ATLAS: No Shrugging These Results (Transcript)

Introductions

Kathy D. Miller, MD: Hello. I am Kathy Miller, Associate Professor of Medicine at the Indiana University School of Medicine in Indianapolis. This is Medscape Oncology Insights, coming to you from the 2012 San Antonio Breast Cancer Symposium. Joining me today is Dr. Hal Burstein, Associate Professor of Medicine at Harvard Medical School and a medical oncologist at Dana-Farber Cancer Institute in Boston. Welcome, Hal.

Harold J. Burstein, MD, PhD: Thanks for having me.

Biggest News: Longer Tamoxifen Duration

Dr. Miller: I wanted to talk about some of the major clinical trial results at this year's meeting that could really change practice. Perhaps the biggest news is that 5 years of tamoxifen may not have been the final answer.

Dr. Burstein: Adjuvant endocrine therapy has been moving toward longer courses of therapy. We saw evidence of this in the MA.17 trial,^[1]reported almost a decade ago. Women who had finished 5 years of tamoxifen continued on nothing or on an aromatase inhibitor, and it was shown that switching to an aromatase inhibitor after 5 years made sense and lowered the risk for recurrence.

An open question about the right duration of tamoxifen goes back 25 years. In the early days of the development of tamoxifen, 1 or 2 years were used. They showed that 2 years was better than 1 year, and then that 5 years was better than 2 years. In the extension study, NSABP-B14,^[2] they took women who had been on 5 years of tamoxifen and who were free of known recurrence and randomly assigned them to ongoing treatment with tamoxifen or discontinuation of therapy.

It was interesting that in the NSABP (National Surgical Adjuvant Breast and Bowel Project) results, stopping tamoxifen was better than continuing tamoxifen. We thought we had maxed out tamoxifen -- that we had reached the 5-year landmark, and that was all the benefit that could be expected from tamoxifen. That's why we had trials like MA.17, which built on 5 years of tamoxifen.

But along the way, concerns were raised about cumulative toxicities, and in particular endometrial cancer and risks for thromboembolic complications. We had been in the habit of thinking of 5 years as the optimal duration, but percolating in the background were some worldwide studies; two ongoing trials were the aTTom trial^[3] and the ATLAS study.

The ATLAS study,^[4]which we heard about this week in San Antonio, looked at women who had been on tamoxifen for 5 years and were free of known recurrence. They were randomly assigned to either stop taking tamoxifen or to continue tamoxifen. The data suggested that women who continued tamoxifen actually did better, had a lower risk for recurrence, and had better breast cancer survival, now extending out to more than 10 years after their original diagnosis.

At Last, ATLAS Results

Dr. Miller: We heard about the ATLAS trial at this meeting a couple of years ago in a presentation that was remarkably free of data.^[5] There was only 1 number in that presentation: a hazard ratio of 0.88. Did we see convincing data this year?

Dr. Burstein: We have heard more impressive data now. The study was also published online this week in The Lancet,^[6] so you can look up the numbers yourself.

It was a very large study that accrued more than 15,000 patients. That group was whittled down to about 6800 women on whom they reported data. Many women either did not have estrogen receptor (ER)-positive breast cancers or had tumors for which hormone receptor status could not be characterized. Furthermore, the study originally permitted a shorter duration of tamoxifen before randomization. So, patients could have had 2-3 years of tamoxifen before randomization. By the time they got rid of those patients and limited it to ER-positive disease, they were left with only about 30%-40% of the original cohort. Having said that, it is still 10 times bigger than the NSABP extension study.

Dr. Miller: And this study was done in the population that we are interested in. It never made sense to address this question in tumors that were estrogen-negative.

Dr. Burstein: Nothing in the overview or anywhere else points to the idea that tamoxifen would benefit hormone receptor-negative breast cancers.

Dr. Miller: On Monday, when you are back in your clinic at Dana-Farber, and your premenopausal patient who has taken 4.75 years of tamoxifen comes in, would you tell her to continue?

Dr. Burstein: It is certainly tempting to. We have gotten into this habit of prolonging the course of the antiestrogen therapy for a while now. The MA.17 data suggest taking 5 years of tamoxifen and then continuing with an aromatase inhibitor.

In my practice, I use a lot of sequence treatment. I give 2-3 years of tamoxifen and then switch the women to 5 years of an aromatase inhibitor. But the question, particularly for premenopausal women, has been what to do, and now we have some data. For a reasonably high-risk, premenopausal woman -- the 37-year-old who had node-positive breast cancer, got chemotherapy at age 32, continues to be premenopausal, and finished 5 years of tamoxifen -- for that patient, these data are compelling, that maybe they should continue tamoxifen for a longer period.

The good news is that for young premenopausal women who have hardly any major side effects, the risk for endometrial cancer is very low. We are selecting a group of women who have already taken 5 years of tamoxifen, with few side effects.

Dr. Miller: That is excellent news, and practice-changing for that important subgroup.

