Late Recurrence in ER- Positive Breast Cancer: Banking on Biomarkers

Joseph A. Sparano, MD; Kathy D. Miller, MD


January 17, 2018

Kathy D. Miller, MD: Hi . I'm Dr Kathy Miller, professor of medicine at the Indiana University School of Medicine in Indianapolis. Welcome to Medscape Oncology Insights, coming to you from the 2017 San Antonio Breast Cancer Symposium (SABCS).

For patients with hormone-sensitive disease, 5 years is not a very good benchmark. More patients have a recurrence after 5 years than in that first 5 years. Those late recurrences have been a vexing problem for decades. I've asked my next guest, Dr Joe Sparano, to give us some recent insights about who might be at greatest risk.

Joe is professor of oncology at the Albert Einstein College of Medicine and associate chair of the department of oncology at the Montefiore Medical Center in the Bronx, New York. Welcome, Joe.

Joseph A. Sparano, MD: Thank you, Dr Miller.

Late Recurrence in Hormone-Sensitive Breast Cancer

Dr Miller: Tell us how you first got interested in this problem of late relapse.

Dr Sparano: As a clinician, it's a problem I face every day in my clinic, and a frustrating one —especially for the patients who have been through the entire process of surgery, radiation, chemotherapy, and 5 years of hormone therapy. Generally, people celebrate after 5 years because they feel that they have completed their therapy.

Unfortunately, it is becoming more and more appreciated that patients who have estrogen receptor (ER)-positive breast cancer are at risk for late relapse beyond 5 years. Recent studies[1,2,3,4] have shown that there is some small benefit by continuing endocrine therapy beyond 5 years, but there are associated side effects. We really need a biomarker to identify who might be most likely to benefit from [continuing endocrine therapy beyond 5 years].

Dr Miller: The problem of late recurrence got elevated in our consciousness with those studies of longer-term hormone therapy. We still see patients who think at 5 years that they are done and home free. It's always a difficult discussion when we remind them that their risk is not zero and that it continues for a long time. For many of those patients, longer- duration hormone therapy is a tough sell. There are costs and toxicity. The benefits are pretty small.

Dr Sparano: Yes, absolutely. Two forces are sort of converging. One is that we have known for at least 20 years about the risk for late relapse. In fact, one of our colleagues, Tom Saphner, published a study in the Journal of Clinical Oncology[5] illustrating that possibly for the first time. More recently, a meta-analysis[6] was published demonstrating that on a large- scale basis, and really nailed it down quantitatively in terms of what the risk is.

We thought that we could deal with this issue by developing better therapies to use up-front. Those have been helpful in reducing early recurrence, but not necessarily affecting later recurrence.

Dr Miller: That has particularly been true of chemotherapy in our hormone-positive patients. It reduces that early peak of recurrence, but has very little impact on those recurrences that happen after year 5.

Dr Sparano: Absolutely. One of the key issues is that there is a large subset of patients with ER-positive disease who do not benefit at all from chemotherapy. Trying to find the sweet spot has been the challenge.

What Drives Late Recurrence?

Dr Miller: Tell us about the biobank that you helped develop to try to tackle this question of what drives late recurrence and who might be at risk.

Dr Sparano: This idea started about 5 or so years ago, as we all became more concerned about the issue of late recurrence. Our colleagues at ECOG-ACRIN came up with the idea of leveraging existing federally funded trials in patients who were being followed long-term and collecting biospecimens from patients who were without evidence of recurrence for 5 years.

Fortunately, we were able to leverage funding from both the BCRF (Breast Cancer Research Foundation ) and the Komen Foundation, to support the establishing of this biobank. We are now starting to bear some of the fruits of that effort.

Evolution over time may be driving late recurrence.

Dr Miller: It's a very pragmatic approach. Many of the costs are decreased because these patients are being followed anyway as part of existing trials. I would assume that those existing trials also collected their initial tumors when they enrolled in the parent trial, so we will have the potential to look at predictors of late recurrence even from the beginning and also whether things change. Evolution over time may be driving late recurrence.

Dr Sparano: Yes, we are pigeonholed into this notion that there is one decision point in a patient's cancer trajectory. That decision point is at the time of diagnosis as it relates to their surgery, radiation, chemotherapy, endocrine therapy, and anti-HER2 therapy. With our ability to evaluate tumor biology through blood sampling, either through circulating tumor DNA or circulating tumor cells (CTCs), we now may be reaching a point to get prognostic and possibly predictive information. Perhaps we should start thinking about a second decision point 5 years after diagnosis to further tailor therapy.


Dr Miller: You are presenting some of the first results[7] of using samples from this biobank. Tell us about those first analyses.

Dr Sparano: We leveraged samples from an existing clinical trial, the E5103 trial, which evaluated the role of bevacizumab as a component of adjuvant therapy, added to chemotherapy. The trial did not meet its primary endpoint, but this is an example of how clinical trials can provide very useful information even when they may not meet their primary endpoint.

We took patients who were recurrence-free at about 5 years and collected a biospecimen from them, mainly from the blood, and included the CTC assay. We accrued 547 patients over a period of about 3.5 years. About a year after we closed the study to accrual, we took our first look at the data.

Dr Miller: The idea that someone 5 years from diagnosis—after that full 5-year standard therapy—would have a CTC detectable is a scary thought. Many of our listeners must imagine that we did not find CTCs. How common was it, and did it matter?

