For Low-Level HER2 Breast Cancer, Trastuzumab Is a No-Go

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. We are coming to you today from the 2017 San Antonio Breast Cancer Symposium (SABC).

The introduction of trastuzumab revolutionized treatment for patients with HER2-positive disease. But those early trials set an arbitrary threshold for considering a tumor in a patient to be HER2 positive. Did we pick the right threshold? That was an unknown question and one that the NRG Research Group tackled in their National Surgical Adjuvant Breast and Bowel Project (NSABP) B-47 trial.^[1]

To bring us the results of NSABP B-47, I'm joined today by Dr Charles Geyer. Charles is professor of medicine at the Virginia Commonwealth University and associate director of clinical research at the Massey Cancer Center in Richmond, Virginia. Welcome, Charles.

Charles E. Geyer Jr, MD : Thank you, Kathy. I appreciate the opportunity to talk about our trial.

Challenges With Diagnostic Tests

Dr Miller: Take us to the background. Why look at trastuzumab in HER2-negative patients?

Dr Geyer: At the time that adjuvant studies were launched, everyone recognized that the real gains would probably come in the early setting, based on the very exciting data that we saw in metastatic disease. But there were issues and a lot of uncertainty about cardiotoxicity. A lot of emphasis was placed on trying to define the population who would benefit.

In the metastatic trials, things were a little softer. They used broader eligibility criteria: 2+ by immunohistochemistry (IHC), fluorescence in situ hybridization (FISH) positive. In those studies, it appeared that the benefit was going to be definitively occurring if it was going to occur—it would occur in those who were amplified for HER2 or who had very strong IHC staining. The trials were set up to target that group.

The critical thing, though, was that these were new diagnostic tests. Technical aspects were still being worked out. Different groups used different approaches. Some like Dennis Slamon's group required central testing with one pathologist, Mike Press, doing it.

Dr Miller: Some even used different antibodies.

Diagnostic Testing in NSABP B-31

Dr Geyer: Different antibodies, different things. NSABP has always been a bit more community-based and pragmatic in our approach. We basically set up NSABP B-31,^[2] our adjuvant study, to rely on the testing being used at the sites, but we recognized that because of the uncertainties around testing and the efficacy of the therapy, we would want to do central review.

The study was written. After 500 patients were entered, we did central testing. We found that 16% of patients that were found to be HER2 positive at sites were HER2 negative in our hands.^[3,4] Sun Paik's lab did that work. We amended the protocol to try to reduce that number. It did not eliminate it, but it dropped about 6%.

At the end of the trial, 9.7% of the patients did not have centrally confirmed HER2-positive disease. Much to our surprise, when we looked at whether there was any benefit, it appeared that the benefit from trastuzumab was as great in those patients who were centrally HER2 negative as in the HER2-positive cohorts.

Dr Miller: I remember the discussion about that. I think a similar thing was seen in some of the other studies that used different testing strategies. There was a lot of talk about why that might be, with some pretty fanciful hypotheses: that this was tumor heterogeneity; that maybe a different part of the tumor was tested.

You mentioned testing difficulties with these new tests. Whenever there are discrepant test results, every pathologist you talk to is absolutely convinced that their result is the one that is correct. We do have to acknowledge that in some of those, maybe the central lab got it wrong and the local lab was correct. Dr Paik would argue. Every pathologist is convinced they are right.

Dr Geyer: There was a lot of work done initially. The groups got together and the pathologists doing the central testing swapped slides to look for concordance amongst the specialists. They did additional ways of assessing HER2 on these HER2 normals, and they did have very low microRNA. They did look truly HER2 negative in multiple ways.

The other thing that is still of interest is that you would think that if the amount of HER2 positivity was important, patients with higher HER2 amplification levels ought to get progressively greater benefit. When you look at the hazard ratios in those studies, they line up. There is no interaction between the effects of trastuzumab and the number of HER2 copies or the ratio. However you want to look at it, it's fixed.

Dr Miller: It's almost like a tipping point. Above a threshold, you get benefit; below the threshold, maybe not.

Dr Geyer: That is what it would be. The thing that we struggled with, and I still do when seeing the results of NSABP B-47, is that biologically there tends to be continuums and not sharp cut-points. In our study, you might think it could be statistical noise or other reasons. But Edith Perez's group saw basically the same thing in N9831.^[5] Because of that uncertainty and the clear significant efficacy of trastuzumab, we felt we needed to do a study looking at this lower end. There were reasons for the cut-point, but at some point you just have to pick something and go with it.

Dr Miller: The reasons, as you mentioned, were shaded towards, "If this works at all, these are the folks it's going to work in." There was less concern about potentially leaving benefit for a few patients on the table, and then missing benefit because too many people were included.

Dr Geyer: Also, the safety signal. That was a lot of what went on. Subsequently, there were other theories. Some suggested that there were stem cells that were HER2 positive.^[6] I guess the pragmatic side is that this is an important question. There is a large relative risk reduction from a relatively nontoxic agent. The number of patients who benefit may go from 15% or 20% up to 60%. We really need to know this. We thought a trial was a reasonable way to proceed.

