Early Stage HR+ Breast Cancer: Optimizing Multigene Assays

Andrew D. Seidman, MD: Hope, there are 2 other issues on assays that I want you to touch on for us. Then we’ll finish talking about CDK4/6 inhibition prospects for early breast cancer. There was a new term that came up at ASCO [the American Society of Clinical Oncology annual meeting] that I was unaware of. The term is ultralow-risk, as applied to MammaPrint. There are some new data with that. Maybe you can talk a little about what that has told us and how it might impact practice. The second issue relates to the use of the Breast Cancer Index, some new data from B-42 and other studies helping us better understand the use of this assay for extended adjuvant aromatase inhibitor [AI]. First, we’ll go with ultralow-risk MammaPrint. What is this, and what do we do with it?

Hope S. Rugo, MD: As you recall, the 70-gene score or MammaPrint is yes or no. What this represents is a further definition of that score, which is incredibly important. There are actually 2 further definitions: ultralow and ultrahigh. We call ultrahigh MammaPrint 2, but in I-SPY, that will be a very important differentiation, too. Ultralow came about when Laura Esserman, MD, and Laura van ‘t Veer, PhD, at my institution were very interested in looking at an older data set of patients where there was long-term follow-up. Some got tamoxifen and some didn’t get any endocrine therapy. They looked at a lot of postmenopausal women to try to identify a score that represented a group of patients whose prognosis was so good that they wouldn’t benefit from anything much.

This was a retrospective analysis, but plus-1 is really low, and minus-1 is really high in MammaPrint, for everybody to be on the same page. If you had a score that was above 0.355, you were ultralow-risk. In that initial publication, it showed that those patients had an excellent outcome both with and without tamoxifen. In these analyses, of which there were 2, one was looking at the extended use of endocrine therapy and later recurrence risk in a population of patients who were evaluated for MINDACT. These patients seem to have better outcomes. If you have a very low risk, it indicates a very indolent cancer.

My caveat on that is we haven’t identified that those patients don’t need endocrine therapy. I wouldn’t take that as the take-home message. We should use endocrine therapy, but it may be that it doesn’t matter which endocrine therapy you use. Then if a patient really doesn’t tolerate endocrine therapy, some is better than none. In terms of disease burden, we don’t actually know what the recurrence risk will be at 10-plus years in this patient population, à la the Early Breast Cancer Trialists’ Collaborative Group analysis. But we know they don’t need chemotherapy. That is absolutely for sure, and they don’t need huge therapy.

Andrew D. Seidman, MD: I want to point out that the end point examined for this is 8-year distant metastasis-free interval, which was excellent at 95% to 98% for these patients who are largely node-negative T1 grade 1/2. I don’t know what 8-year distant metastasis-free interval means for a younger patient in this disease category.

Hope S. Rugo, MD: It’s important to keep in mind that very few young patients have ultralow scores. It shows you in our clinical practice, in our metastatic studies, and the studies of ovarian function suppression that young women tend to have more highly proliferative disease, and they don’t generally fall into this ultralow-risk category. You’re right; we don’t know what to do with them. Although we all wish that the Breast Cancer Index would give us an absolute yes or no answer, it doesn’t seem to have done that either.

The analysis of the NSABP [National Surgical Adjuvant Breast and Bowel Project] extended-duration endocrine therapy with AI was quite complicated. It’s interesting. It looked like there was almost a carry-over effect, that if you had a high-risk score and you gave extended-duration aromatase inhibitor therapy, you saw the benefit later, as you might expect, because you get a carry-over from your first 5 years for the first few years. It complicates it and suggests that you may be differentiating the group of patients who benefit from longer-term aromatase inhibitor therapy, but that we’re going to need even longer follow-up of the NSABP trial to really understand that true benefit.

Andrew D. Seidman, MD: Now, we’ll shift gears from MammaPrint to the Breast Cancer Index. What do you think of the evolving data on using that gene ratio to help parse the decision for 5 vs 10 years? Who are you thinking about? Is it the N1 patient, the N2 patient, the N0 patient?

Hope S. Rugo, MD: That’s what I was just describing from the carry-over effect with the Breast Cancer Index, the fascinating data that tell us a lot about the natural history of hormone receptor-positive breast cancer. Interestingly, in the MA.17 trial, if you took tamoxifen, we saw a difference from extended endocrine therapy with letrozole at 3 years. But in a striking contrast, if you’d had an AI before, you couldn’t differentiate benefit based on the Breast Cancer Index in the first few years after randomization vs taking a placebo vs letrozole. I thought it was really interesting that there’s a carry-over effect from the AI. We presume that it seems to play a role there. This was an unplanned analysis. Then you did see a separation of the curves later, suggesting that you have your impact later, which is fascinating. It suggests that we need longer-term follow-up.

Where would I use the score? It’s hard to know, because in patients who have node-positive disease and seem to have a higher risk of later recurrence, they continue to have a higher risk throughout the time period. You might just want to give the drug if they tolerate it. I’m not sure that the score is helping me decide enough. We haven’t seen this with the NSABP trial. The score might be more beneficial in the patients where we have equipoise, a T2N0 cancer, more than the node-positive. It’s a confusing situation, and I’m interested in what others think.

Andrew D. Seidman, MD: Clearly, we can let the assay alone drive the decision, but probably none of us do that. We factor in how the patient felt during those first 5 years on an AI, what their bone health is like, etc.

Dejan Juric, MD: What is clear from all of this, Andrew and Hope, is that we need something much more than these static profiles or markers. We are desperate to understand something that’s functional about these tumors and use of these dynamic markers. Obviously, in its first iteration, we rely on the old Ki-67 approach, but looking at delta after initiation of a certain period of systemic therapy, to me, that’s exciting. To this extent, the ADAPT and ALTERNATE trials are somewhat interesting because they’re trying to get to this dynamic dimension of these tumors. Another thing that’s really fascinating to me is that we are using staging or grading or biology insights from these profiles to infer the behavior of these tumors. What we really care about is the metastatic clone that could be hiding somewhere. To this extent, further studies in minimal residual disease and the detection of that residual disease with highly sensitive liquid biopsy approaches are where we should go next. I’m sure Aditya can speak about some of the investigations in this space.

Andrew D. Seidman, MD: You’re absolutely right. We’re making decisions based on a gene profile from a tumor that was resected 5 years ago, as opposed to what’s happening at the moment.

Transcript Edited for Clarity