Lisa A. Carey, MD, comments on the recent I-SPY 2 and KRISTINE findings, the I-SPY 2 program as a whole, and the work ahead in the neoadjuvant landscape of HER2-positive breast cancer.
Lisa A. Carey, MD
While recent findings from the I-SPY 2 trial have shown potential with the combination of ado-trastuzumab emtansine (T-DM1; Kadcyla) and pertuzumab (Perjeta) for patients with HER2-positive breast cancer, the neoadjuvant space still has a lot of work ahead, according to Lisa A. Carey, MD.
Results presented at the 2016 AACR Annual Meeting1 showed that, out of the 249 patients enrolled on the I-SPY 2 study, 54% of those who received T-DM1/pertuzumab experienced a pathological complete response (pCR) rate compared with 22% of those who received the combination of paclitaxel (Abraxane) plus trastuzumab (Herceptin).
This suggests that T-DM1 could increase overall survival (OS) in this patient population, but Carey adds more research with the regimen needs to be conducted.
“The I-SPY 2 paradigm is a really interesting one,” Cary explains. “They are really pushing the envelope on trying to help with developing drugs and regimens in the neoadjuvant setting. I do worry how often the assumptions are going to hold true. In this particular case, we will see.”
Moreover, findings of the KRISTINE trial,2 which were presented at the 2016 ASCO Annual Meeting, demonstrated that neoadjuvant treatment with docetaxel, carboplatin, trastuzumab, and pertuzumab (TCHP) was associated with a higher pCR rate compared with T-DM1 plus pertuzumab in HER2-positive patients, at 55.7% and 44.4%, respectively.
OncLive: What are your thoughts on the I-SPY 2 trial of T-DM1/pertuzumab versus paclitaxel/trastuzumab in the neoadjuvant setting?
In an interview with OncLive, who is the medical director of the University of North Carolina Breast Center and chief of Hematology/Oncology and physician-in-chief of the North Carolina Cancer Hospital, comments on the recent I-SPY 2 and KRISTINE findings, the I-SPY 2 program as a whole, and the work ahead in the neoadjuvant landscape of HER2-positive breast cancer.Carey: In this particular trial, the control arm received trastuzumab and paclitaxel for 12 weeks, then they got AC (doxorubicin and cyclophosphamide), and then they went to surgery. In the investigational arm, patients received T-DM1 and pertuzumab, so they switched out 2 drugs. There is a hidden chemotherapy kind of tucked inside T-DM1.
The estimates of the performance of the T-DM1 arm outperform the estimates of the control arm with a certain level of statistical certainty—99% based on what they had seen. Based on the early signals of this, it was much more likely that the T-DM1/pertuzumab arm would outperform. That’s what the authors reported.
In context, their control arm didn’t perform very well. The T-DM1 arm was estimated to be 52% and the control arm was at 22%. Those 2 differences may be true, but I’m not entirely sure why the control arm underperformed. I’m a little bit concerned that T-DM1 seems to underperform relative to what we would expect these days with a trastuzumab-containing regimen.
That said, this clearly is a positive trial. What they have is a difference that is based on their statistical assumptions—that T-DM1/pertuzumab would outperform paclitaxel/trastuzumab. I’m a little worried about the underperformance of the control arm, which only has 30 patients in it.The KRISTINE study is a randomized neoadjuvant clinical trial comparing TCHP with T-DM1/pertuzumab in HER2-positive breast cancer—very similar to the I-SPY 2 trial. They received it for 6 cycles. In that setting, TCHP outperformed T-DM1/pertuzumab in pCR significantly more than 10%.
On the other hand, while the pCR rate was in favor of the TCHP arm, toxicity was in favor of the T-DM1/pertuzumab arm, with serious adverse events at approximately 5% [versus] almost 30% in the TCHP arm.
Can you explain the purpose of the I-SPY 2 program?
There is going to have to be some thoughtful consideration of how T-DM1 plus pertuzumab moves forward, given a pretty good-sized study that used TCHP where it was a significantly lower pCR rate.Basically, it is a framework for trying to identify promising drugs or regimens but it is a very statistically complicated framework. Essentially, the trials are mostly made of up clinical subsets by estrogen receptor, progesterone receptor, and HER2 receptor [status]. Low MammaPrint-score luminal tumors are excluded, so some of the low biologically risk tumors are not selected.