NSABP B-47

Dr Miller: Let's get to the details of NSABP B-47. How many folks participated?

Dr Geyer: Total accrual of women participating in the trial was 3270. We had learned some things. NSABP B-31 excluded node-negative patients; N9831 included them. So NSABP B-47 included higher risk node-negatives. If you were ER negative, you had to be T2. If you were ER positive [and] if you were T2, you had to be high-grade or have a high recurrence score. Or the patient had to be node positive. As it turned out, 80% of our patients were node positive. We were successful in accruing what would seem to be a higher-event-rate group of patients when you look at the characteristics table.

Dr Miller: Were all these folks what we would consider HER2 negative or did you try to find low-level expressers?

Dr Geyer: Patients had to have some HER2 presence on their cell. It had to be IHC 1+ or 2+. We did not include the zeros. We decided to focus on where there was at least some evidence of the target on the cell surface. The zeros seemed to be a bridge too far for us.

Patients had to be unequivocally negative on FISH; they had to be < 2, and < 4 copy numbers. We did not want to get into the equivocal areas. We would not have had enough of the patients we wanted. We were trying to be definitive and [include] what everyone would agree would be considered HER2 negative. That is what we did.

Dr Miller: Do not make us wait any longer. Is this going to change practice?

Dr Geyer: It will inform practice. It really does eliminate the idea that trastuzumab might have a role for HER2-negative patients. The study was unequivocally negative. I had joked, as we were putting it together, that I wanted us to do yellow and blue lines so we could see one green line.

There were no trends, which a lot of us thought we might see. Assuming we had a continuum of levels of FISH expression and IHC, which is what we expected in a large study, we thought maybe the 2+ might have something where the 1+ do not. When you look at the forest plots, everything lines up. There is just no benefit in this patient cohort in this prospectively defined study.

Dr Miller: Were there any new or different safety issues in this population?

Dr Geyer: No. We were successful. We purposely left it to the investigator to decide whether they wanted to use non-anthracycline- or anthracycline-based regimens. We learned from NSABP B-31 that patients who had ejection fractions of 50% to 54% had more heart failure. Older and hypertensive patients [were excluded]. We put in some things that we felt we needed to exclude for safety. We actually saw less cardiotoxicity in this study than we saw in NSABP B-31. No new signals.

Dr Miller: Where do we go from here in our patients with HER2-negative disease? We have seen a succession of adjuvant trials. Adding gemcitabine or adding capecitabine did not move the bar. Adding trastuzumab did not move the bar. Adding bevacizumab did not move the bar. We seem to be stuck and not making progress in this large group of patients. Where are we headed next?

Dr Geyer: Of course, you are speaking to HER2 negative, the triple negative.

Dr Miller: The HER2-negative group. The ER group is a bit different; we have made some inroads there. We have thrown a lot of things at the triple-negative group which have not made any impact.

Dr Geyer: There is a lot of interest in the I-SPY^[7] results, which are moving checkpoint inhibitors into that space. We will see. I have a lot of enthusiasm for the immune therapies. It's not just finding a better drug in a category where we have had a number of drugs; it's a new approach. We have all thought about agents to help release the immune system's activities. I'm very excited about that. We are launching a study with GBG collaborating on a neoadjuvant study in triple-negative disease looking at atezolizumab. Others are doing the same things.

Can We Learn More From NSABP B-47?

Dr Miller: We will bring you back in a few years to talk about that study. One last question for you about the NSABP B-47 trial: I assume you collected the tumors, so we may still learn things from this large effort?

Dr Geyer: Yes. We thought the correlatives would be extremely important in this. We were kind of thinking that we might find a cut-point. Was that first cut-point right? It appears that the answer is yes at this point. That was an important part of the study. We are now thinking, with those lines so on top of each other, what could we learn?

Dr Miller: The samples might be useful for things that were not related to the intention of the study.

Dr Geyer: In terms of correlative studies that might find a subset that was benefiting, I do not have a lot of optimism.

Dr Miller: It seems unlikely, given how close the results were.

Dr Geyer: They were very close.

Dr Miller: It was an important hypothesis. They were important data even though it did not turn out the way you were hoping.

Dr Geyer: Positive studies are always much more exciting. There are trials that need to be done. The thing that always causes a lot of difficulty is a trial that does not really answer [a question]. Had we had an insignificant trend, I would have felt much more like I was not sure about things. But this gives an answer and provides information.

Dr Miller: Oncology lost one of our heroes in Don Coffey this year. He was famous for teaching all of us when we trained with him that you should never assume those things that you can prove. This was an assumption and you proved it. Thank you for doing that. Thank you for joining us today to talk about these results. And thank you to our audience for joining us as well. This is Dr Kathy Miller, reporting from SABCS.