Kathy D. Miller, MD: Hi. I am Kathy Miller, professor of medicine at the Indiana University School of Medicine in Indianapolis, Indiana. Welcome to Medscape Oncology Insights, coming to you from the 2016 Annual Meeting of the American Society of Clinical Oncology (ASCO).
One of the longest-running questions in breast cancer is: For how long should patients receive an aromatase inhibitor (AI) in the adjuvant setting? We finally have some data from a study presented at this meeting in the plenary session that addressed this question head-on. Here to discuss the results with us is one of the study authors, Dr Hyman Muss, professor of medicine at the University of North Carolina and the Lineberger Comprehensive Cancer Center in Chapel Hill, North Carolina. Welcome, Hy.
Hyman B. Muss, MD: Thank you for inviting me.
Backstory: A Hard Sell for Longer-term AI Therapy
Dr Miller: Tell us first about the original MA.17 study.
Dr Muss: Probably about 10 years ago or longer, Paul Goss got this idea to look at extended adjuvant endocrine therapy for women with breast cancer. At that time, 5 years of tamoxifen was the standard I am old enough to remember a National Cancer Institute alert—don't give it for 10 years—because there were several studies suggesting that longer was worse. Subsequently, Paul thought that because relapses of hormone receptor–positive breast cancer were continuous, like a low-grade lymphoma, that maybe if we had longer-term endocrine therapy—like when taking statins for long periods of time—we could do better.
He designed a wonderful study[2,3] where, after 5 years of tamoxifen, women were randomized to letrozole or a placebo. They took that for another 5 years. It had 5000 patients and showed a substantial risk reduction for women on letrozole. That really changed our mind about the duration of endocrine therapy. It was a great step forward and a great idea.
Dr Miller: Today that seems so simple, so common. At the time that study was being designed, there was not a lot of enthusiasm for it.
Dr Muss: No, it was a very hard sell. I remember as I got involved through the Cancer and Leukemia Group B (CALGB), now the Alliance, that Larry Norton and I were co-chairing the committee, and we did not know who of the young people we could assign to do this. Larry said, "Why don't you do this, Hy? Because otherwise this may go anywhere."
We did not think it would be very popular to randomize these people. They had already had 5 years of tamoxifen, and many of those women had side effects. It was a very hard sell. When we started, accrual was very poor, but for some reason—and I am really not certain why—it slowly picked up and then really accelerated.
Dr Miller: That original study taught us that 5 years of an AI after an initial 5 years of tamoxifen was helpful. That was not the end of the story; it then became the MA.17R study. What was the next step?
Dr Muss: We realized from the overview data that the relapse rate is nearly continuous. There are still relapses after 10 years for many of those woman who took the endocrine therapy. A lot of our patients were coming in at the same time and saying, "You know, I am really not having a lot of side effects with these meds. I had four positive nodes and you all told me that I was likely to do very poorly. If there is not really a problem and you do not really know the answer, why don't we continue this?" We said, "That is not a bad insight." We talked about it and subsequently designed the MA.17R to randomize patients—the majority of whom had now been on 10 years of endocrine therapy (5 years of tamoxifen, 5 years of letrozole)—for another 5 years. Those women had strength of steel to accept this subsequent randomization.
Is Longer Better? Somewhat
Dr Miller: The real crux of the design was 5 years versus 10 years of an AI. Is it better?
Dr Muss: It is better. I think we have got a lot of follow-up to do. This was approximately another 2000 patients randomized. We learned from these data that subsequent 5 years of letrozole was associated with a few percent improvement in relapse-free survival. We do not see any overall survival benefit.
I think we have to look back. If you look at the 10 years versus 5 years of tamoxifen, you do not really see the benefit in the 5 to 10 years. It took the 10 to 15 years to see the carryover.[5,6] There are a lot of people who say, "Are the patients going to survive that long?" A lot of these are older patients, but many will. In fact, what is really impressive is that in this study, the overall survival with 6 years or so of follow-up is like 95%. It is terrific. In years to come I hope this therapy will continue to decrease the rates of distant metastases and lead to a survival advantage. That few percent may be very well worth it to a lot of women—higher-risk women initially—and I think it is important.
