Pertuzumab and the Curious Case of Accelerated Registration

Kathy D. Miller, MD; Clifford Hudis, MD; George W. Sledge, MD


February 09, 2015

This feature requires the newest version of Flash. You can download it here.

Editor's Note: Should neoadjuvant studies serve as a route to accelerated registration? The intention to make active drugs available to patients faster is commendable. Yet, ironically, accelerated registration can hinder patient accessibility to viable therapies when drugs, such as pertuzumab, are approved for use in the neoadjuvant setting—before mastectomy—but not after. Three noted leaders in cancer research—Drs George Sledge, Clifford Hudis, and Kathy Miller—discuss the viability of accelerated registration and go in search of loopholes in the labeling of pertuzumab that allow patients to receive the drug in the adjuvant setting.

Is the Neoadjuvant Setting a Viable Proving Ground?

Kathy D. Miller, MD: I'm Kathy Miller, professor of medicine at the Indiana University School of Medicine in Indianapolis. Welcome to this edition of Medscape Oncology Insights, coming to you from the 2014 San Antonio Breast Cancer Symposium.

The neoadjuvant setting increasingly has become a testing ground for novel therapies in the hopes that this will allow us to show efficacy and avoid a much larger and lengthier process of testing in the adjuvant setting. But is this really a viable strategy? Can or should those neoadjuvant studies serve as a route to accelerated registration?

To discuss this provocative issue, I have invited two leaders in breast cancer research to join me: Dr Clifford Hudis, chief of the breast cancer medicine service at Memorial Sloan Kettering Cancer Center in New York, and Dr George Sledge, professor of medicine and chief of the division of oncology at Stanford University, and professor of medicine and chief of the division of oncology at the Stanford Women's Cancer Center at Stanford. Neither of you are short of opinions on this subject. What do we learn in the neoadjuvant setting?

Clifford Hudis, MD: We may be asking a lot out of a relatively small set of data, but we do learn a good bit in the neoadjuvant setting, just not everything. The neoadjuvant setting is a very efficient way to potentially screen drugs for activity. It's a very important and efficient way to do early translational and correlative science so that we can understand what works, why it works, and why it doesn't work.

But in your introduction you alluded to the possibility of also learning from that setting what works long-term for disease control, leading to improved survival. There, the jury is still more out than in on whether that is an effective strategy.

Dr Miller: The US Food and Drug Administration (FDA) jury is in, at least with their draft guidance,[1] and they have now given accelerated approval to at least one agent on the basis of neoadjuvant testing.[2] George, are you comfortable with that as a strategy to accelerated approval?

George W. Sledge, MD: I'm comfortable with it as a hypothesis from a scientific standpoint. First, we know that if the patient has a pathologic complete response (pCR), that person as an individual has a very good prognosis. pCR is associated with a very good rate of long-term survival, and I think that is consistent across every study that we have ever done. But the real question is, if we are looking at arm A vs arm B in a neoadjuvant trial, and we see a delta between arm A and arm B, does that correlate with improved disease-free survival (DFS) in an adjuvant trial? Does it correlate with improved overall survival in an adjuvant trial? There, the data are conflicting at best, but in general they are probably negative. Patricia Cortazar at the FDA, and an international group of investigators, brought together all of these neoadjuvant trials in which arm A had been compared with arm B.[3]

In general, when one does a statistical analysis of these trials, it's very difficult to come up with a delta between two trials that correlates with a delta in DFS. As clinical scientists, that is what we need. We would love to have some formula that would say that this, in fact, is what happens. This improvement in pCR rate is associated with this improvement in the DFS rate. We can't do that.

FDA Approval: The Curious Case of Pertuzumab

Dr Miller: In the absence of that, however, we still have this draft guidance in effect. We have a drugpertuzumabthat is currently approved for the neoadjuvant setting but not for the adjuvant setting. For our readers in practice, that puts them in a difficult position when a patient comes to see them. We have heard about some difficult situations, and I would love to hear how the two of you would handle them.

The first is a young to middle-aged woman, exceedingly healthy, who had a 5-cm tumor with an obvious palpable node. She told her surgeon, "This is making me crazy. I want a mastectomy and I want it now." The surgeon honors her wishes and takes her directly to the operating room. She is now seeing you, but had you seen her in the neoadjuvant setting, neoadjuvant therapy [with pertuzumab] would have been quite reasonable. What do you do now?

Dr Hudis: I have spent a lot of time on this. I spoke at length to many of the good people at the FDA in preparation for the labeling of pertuzumab because I had real concerns, and this case highlights them. Let me expand on that for a minute. Miss Smith walks down Second Avenue in New York City. She has a 3-cm cancer, she's HER2 positive. She makes a right-hand turn into facility A, has a mastectomy, and is a candidate for conventional trastuzumab-based adjuvant therapy.

She makes a left turn into institution B, and they refer her for neoadjuvant therapy. In that second case, per the label, she gets pertuzumab [before her mastectomy]. The public health reason for accelerated approval, it seems to me, is to make a difference in people's lives. The label given to the drug stipulates four to six cycles of therapy.