Dr Sparano: In patients with metastatic breast cancer where the assay is commercially available and cleared by the US Food and Drug Administration, we know that about 60% of patients have detectable CTCs, and that the presence of a high burden of CTCs is prognostic. That is point number 1.

Point number 2 is that we know from earlier- stage disease in patients who have had surgery but have not yet had adjuvant chemotherapy that about 20% of patients had detectable CTCs.[8]

We know that the presence of those cells provides independent prognostic information. There is clinical validity with that association, but not clinical utility, because there's not really much we can do to act upon it.

First Results of Biobank

Dr Miller: At that point, those patients who have just had surgery have not yet had any systemic therapy. You might also wonder if those patients who have a CTC at that point are the ones who really need our systemic therapies. [They are] hopefully the ones who derive greater benefit. But this study looked 5 years down the road. Could we still detect CTCs at that time point?

Dr Sparano: Our hypothesis was that we would see a detection rate of about 5 %-10%. We designed the study to distinguish that versus a rate of less than 1%. We built in an early stopping rule to stop if we had a very low detection rate of less than 1%.

When we did our interim analysis, we found that it was not less than 1%, so we were able to continue. When we did the final analysis, we found that about 5% of patients had detectable CTCs, which was the first study objective—to get a sense of what that prevalence was at that particular time point.

Dr Miller: Were there any obvious clinical factors, such as more extensive disease biologically or more aggressive disease that predicted finding a CTC at roughly year 5?

Dr Sparano: The simple answer is no. The more complex answer is when we did a multivariate analysis looking at tumor size, nodal status, grade, ER expression, or use of endocrine therapy in patients who had hormone receptor-positive disease, there were no differences between the CTC-positive and the CTC-negative groups. They could not be picked out on the basis of clinical features.

Risk Associated With Positive CTCs

Dr Miller: This potentially makes the assay more important. It's frightening for someone who sees patients in clinic to think that we cannot distinguish on clinical grounds that ongoing risk. What was the risk associated with finding a CTC?

Dr Sparano: We were expecting there to be an increased risk—that was the hypothesis driving the study. The level of risk stratification was really pretty striking. There was a 20-fold higher risk for recurrence in those who were CTC-positive compared with those who were CTC-negative. The follow-up of the study was relatively short for ER-positive disease, but substantial, in that we could estimate 2-year recurrence rates. The positive predictive value for a positive test was 35% at 2 years.

Dr Miller: For women with a positive test, 1 in 3 had recurrence within the next 2 years.

Use of Biomarkers to Guide Therapy

Dr Sparano: Yes. That is obviously a very important finding and raises the question as to whether we should be using this as a biomarker to identify high-risk patients to possibly test other strategies to reduce risk.

Dr Miller: For the second decision point you mentioned, maybe this is a test you do at that point to identify patients who [need] secondary adjuvant therapy or some other intervention to reduce that risk.

Dr Sparano: I think that is a very important component of it. Perhaps an equally or more important component is the negative predictive value of the test. The negative predictive value was 98%, meaning that only 2% of patients who had a negative test had a recurrence by 2 years. This means that we might be able to use this as a biomarker to spare the continued use of extended adjuvant therapy.

Bear in mind that in the low-risk registry of TAILORx,[9] the 5-year risk for distant recurrence was 1%, which was deemed sufficiently low enough that clinicians and experts felt they did not need to do a randomized trial in that setting to demonstrate that you did not need chemotherapy.

Here we are dealing with a different situation. We are not dealing with a toxic chemotherapy regimen. We are talking about continuing an existing regimen that someone has been on—an orally administered drug. But many patients do have issues tolerating the drugs.

Dr Miller: There are toxicity and bone health concerns.

Dr Sparano: Yes. I think patients would [take] the opportunity to stop the therapy, if they felt they could do it in a safe way.

Establishing Clinical Utility

Dr Miller: These really are striking results, but not yet ready for people to go home and start looking for CTCs in their long-term follow-up patients. Where do you take this next to get us to that point?

Dr Sparano: You are absolutely right. People get confused with clinical validity. This study provides level 1 evidence showing clinical validity —that is, the association of this biomarker with recurrence. But that does not always translate into clinical utility. Does the information provided by the test provide additional prognostic information? Does treatment change as a result of that? Do patients benefit from that change, which is the most important factor?

We are on solid ground with the clinical validity part, but we will need to do more studies to establish the clinical utility.

Dr Miller: I describe this to patients as, "If you do the test now, you will tell me to worry more or worry less. But you will not tell me what to do about it that would change how much I need to worry about you." The next step is to look at interventions in this group, and other biomarkers because we have a lot more material in the bank.

Dr Sparano: This may not be the only biomarker. First of all, we only assess the CTCs at one time point. The assay may have greater negative predictive value if we assay at serial time points, say after the completion of endocrine therapy. We can use it in combination with other biomarkers, such as circulating tumor DNA, looking for specific mutations (for example, in ESR1 ) that might indicate whether patients may respond better to a different type of antiestrogen therapy.

There is quite a bit of work to do. The results of the study lay a foundation for doing that work in a more enlightened way.

Dr Miller: Thank you, Joe, for joining us. The biobank is going to teach us so much over the next several years. Thank you for coming to share these first results.

Dr Sparano: My pleasure. Thank you for the invitation.

Dr Miller: And to you, our audience, thank you for joining us as well. This is Dr Kathy Miller, reporting from SABCS.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.