"

Next Article:

   Comments ()