This transcript has been edited for clarity. For more episodes, download the Medscape app or subscribe to the podcast on Apple Podcasts, Spotify, or your preferred podcast provider.
Kathy D. Miller, MD: Hi. I'm Dr Kathy Miller. Welcome to the Medscape InDiscussion podcast series on breast cancer and HER2. Today, we're going to tackle the topic of central nervous system (CNS) involvement in human epidermal growth factor receptor (HER2)–positive breast cancer, trying to understand both the biology and managing patients with this problem.
I am so excited to introduce you to Dr Nancy Lin. Dr Lin is professor of medicine at Harvard Medical School, and she also serves as the associate chief in the Division of Breast Oncology at Dana-Farber Cancer Institute in Boston, Massachusetts. Welcome, Dr Lin.
Nancy U. Lin, MD: Thank you for having me.
Miller: I know CNS involvement in breast cancer in general, and particularly in HER2-positive disease, has been a passion of yours for a number of years. Can you help us understand the scope of the problem? What proportion of patients with HER2-positive disease develop CNS metastasis?
Lin: When we look at patients who have early-stage breast cancer, fortunately, it's relatively uncommon, although we do see it. In the KATHERINE study, which looked at trastuzumab emtansine (T-DM1) post–neoadjuvant therapy for patients with residual disease, about 6% of patients developed isolated CNS recurrence as their first site of recurrence.
That doesn't sound like a lot, but it was actually more than half of the recurrence events that occurred in the trial. If we look at the metastatic setting, we do see that over time, more and more patients develop brain metastases. And so some of the older estimates are that up to half of patients will develop brain metastases over the course of their lifetime.
My colleague, Dr Sarah Sammons, has recently done an analysis of the Flatiron Health database — these are real-world data. There are over 18,000 patients with metastatic breast cancer in that database. If you look at patients with HER2-positive breast cancer, around 10% have brain metastases when they're first diagnosed with metastatic disease before they start first-line therapy.
By the time patients get to third-line therapy, about one third of patients have brain metastases in addition to the patients who had it at the outset. So we are definitely seeing it, and we're still seeing this in the modern era.
Miller: Do we know why CNS metastasis seem to be more common in the HER2 subtype than in other phenotypes of the disease?
Lin: We don't fully understand this. It does seem that there is truly a biological propensity. It's not just a matter of the brain being a sanctuary from trastuzumab; there really does seem to be some biological reason why HER2 predisposes to brain metastases. But we don't fully understand, beyond that, what that's mediated by.
Miller: I remember a study that one of my former pathology colleagues, Dr Sunil Badve, looked at trying to answer that question. They were able to find some genetic patterns that predicted early brain metastasis — so those patients that you mentioned for whom this is a first site of metastasis vs those where it's a very late phenomenon. But trying to look at CNS involvement or not just wasn't successful.
Lin: It's interesting. That's exactly right. We've also tried doing this and, like estrogen receptor (ER) status, ER-negative status predicts earlier brain metastases vs later in the HER2-positive space. But again, many people have tried putting together signatures that predict CNS relapse, but when you try to validate them for the ever vs never question, it turns out that they don't validate.
Miller: Do we know if this is getting more common over time? There have been dramatic improvements in management of the systemic disease. Does that mean we're seeing more patients with CNS involvement?
Lin: I think if we compare with the pre-trastuzumab era, the answer is yes. Remember that before trastuzumab came around for metastatic HER2-positive disease, the median survival was only about a year and a half. Patients were not often living long enough to develop brain metastases. I think that's obviously different now in the HER2-positive space, where patients live for many years.
One of the things that has been sobering when we've looked at the rates of CNS metastasis over time is that we don't really see a plateau. Patients develop brain metastases in the third-line setting, in the fourth-line setting, in the seventh-line setting, in the tenth-line setting. So it seems to be quite different from small cell lung cancer, where you see these early peaks and then if you make it past the early peak, patients are often okay. This really seems quite different. Patients remain at risk really long-term.