Longer-term AI Therapy Raises Concern Over Bone Health
Dr Miller: Whenever we talk about longer-duration therapy and its benefits, we also have to think about toxicities or the harm that we might be doing. I know that a big issue with AIs is bone health. What are the consequences to the bones?
Dr Muss: That was a concern. There was probably double the fracture rate in the women who took the AIs for 5 more years versus placebo. Some of these were not fractures that you would expect with osteoporosis and bone loss. I think I am a better bone doctor, as I am sure you are, Kathy, from learning about dual-energy x-ray absorptiometry scans (DEXAs).
Dr Miller: I am not sure I am better, but I pay a lot more attention to bone health than I did 5 years ago. I order a lot more DEXAs, and I spend a lot more time looking at them and intervening than I did even a couple of years ago.
Dr Muss: There were other fractures here, as well as some osteoporotic fractures. A lot of the women in this trial, probably close to half, were taking bisphosphonates. It is interesting because that could have ameliorated the relapse rate as well. It's hard to know. But [the fracture rate] is a concern and we have to be very vigilant when we extend this therapy to monitor their bones well. It can be very challenging to do that.
Dr Miller: Other toxicities that have come up in some previous AI studies—usually ones comparing AIs to tamoxifen—were a potentially increased risk in cardiac events, and an increase in hypercholesterolemia that might be associated with those cardiac events. Were those issues seen in this study?
Dr Muss: Not really. The cardiovascular events were very small in this study. In MA.17, where we had 5000 patients and compared letrozole versus a placebo, there really was not even a hint of any increased cardiac risk. In some population studies of people taking tamoxifen—people have looked in things such as the Danish registry—there was a lowered risk of dying of cardiovascular disease. You are probably comparing an AI, which has no increased cardiovascular risk, with tamoxifen, which, in the older studies, was found to have a small cardiovascular benefit. I really think it is safe from the cardiovascular point of view.
Should Patients Who Finished AI Therapy Start Up Again?
Dr Miller: I am imagining that our listeners, who will be going to clinic after seeing this, will all face patients who finished 5 years of an AI, 6 months, a year, even 2 years ago. Are you going to be talking to those patients in your clinic about these data and whether they should restart an AI?
Dr Muss: It is a great question. We have to give women who participated in this study great credit because some of them may have been randomized to tamoxifen or not many years before. These women obviously did not get terrible arthralgia or myalgia from AIs, and they were able to tolerate it. I am sure that a small percentage were so petrified to stop that they would tolerate any toxicity to stay on.
There may be a high-risk patient with modest symptoms who stopped it 6 months ago, and you and the patient discussed it and you said, "Okay, we do not have the data yet. It is okay to stop, but I will revisit it with some." We showed in MA.17R that when we broke the code early for the relapse rate, and put women back on letrozole who were on placebo, they derived some benefit. That benefit was over a 5-year period, and some women derived benefit. It is not unreasonable. If you have a high-risk patient doing well, who stopped AI therapy within a year or two—kind of what we did with Herceptin® studies—it would be very reasonable to discuss [resuming AI therapy].
Special Considerations for Older Breast Cancer Patients
Dr Miller: I have to ask you about another interesting aspect of this study that matches nicely with one of your interests in older patients and geriatric oncology. This really has become the largest collection of data on older patients, just by virtue of when they started tamoxifen. They were on for 5 years, and they have now been on for 15 years. The average age of patients in the MA.17R study was 65. I do not think that we have another study anywhere in oncology where the average age is in the mid-60s.