Dr. Burstein: It begs several questions. Are 10 years of tamoxifen as good as 5 years of tamoxifen and then 5 years of an aromatase inhibitor? We don't know the answer to that.

The other thing that should not be inferred from this trial is that 10 years of an aromatase inhibitor is the new standard. Some people are looking at this and saying, "Well, MA.17 says 10 years is good, ATLAS says 10 years is good, so I am going to leave women on an aromatase inhibitor for 10 years." That is problematic for a couple of reasons. We don't have the data, and it could be that you only need 5 years for an aromatase inhibitor. In many women, the aromatase inhibitors cause substantially more side effects, and we have all seen patients with bone and joint symptoms, osteoporosis, genitourinary symptoms, and so forth. Before committing a woman to 10 years of that drug, I would like to see more data, and there will be studies looking at that question.

Dr. Miller: Two large studies already finished enrolling patients a few years ago, so those answers will be coming.

Does Local Recurrence Merit Systemic Therapy?

Dr. Miller: Another, perhaps not terribly common scenario, but one that we all struggle with, is the woman who has an isolated, local, regional recurrence: either a breast recurrence after breast-conserving surgery, an isolated chest wall recurrence, or isolated axillary recurrence. These patients receive additional local therapy, but should they receive additional systemic therapy? Is this analogous to adjuvant therapy?

Dr. Burstein: This is an area where we have had no data. People were very interested to see the results of the CALOR trial,^[7] which was an effort between the Breast International Group, IBCSG (the International Breast Cancer Study Group), and the NSABP. The study included about 160 women, all of whom had had isolated regional recurrence such that they could be surgically debulked. These women all had been rendered free of cancer recurrence. If they had recurrence in the breast, they had a mastectomy; if it was in the chest wall or the skin, it had been excised. They were randomly assigned to either chemotherapy or no chemotherapy, and they received appropriate endocrine therapy. Women with HER2-positive breast cancer received trastuzumab.

There was a compelling difference in disease-free survival favoring the introduction of chemotherapy. In fact, that translated to an emerging difference in overall survival. Perhaps not surprisingly, most of that benefit was in women with estrogen receptor-negative cancers, but this is the first study we have seen in a very unusual patient population, but it is not unique. In the ATLAS study, one third of events were in ipsilateral breast cancer recurrences. If you follow women for long periods, you see this. This establishes a principle that a second go-around of adjuvant chemotherapy can be appropriate for patients in this very unusual circumstance.

Dr. Miller: This trial doesn't address the patient with a contralateral breast primary tumor 2, 3, or 5 years after her initial therapy. We have similar struggles and similar discussions in those patients.

Dr. Burstein: Absolutely, and you have to look at the tumor biology. Is it going to be sensitive to a hormone, did it arise while the patient was on adjuvant therapy, such that it is not sensitive to a hormone, when did the patient last receive chemotherapy, has chemotherapy already been maxed out? Is the patient already on an anthracycline and a taxane, or is there an opportunity to introduce a new class of drugs?

The CALOR study didn't really address that. It ended up being a smaller study than they had planned. They planned on about 1000 patients, but after running it for 10 years, they had nowhere near the accrual they wanted to get, so they decided to analyze by a median follow-up endpoint of about 4 years.

Dr. Miller: If you have an international study for a problem that is not so rare, 10 years, and 160 patients, a huge selection bias goes into who enters that trial. From looking at the patient population they enrolled, does it look like the population that you see in those situations?

Dr. Burstein: They did a pretty good job, even though it was a small number of patients. In fairness, people weren't randomizing patients because we all thought we knew the answer. The answer was to give chemotherapy. Now, that turned out to be correct, but this is the first real proof we have had of that.

Dr. Miller: Some of us thought we knew the answer, and it was not to give chemotherapy. So for some it was confirmation, and for others it was a change.

Dr. Burstein: It's why we still need to do randomized studies, even if they are small, but well-designed ones.

Bevacizumab: Gone for Good?

Dr. Miller: There were 2 other studies that were not necessarily practice-changing because their results were negative. We have this ongoing issue of what, if anything, we should be doing with bevacizumab. Two trials have looked at this now; one in the metastatic setting, and the other in the adjuvant setting. Let's talk first about the LEA trial.^[8]

Dr. Burstein: Talking to you about bevacizumab is surely taking coals to Newcastle, because you have been a pioneer in this field, but data are still coming in from trials that began several years ago in the metastatic setting.

The LEA study looked at endocrine therapy with or without bevacizumab. Patients could be offered letrozole or fulvestrant, then were randomly assigned to receive bevacizumab or not, and the endpoint was progression-free survival. There was a modest difference in progression-free survival, on the order of about 3 months -- very consistent with what we have seen in other studies, with no suggestion of a signal for overall survival improvement. This will be interpreted, as have many of the studies in the metastatic setting, as suggestive but not definitive, and probably not enough to rescue bevacizumab for breast cancer.

Dr. Miller: Like all of the other studies in an unselected population, there was not enough benefit to recommend use.