Dr Miller: Because it was a neoadjuvant trial. The trial stopped at surgery.

Dr Hudis: Right, and those were the data they got; I understand that completely. However, to return to George's point, the guidance document and the actions of the FDA call for confirmatory data from the adjuvant setting. The thing about the pertuzumab label is that everybody knew there was a survival advantage in metastatic disease, and I don't think it's a stretch to call it unprecedented.[4,5] It's one of the biggest survival deltas that we have seen, and that's supporting [of neoadjuvant approval]. The APHINITY trial was finished[6]; those data will come in. So you get accelerated approval. Here is the problem: Everybody who gives four to six cycles of pertuzumab for the next 2 years waiting for APHINITY will wake up—and let's just say APHINITY is positive—and ask, did they save a single life? The answer can't be known because the APHINITY data will be for a year of therapy.

I have given you a long walk in terms of the answer. What we do (and I feel strongly and passionately about this): I give a year of pertuzumab to anybody who would have qualified for pertuzumab one way or the other. I don't care whether it's preoperative or postoperative. We actually pushed a little bit so that the National Comprehensive Cancer Network guideline would be just a little vague on this issue for that reason. I have a truly moral objection to the notion of giving only 4 or 6 months of a drug when you believe it's lifesaving, because that [4 to 6 months of therapy] might not be.

Dr Miller: At this point we don't know whether it's lifesaving at all. I'm going to take the liberty of saying that the three of us would be surprised if the APHINITY trial is not a positive trial. But it does put practitioners in a quandary. It has the potential to put patients at financial risk for different coverage.

Cliff is going to give that woman who had surgery first a year of pertuzumab. If she takes a bigger left-hand turn all the way across the country to your office in Stanford, what are you going to do with her?

Dr Sledge: I'm going to do what her insurance company allows me to do because these are very expensive drugs. But Cliff brings up a very important point. One of the issues with neoadjuvant therapy is that it doesn't teach us very much about duration. If we look at trastuzumab, for instance, we know (or at least we think we know) that 6 months of trastuzumab in the adjuvant setting is not as good as a year of trastuzumab in the adjuvant setting.[7] As new drugs and new biologics come along for the treatment of breast cancer and are used in the neoadjuvant setting, the neoadjuvant setting is probably not going to be the best place to answer questions about duration. In fact, we may lose in the long run if we choose the wrong duration.

Gambles and Predictions

Dr Hudis: You bring up a point, and I feel a certain tension about it. The problem is that I may be making a mistake. We all may be making a mistake. But we have to go to the clinic and make a real decision for a real patient. I would rather make the mistake for 2 years with a little overtreatment—it's expensive—than make the mistake of undertreatment (if AFFINITY turns out to be positive). That's just a value judgment.

Dr Miller: But with undertreatment, we have no way of knowing whether the undertreatment helped her a little bit in terms of her overall survival, but maybe not as much as a full year of treatment. That would certainly be what we saw with the shorter duration of trastuzumab.[7] Those patients [who had trastuzumab for 6 months] didn't receive no benefit, but it just wasn't quite as much [as those receiving trastuzumab for 1 year]. But we won't know. Did she receive any benefit at all, did she receive partial benefit, or did she receive the full benefit? If the reason to treat people in the neoadjuvant setting is the same as the reason to treat them in the adjuvant setting, this dichotomy creates a tension.

Dr Hudis: If the reason for accelerated approval when the data are evolving is to make a public health impact, the public health impact is not the pCR; the rate of breast conservation is no different on the two arms of those trials. These are not patients by and large with inoperable disease being salvaged or saved from mastectomy. The practical impact of the pCR on those patients is not really detectable, except for good news. The long-term implications of that good news are exactly what George described before. My issue is, in a way, more nuanced. I am not sure whether it makes any difference just to get a bump in the pCR rate.

I want to be very clear about this: My prediction, but I don't have any data, is that pertuzumab will be effective as adjuvant therapy. We could consider lapatinib in the same discussion, and if you are careful about it, you can find a delta in pCR for lapatinib pretty much like pertuzumab. We know from last year at the American Society of Clinical Oncology annual meeting that that doesn't necessarily translate into an impact in long-term outcome.[8] So the principle isn't there for me, which brings me back to why do it? Do it to save lives. If you are going to save lives, you have to take a gamble on what is being tested to save lives.

Dr Miller: We have surgeons who tune in as well. If Cliff's Miss Smith turns right into that facility and has her immediate surgery, has the surgeon done something wrong?

Dr Sledge: No, the surgeon has done what the surgeon thinks is appropriate. In general, wherever we have tested it, neoadjuvant therapy and adjuvant therapy give you identical survival outcomes.

Dr Miller: If the therapy is the same, which is our problem here.

Dr Sledge: And that is the question: Is the therapy going to be the same? At least in the short run, the therapy won't be the same. It's not that the surgeon has done something wrong; it's perhaps that the system has done something wrong.

Does pCR Equal DFS?