Miller: This feels a bit stochastic, in that if you imagine you have microscopic CNS metastasis, then the longer you live, the more likely they perhaps grow to the point where we can detect them.
Lin: I wonder if that may be part of why it seems to have been easier to identify potential signatures of early CNS relapse vs the late ever vs never question, because you think molecular features may or may not end up predicting stochastic events, but they may predict early biological predisposition.
Miller: It reminds me of a study that some of my colleagues did in the antiangiogenic therapy era where we compared the rates, and the survival of patients with asymptomatic CNS disease just found on screening with a cohort who had presented with symptomatic metastasis and their survival appeared identical.
It really suggested a stochastic phenomenon in which if the metastases happened to land in an area of important function, you presented with symptoms. Otherwise you might be found on screening.
Lin: I actually cite that paper quite frequently, because obviously this question about screening comes up all the time, both at meetings and tumor boards and with patients. Should we be screening patients with metastatic disease with serial brain MRIs in the asymptomatic situation? Your study and others have not shown survival differences associated with screening.
I think that some of the other endpoints that are important to patients are difficult to discern from retrospective studies. A number of people, including my colleague Ayal Aizer at Brigham and Women's Hospital and also colleagues in Canada and South Korea, have launched prospective screening studies where additional endpoints are collected, including neurocognitive outcomes and functional outcomes.
I think those studies are very important because obviously there are outcomes that are very important to patients beyond survival. But it's really hard; you can't go back in a retrospective study and assess neurocognitive outcomes. I think those studies will be very important to try to get at this question in a more comprehensive way.
Miller: I would love your thoughts on the surveillance and the screening because I personally have posttraumatic stress disorder (PTSD) from the screening from the antiangiogenic therapy. It was a very different era and our systemic treatments have come a long way. So the results might be very different. What was so sobering for us is that as best we could tell, none of the patients with CNS metastasis that we identified developed symptoms.
Now, all but two of them were radiated after the CNS metastasis were found. So what's tough to know is how many patients might have developed symptoms if they had not been treated.
But it appeared that they all died of progression of their systemic disease, not of the CNS metastasis. And because that then limited their therapies, because so many trials excluded patients with CNS disease, it was really hard to feel like we helped those women who we found things in.
Lin: I feel the same way and I have the same PTSD because I also screened patients for the various bevacizumab trials and also many of the HER2 trials, and particularly at the beginning, I also screened for brain metastases and those patients were unfortunately excluded from the trials. So, I hear you.
I think one of the nice things about these ongoing studies is that they do separate out the subtypes, because I think the answer to the surveillance question may be different in patients with HER2-positive disease vs those with triple-negative breast cancer, where post–brain metastases survival is very different.
Part of the motivating data for this prospective study was kind of a natural experiment where they looked at patients with non–small cell lung cancer vs those with breast cancer, and the presenting features at the time they developed brain metastases as well as the treatments they received. It's a natural experiment because the lung cancer patients are generally surveyed and the breast cancer patients are not; obviously it's two different diseases. What Dr Aizer found was that the patients with lung cancer tended to present with fewer lesions and smaller lesions and were less likely to need whole-brain radiation, et cetera.
That's part of the motivating idea behind this prospective study. The way the study was designed, there's a separate cohort for triple-negative patients because the questions and the value is going to be quite different potentially in that group of patients compared with patients who potentially have longer survival after brain metastases, such as the HER2-positive patients.
Miller: Well, there certainly have been many more advances and improvements in treatment and extending overall survival for patients with metastatic HER2-positive disease and triple-negative disease. If there is a group where surveillance might be helpful, whether it's improving survival or preventing symptoms, it does feel like the HER2-positive group might be the sweet spot where that would be useful. But I won't be doing it until you show me data that it works.
Lin: I'm with you.
Miller: We've mentioned the blood-brain barrier a couple of times, and I also spend way too much time with radiologists who point out that the edema we see around the brain metastasis means the blood-brain barrier is not functioning very well. Is it really still accurate to talk about it as a barrier in that way?