Dr Muss: Very unusual. Some of the extended endocrine studies going on have older patients, because they initially took postmenopausal patients. This is really more like the real world of patients we see, because the average age of a diagnosis of breast cancer in the United States is about 62. The vast majority of women who do not survive it are 65 and older. This is really representative of the population we see, and not just here in these studies but especially in many of the chemotherapy studies where we have a paucity of older patients.
Here we have some good data looking at second cancers as we get older—your greatest risk of getting a second cancer is a prior cancer. For a lot of these patients [in MA.17R], equally balanced in both arms, a lot of the events were second cancers, not just relapses. It is very realistic and we are going to be very challenged as oncologists to pay attention to comorbidities and other cancers, in addition to giving patients the best advice about breast cancer.
Therapy on the Basis of Functional Rather Than Chronological Age
Dr Miller: When we approach an older patient in our clinics, someone at least in their later 60s, into their 70s, maybe even into their 80s, what are some of the simple key assessments that we should be paying attention to that would help us serve those patients better?
Dr Muss: Thank you for asking me that. It is a passion. One thing we have got to get used to thinking about is not age but life expectancy. The first thing is, you see someone 75 years old. I think we know that 65-year-olds are young people now. At least I hope so.
Dr Miller: They're getting younger all the time.
Asking the Right (Specific) Questions
Dr Muss: Right. When you get a 75- to 80-year-old, you are using their chronologic age to think about how to manage them. What you really want to know is their functional age and physiologic age. There are some terrific Web calculators. Just Google ePrognosis or put it in your browser and it will give you models for community-living adults. It will ask questions that you are not used to asking like, "Can you walk a block? Can you take care of yourself?"—things that medical oncologists do not always ask. It will separate that patient who may have an estimated survival of 5 years from someone who has [an estimated survival of] 15 years and is playing tennis, etc. That will change how you frame your discussion with the patient.
Dr Miller: I have seen a couple of those Web models, and what struck me is that the most common thing was how specific the questions were, which is something I have tried to teach our fellows even with younger patients when they are trying to assess someone's performance status. If you just ask, "How are you doing at home?" they reliably shake their heads and say, "Oh, I am doing fine." Sometimes they are doing fine because they have wonderful family support that is picking up a lot of the slack and doing a lot of things to keep them safe and successful at home. When you really get down to: Who does the housework? Who does the laundry? Who fixes lunch? Do you need help with getting up a flight of stairs? Do you need help with showers? They are really not doing nearly as well as the "Oh, I am doing fine" answer.
Dr Muss: You have got to ask them. Older people want to please you and me. It is still a very strong patient-doctor relationship—they are not clients, they are still patients. You have got to really ask them and you really should do some type of little mini-mental status to make sure they are with it. A lot of times you can be fooled, and what we have learned is that performance status is very crude. You do need to ask those specific questions: Can you walk and get your own groceries? Do you pay your bills? Who does the laundry? We are not used to doing that—we are used to saying, "Relapse-free survival hazard ratio is 0.62."
Matching Treatment to Patients' Goals
Dr Muss: We have got to really think of the basics in their lives and their goals—not as much with endocrine therapy, which we are talking about; but with chemotherapy, we could change someone who has a high probability of getting neuropathy from being a community-living adult into someone in assisted living on the basis of a clinical trial showing a very tiny benefit that had very few people in their age group. You are right on the money: We need to spend the time; we need to make the time.
Dr Miller: Thank you, Hy, for joining us to talk about the MA.17R trial and new information to really inform discussions about duration of endocrine therapy—discussions many of us have been having for a long time without data to guide us. And particularly for talking about your interest in our older patients.
Dr Muss: Thank you very much for having me.
Dr Miller: To you, our listeners, for joining us for this edition of Medscape Oncology Insights. This is Kathy Miller, reporting from ASCO 2016.
Medscape Oncology © 2016 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Kathy D. Miller, Hyman B. Muss. Should High-Risk Breast Cancer Patients Go Back on an AI? - Medscape - Sep 09, 2016.