Dr Miller: The other area where there has been enthusiasm for pCR as a surrogate is in the triple-negative setting. At this point, no drug is approved in that setting on the basis of pCR, although there is an ongoing phase 3 trial with that as the hope, looking at a poly ADP ribose polymerase (PARP) inhibitor in triple-negative patients.[9] Do we have any greater sense of that correlation between pCR and long-term outcomes in a triple-negative vs HER2-positive patient?

Dr Sledge: If we look at the Cortazar analysis, the answer is no. In the Cortazar analysis, it didn't matter which subtype you had. They were unable to show that the delta pCR equals delta DFS for any of the particular subtypes. Currently we would have to say that the answer to that is no.

Let me say something in fairness to the FDA, because the FDA, in their draft guidance, were quite frank about saying that in approving something on the basis of neoadjuvant therapy, they were approving on the basis of the totality of the data. The totality of the data for pertuzumab, for instance, includes a metastatic overall survival advantage. It includes a large amount of safety data. It includes the fact that, as Cliff mentioned, a phase 3 adjuvant trial is already in the bank and waiting to be evaluated,[6] but we are not going to get that every single time. We were lucky to get that with pertuzumab, but we certainly are not going to get that with every drug that comes along. That is a big question.

Dr Hudis: I agree. The other point is that the subtypes do matter, which is indirectly what you are asking about. One has to also be aware of the fact that some of the most important drugs from a public health point of view don't give us pCRs in the ER-positive setting. The draft guidance steers you away as an investigator or a company from thinking about ER-positive patients for this approach.

Dr Miller: The other thing that this means is that we will have to consider as a community the very real possibility—perhaps not with pertuzumab but with some agent at some future time—that accelerated approval on the basis of improvement in pCR, in a subsequent study that is appropriately powered and followed for a long-term outcome, will not pan out. What happens at that point?

Dr Hudis: It's an interesting question because in the case of pertuzumab, it's perhaps less damaging than it would seem at first. The drug is approved; it has other indications, and so if one were to withdraw the label for that indication, no harm is done.

Dr Miller: But for some drugs, this [accelerated approval for use in the neoadjuvant setting] could be their first route [to approval], and that [a negative trial for long-term outcome] could remove an agent entirely.

Dr Hudis: It could, although George's point about the totality of the data is important. It would be surprising if a drug garnered approval only in the neoadjuvant setting, on the basis of the totality of the data at the moment, and by definition be limited to just pCR results. It seems to me unlikely. Maybe one could imagine that happening, but I don't see it as happening so easily.

Considering the Harms

Dr Miller: I am also thinking about the harm that we have done in that time, because in however long a time it is between accelerated approval and ultimate study results, and withdrawal of that approval regardless of whether it is just a change in indication or a complete removal from the market, there will be many—likely thousands—of women who have been treated, some of whom individually may have benefited, who we won't be able to identify. But some women have clearly been harmed by toxicities. No matter how well tolerated drugs are, they all have some toxicities. In some patients, these are particularly severe. What do we as a community say to those women when they come back and ask us?

Dr Sledge: We have been there before. We were there before in the 1990s with high-dose chemotherapy and stem cell transplantation, when the popularity contest of the moment condemned thousands of women to therapy that did not benefit them and arguably harmed them. It is the nature of clinical research. I always remember Ben Bradley's discussion of newspapers. He said that they were the first rough draft of history. Our clinical trials are probably the same thing. They are the first rough draft—and sometimes very rough—of medical history. We learn as we go along. Hopefully we learn carefully, hopefully we don't harm our patients, and hopefully we don't overexpose them to therapies that will harm them.

Dr Hudis: You captured the real tension that the regulators have to confront and manage. No matter what it might sound like, this is the reason that I have tremendous respect for them. This is a very delicate balancing act between the pressure to do more, better, and faster than we have before, and the inevitable backlash against the occasional unfortunate harm. Getting it right is probably not possible. It's a never-ending game of lefts and rights and adjustments. We are all in this together.

Dr Miller: For pCR, you just gave a great analogy. pCR is our first draft of efficacy in an early-stage setting. First drafts are sometimes good drafts. Sometimes they are substantially rewritten before they come to their final versions.

Dr Hudis: That's true. I think it will turn out that having pertuzumab labeled for the neoadjuvant setting was a good thing, although I would have preferred a more ambitious or aggressive stance. But that doesn't mean it's a principle that is established. It may not be. It may just be in the circumstance that all of the forces were aligned.

Dr Miller: Do you think approval in the neoadjuvant setting a good thing?

Dr Sledge: It is for pertuzumab. I go back to that phrasethe totality of the data. In the metastatic setting, we have never before seen a drug that improves something by as much as pertuzumab has in terms of overall survival. One has to think that it is going to translate to the adjuvant setting. Perhaps we will be wrong, but I doubt it. This is an exceptional difference. If it was a 2- or 3-month difference, one would worry.

Dr Miller: Thank you both for joining us for this very important, thought-provoking discussion. Thank you to our audience for joining us in this edition of Medscape Oncology Insights. This is Kathy Miller, reporting from San Antonio.


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.