Lin: I kind of distinguish the blood-brain barrier and the blood-tumor barrier. Also, what is the goal of treatment? The reason we give gadolinium is because the blood-tumor barrier is disrupted, and tumors enhance because the blood-tumor barrier is disrupted and gadolinium can get out.
We know that the blood-tumor barrier allows molecules in that the normal blood-brain barrier would not. And we know this both for small molecules. So dabrafenib, which was developed for melanoma, was purposely engineered such that it would not cross the blood-brain barrier because they were worried about the toxicity — and lo and behold, it turns out that patients with brain metastases did respond. Not with prolonged responses, unfortunately, but they did have transient and very convincing responses.
The same is true of lapatinib, which we always sort of think of as crossing the blood-brain barrier. But in animal models, if you look across intact blood-brain barrier without a tumor in place, you don't see a lot of lapatinib crossing in through the intact blood-brain barrier.
Obviously for the antibody-drug conjugates, that's even more true because they're very large. But of course we now know that many of these drugs, including drugs like T-DM1 and trastuzumab deruxtecan (T-DXd), do lead to CNS responses. So enough is getting in that we're seeing a biological effect. I think the other question really has to do with prevention and whether drugs that get through the intact blood-brain barrier better are necessary to get the optimal prevention effect.
If we look at anaplastic lymphoma kinase–positive lung cancer, for example, the more the CNS penetration of the successive generations of drugs, the more we're seeing a real CNS prevention effect. One of the things that I wonder about the KATHERINE trial, which was that post-neoadjuvant T-DM1 study where there was not a difference in the CNS relapse rates between the two arms, is maybe we need something that gets in better in the prevention setting.
We'll have to see. It is possible that just by better control of systemic disease, you don't have seeding to the CNS and so you don't need that CNS penetration. But if there is micrometastatic disease already preexisting in the brain, you would think that the drugs have to get in in order to prevent it from growing out.
Miller: Prevention sounds a bit odd here, because we're treating people with active metastatic disease. So it feels more like we're talking about preventing outgrowth of subclinical or microscopic disease, which I would assume would mean if you screen those patients, we wouldn't see it because the area is small enough that the blood-tumor barrier is still intact. So it's not enhancing. It does raise a question of do drugs have to get there? Do they just have to control the systemic disease? It seems a difficult thing to study.
Lin: I think so. And I think that part of it is just, in terms of the way our clinical trials are designed, it's hard to get excited from a purely clinical perspective in patients who don't have brain metastases to do MRI scans on a regular basis timed with the CT scans. But from a scientific standpoint, that's the best way to understand whether drugs prevent brain metastases, because we know that if we wait for symptomatic brain metastases [to develop], that significantly underestimates the true prevalence of brain metastases.
I think the best we can do in most trials now is to look at the appearance of symptomatic brain metastases, because that's how most of the brain metastases are discovered, and try to compare arms. I do think that's an interesting secondary endpoint to include in large randomized trials, to get an understanding to some extent of whether various different interventions might delay that onset of symptomatic CNS disease.
Miller: I want to take you to some of the treatment questions that come up. Whether patients come to attention because of symptoms or because someone went looking for trouble, when you find it, you have to figure out what to do about it. So when it's first diagnosis of CNS metastases, how do you think through, do I call the neurosurgeon? Do I call the radiation oncologist? Do I just switch their systemic therapy to something that you've shown has activity? What are the considerations you're running through?
Lin: That's exactly right. I think about each of the disciplines and what their role might be. The first and usually most straightforward question is, is there a will for surgery? That's generally the patients who have a large symptomatic lesion. It could be a single lesion.
There might be other lesions, but if there's a dominant lesion that's clearly causing acute problems, that would be a reason to involve neurosurgery for an assessment. Surgery nowadays is so amazing. Patients recover unbelievably quickly for the most part. I definitely do consider that often as my first branch point.