Dr. Burstein: We always make these tributes to the idea that we need a biomarker. There must be a marker that predicts who benefits, but it has been hard to find it so far.

Dr. Miller: One suggestion for how to make this work is to move to the adjuvant setting, where resistance pathways aren't as well established, and perhaps focus on patients with more aggressive biological disease -- particularly triple-negative disease, where this might be important. And that takes us to the BEATRICE trial.^[9]

Dr. Burstein: Your data and others in the metastatic setting suggested that the gains with bevacizumab might be more profound in the triple-negative setting. In the BEATRICE study, an adjuvant trial of higher-risk, triple-negative breast cancer, patients were offered chemotherapy with or without bevacizumab. The bottom line is that the introduction of bevacizumab does not have an impact on disease recurrence or overall survival. This was a bit of a disappointment, because if we believe the angiogenesis hypothesis, we might think that the adjuvant setting is the place where bevacizumab should work, before the blood vessel supply is established. At the moment, however, there is no strong signal.

Dr. Miller: If there was a biomarker, would it be possible to revive this agent, or are there just too many negative studies, too many bad feelings, and too many concerns for this drug to come back?

Dr. Burstein: My guess is that bevacizumab is not coming back in breast cancer. The field has moved on in other directions. We have exciting opportunities in mTOR (mammalian target of rapamycin) pathway inhibition. There were some very provocative data^[10] here on CDK46 inhibitors in the metastatic setting for ER-positive breast cancer. HER2-driven breast cancer has taken off in its own direction.

If there is a robust biomarker out there, then people haven't found it yet, because it's been tough to see. It could happen, however. Look at colon cancer, where the whole KRAS story has flipped on its head, the way we thought about using some of these interesting biologic drugs on the basis of well-done, retrospective analyses.

Trastuzumab: 1 Year Is Good, 6 Months Is OK

Two more studies that I want to mention deal with the duration of trastuzumab question. Some of these were presented earlier this fall at the European meetings.

Dr. Miller: This has been an important question, because most of the original trials used 1 year, a nice round number, but one that was admittedly completely arbitrary.

Dr. Burstein: Yes, I used to joke that the scientific rationale for the duration of adjuvant trastuzumab was how long it takes for the earth to orbit all the way around the sun -- which is a true scientific observation, but has very little to do with breast cancer.

Dr. Miller: There isn't a better explanation.

Dr. Burstein: There isn't, but it turns out miraculously, that was correct. We saw 2 results; one from the HERA investigators,^[11] a study that had looked at zero vs 1 year of trastuzumab, and then had a third group that received 2 years of trastuzumab. No benefit was found for 2 years of trastuzumab beyond that achieved with 1 year.

Dr. Miller: That result was the clean one. There was no hope of a trend anywhere in sight that longer therapy was better. Shorter was messier.

Dr. Burstein: It was more complicated. The French consortium put together the study they called PHARE,^[12] and in that study they compared 6 months vs 12 months of trastuzumab. Technically, this was a noninferiority study, and it did not prove noninferiority.

Dr. Miller: Correct. But it did not prove inferior.

Dr. Burstein: Right, but if you look at the curves numerically, there was about a 2% benefit for the 1-year group over the 6-month group and hardly any difference in toxicity. The French group internally decided -- and I agree with this -- that given the safety of the drug and the lack of other compelling reasons, it still seems that 1 year should be the recommendation.

Having said that, these data inform some patient issues, because occasionally we see patients who, for one reason or another, can't get a full year of trastuzumab. Now we can point to these data and say, 6 months is getting you almost all of the benefit, you don't lose a lot, but it probably will remain the standard.

Dr. Miller: Two other large studies are completed, looking at shorter vs longer duration of treatment, and even shorter, looking at 9 weeks vs 12 months in one study. These other 2 large studies, each with about 3000 patients, are the results we are waiting for. I wonder whether these results will be interpreted and put in practice differently in different regions of the world.

Dr. Burstein: It's going to be a cost issue in some parts of the world. The back-of-the-envelope calculation suggests that 6 months costs about one half that of 12 months of trastuzumab.

Dr. Miller: From a public health perspective, you could treat twice as many patients for almost exactly the same gain. If I were a public health minister in a developing country, I might wonder whether I wouldn't do my entire population more good by treating twice as many people for 6 months.

Dr. Burstein: That is an interesting way of thinking about it. We also know from the older FinHer study^[13] that even 3 months of chemotherapy is effective. So the marginal benefits of these longer durations are probably pretty modest from a statistical point of view. In the United States and Western Europe, 1 year is the standard, and it's here to stay.

Dr. Miller: I would agree. A year, maybe shorter in some patients or some situations -- and certainly, there is no reason to be concerned if someone needs for their own reasons to stop short of 1 year.

Dr. Burstein: Absolutely.

Dr. Miller: Well, Hal, it has been a pleasure. Thank you for joining us today, and to you, our audience, thank you for joining us for this edition of Medscape Oncology Insights.