The next question is, is radiation the optimal option in this situation or should we just switch systemic therapy? I will say, I don't make these decisions in a vacuum. I do send patients to radiation oncology, and then we discuss it as a team. But some of the considerations are, if somebody has a significant disease burden extracranially that's progressing and the intracranial disease burden is not dramatic and I'm going to have to switch their systemic therapy anyway, I will often switch the systemic therapy and just watch the brain. So, for example, somebody who's on trastuzumab and pertuzumab (HP) maintenance and has dramatic liver progression, we're going to switch them to T-DXd anyway. And we know that it has CNS activity, so I might just watch that patient and do a short interval scan. For patient who has stable extracranial disease, if there's a relatively straightforward stereotactic radiosurgery (SRS) option, and this is their first presentation, I will usually do SRS first and leave the systemic therapy alone because although we don't have great data, the data do suggest that patients probably get somewhere in the 6- to 12-month range of additional time on their initial systemic therapy before we have to make another switch. Some patients even have much longer than that.
If a patient has enough disease where we're thinking about whole-brain radiation, and they have a good systemic option (eg, they haven't received T-DXd or tucatinib before), I will discuss with those patients that whole-brain radiation would still be considered our standard of care at this point, but that I think it's reasonable to consider a systemic option to try to defer the need for whole-brain radiation, given our knowledge of the long-term toxicity of whole-brain radiation and the fact that the HER2-positive patients are the most likely to see that long term toxicity because they're the most likely subtype to survive out to that time point.
Miller: I think the group that we all fear the most and struggle with the most with are the ones with isolated CNS progression. I can think of several patients in my practice who never had systemic progression from the day I first gave them trastuzumab for their metastatic disease, and their only problem was continued progression of CNS metastasis. There were initial guidelines saying thou shalt leave the systemic therapy alone and use local modalities. But those guidelines were before we understood the activity of some of the systemic treatment options. Have the guidelines changed?
Lin: I think the guidelines are evolving, and I think that part of the evolution has to do with more clear delineation between initial management and subsequent management. What do you do when the patient first presents with their brain metastases? Let's say they have isolated CNS disease. Most people still today would give radiation if there's an SRS kind of option. But then what happens when the patient progresses again in the CNS? Do we keep radiating them 10 serial times, or do we start trying to add in systemic therapy?
The other concept that is coming more to the forefront is this concept of brain metastasis velocity. This was developed in the radiation oncology literature, but I think it's very useful. Essentially what the investigators did is they looked at patients who received SRS and then they looked at how many lesions occurred over the first year after the SRS. Not surprisingly, patients who had more lesions occur ultimately needed whole-brain radiation at higher rates than patients who had fewer lesions occur in that same timeframe.
So I borrow that as I think about systemic therapy. For a patient who has their second brain met progression event 18 months after initial SRS, and they have two lesions that are very small, I will often just send them for more radiation, to be honest, because systemic therapy also has toxicity. But that patient who has these repetitive events — on every other scan or every scan, you start seeing new lesions — it seems unworthwhile to me to chase them with SRS. Systemic therapy there, to me, makes the most sense. People often ask for prospective trials and randomized trials, but the devil's in the details.
There's so many different ways people can present. There's all these different scenarios, and so I think that for some of these questions, we're ultimately going to have to rely on common sense and expert consensus and real-world data. I don't think that that's a terrible thing to do.
Miller: Early in my career, if you had brain metastases, if it wasn't surgically resectable you got whole-brain radiation. And I think because survival in general was so much less, we didn't worry much about the toxicity of whole-brain radiation in adults. I think we've clearly seen since then that from a cognitive standpoint, whole-brain radiation is not a good thing to do to an adult, and I wonder sometimes if we've almost become so fearful of it that we try to do everything we possibly can, even when the scan and the tempo of the disease are telling us that's what's going to be needed.
Do we have better data on what the cognitive effects in adults are and over what time course we start to see those effects?
Lin: Yes; you can actually see the effects quite early on. In some of the trials, like the trials looking at memantine or the trials looking at hippocampal avoidant whole-brain radiation, even at the 3- and 6-month time points, you do see differences in neurocognitive test performance between the neurocognitive-sparing approach, whether it's memantine or hippocampal avoidance, and the other approach. It's the same with the whole-brain vs SRS studies; you can start seeing effects relatively early.
But the profound effects that we see tend to happen late. And those patients just can't function anymore. I've seen people who have been confined to a wheelchair because they have gait instability that is so profound and their ability to function in an independent way is completely decimated by the late effects of whole-brain radiation.
I have seen enough of that, that it obviously gives me pause, but those are effects that are years out. And so again, in that patient who has a very rapid tempo of disease, where you think that their survival is likely going to be a year or less, yes, there is short term toxicity with whole-brain radiation. Yes, we have to take it seriously, but there's also cognitive toxicity from CNS progression. So we have to balance that. And in that patient who has rapid progression of disease and doesn't have great systemic options, you have to control their CNS disease. That will also help their cognitive function, at least in the short term.
We're really starting to think about what is the life expectancy of that patient, and that really affects my enthusiasm or not for pursuing whole-brain radiation.
Miller: We haven't talked about leptomeningeal disease, which happens a lot when we talk about CNS metastasis. We tend to leave that bad boy off to the side. Have we made any progress in managing patients with leptomeningeal disease?
Lin: We actually have made some progress, which is kind of amazing to say, because for many years, the answer was no. From a local therapy standpoint, there's this really interesting data for proton craniospinal radiation. The role of that is still being really defined. We're not totally sure where to use it, but it does represent an improvement over the options that we had.
Then systemically, we have a number of different options. There have been now several phase 2 trials of intrathecal trastuzumab in patients with HER2-positive leptomeningeal disease (LMD). Probably the most solid endpoint for LMD trials is overall survival, because the response criteria are constantly being debated. The overall survival in the intrathecal trastuzumab trials was between 8 and 10 months. Even in HER2-positive LMD patients, if you look at real-world data, most of the studies show survival of less than 6 months. So there's something there; there were responses.
If we look at the tucatinib-capecitabine-trastuzumab regimen, that was looked at in a Translational Breast Cancer Research Consortium study that was presented at the ASCO meeting earlier this year. That study not only showed partial responses by the Response Assessment in Neuro-Oncology (RANO) LMD criteria, but also they asked patients to do quality-of-life assessments and the physicians rated their symptoms and signs. They showed not only stabilization, but actually pretty dramatic improvement in quality of life and symptoms with this regimen and a survival in the 10-month range from the time of study entry.
And now we have some data for T-DXd. Our center collaborated with Duke, and we identified a small number of patients — eight patients with HER2-positive leptomeningeal disease. Many of those patients actually did achieve clinical benefit. I have some patients with LMD who have been on it for more than 2 years, with positive cytology and good disease control. I would not have expected that, but we are actually seeing something quite interesting in that population as well. I do think that there's starting to be some movement, and at least we can offer patients some options that were previously not available.
Miller: Thank you all for joining us. Today, we've been talking with Dr Nancy Lin about CNS metastasis in patients with HER2-positive disease. This is an area where we've made huge progress and really requires a multidisciplinary approach, involving the medical oncologist, radiation oncologist, and at least for some patients, neurosurgery. This is also an area where we're going to continue to think about surveillance and perhaps that may have a role for our patients in the future.
Please take a moment to download the Medscape app to listen and subscribe to this and other podcast series on breast cancer and HER2. Dr Lin, thank you for joining me.
Lin: Thank you for having me.
Miller: This is Dr Kathy Miller, for the Medscape InDiscussion podcast.
Resources
Brain Metastases in HER2-Positive Breast Cancer: Current and Novel Treatment Strategies
Raising the Bar for Real-World Data in Oncology: Approaches to Quality Across Multiple Dimensions
Anti-Angiogenic Therapy: Current Challenges and Future Perspectives
ALK-Positive Lung Cancer: A Moving Target
Radiotherapy for HER 2 Positive Brain Metastases: Urgent Need for a Paradigm Shift
Stereotactic Radiosurgery and Local Control of Brain Metastases From Triple-Negative Breast Cancer
Medscape © 2024 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Managing Central Nervous System Metastasis in Patients With HER2-Positive Breast Cancer - Medscape - Oct 22, 2024.
Comments