The trials have sort of a backbone; they have standard regimens to which they compare a novel regimen, and they do it in an iterative fashion. It’s really looking at a hint of activity with an estimated pCR rate in the neoadjuvant setting. It’s not a real pCR; its not powered for a pCR endpoint. When one is considered a positive I-SPY 2 trial, it graduates to a larger definitive trial. It’s a way of being parsimonious about resources, because if it looks like there is only a 10% chance that the regimen is going to be positive in a larger trial, you’re not going to design a big large trial.
If you have 6 options, you want to go through a bunch of them, and you want to pick the ones you want to take into larger trials, they are going to cost a lot of money—so this is a good way to do it. It’s a statistically intensive way to do it. It doesn’t make you go through a lot of hoops in order to identify something that looks promising.
What it doesn’t do is give you a definitive accurate answer; it gives you this regimen that has an 80% likelihood of being positive in a larger neoadjuvant trial. It also doesn’t tell you that the patients are going to live longer. There is some controversy about the neoadjuvant approach and the way it can be interpreted as conferring. If a regimen is better in the neoadjuvant setting, what does that mean in terms of the likelihood of it being better in the adjuvant setting?
pCR, which is a traditional endpoint for neoadjuvant trials and is the endpoint for the I-SPY 2 trials, is not clinically meaningful. It’s a proxy and is an intermediate biomarker for relapse and survival. The question is, on a trial level basis, “If you see an augmentation in pCR, will that translate into an augmentation of relapse-free survival and OS?” In some cases it does, and in some cases it does not.
Are there other clinical trials in the neoadjuvant space that you are excited about?
The I-SPY 2 trial is a couple of steps away, but it’s a really good way of narrowing down your options—picking a winner, in one sense.Right now, there aren’t a huge number of them that are outstanding.
The role of T-DM1, where it is, and where it belongs, is still a work in progress. In the metastatic setting, it did not prove to be superior to a pertuzumab-based regimen. I think we are the most interested in seeing the results of the APHINITY trial, which is a HER2-positive study in the adjuvant setting.
There was a neoadjuvant trial called NeoSphere, which looked at pertuzumab plus chemotherapy and lapatinib, and it showed a statistically significant improvement in pCR. It was a high enough difference—about 20% difference in pCR rate—that there is excitement that this may improve relapse and survival in the adjuvant setting, which is what APHINITY was testing.
What are your thoughts on neoadjuvant treatment?
Pertuzumab does improve survival in the metastatic setting, in the first-line setting, so it is a drug that holds real promise in this realm. That will come out in 1 to 2 years, and that’s probably the most high-profile thing.For many of us, the question that remains is, “Who really needs it? Who really needs all of these drugs?” There is clearly a group of patients who don’t need 3 chemotherapies and 2 HER2-targeted drugs, and there are a bunch of ongoing studies. TBCRC-006 and TBCRC-023 were 2 trials that explored only HER2-targeted agents—not chemotherapy—in the neoadjuvant setting. In both cases, they found some patients to have a pCR, suggesting that there are some tumors that really respond well to this.
In the NeoSphere study, there was an arm where patients received only trastuzumab and pertuzumab, and there was about a 17% pCR. This is interesting because, you wonder, “Who are these patients who arguably don’t need any chemotherapy?” It is certainly a suggestion that there are some tumors for which you may be able to omit chemotherapy altogether. It would be lovely to be able to figure out who those patients are.
In a follow-up to the ATEMPT trial, T-DM1 may allow us to omit the chemotherapy because it essentially delivers chemotherapy, but in a “Trojan horse” kind of way. I think they have approved 300 out of 500 patients, so that one we’ll wait and see. We got a few years, but it is very promising. Personally, I have a patient on that study who ran the Boston Marathon on it.
In the neoadjuvant space, what do you think is on the horizon in the next 5 to 10 years?
HER2-targeted therapy goes on for 1 year, so if patients want to avoid 5 drugs or losing their hair or not feel well for 1 year, that would be a nice thing. Though T-DM1 is infusion-based, you can narrow down some of the toxicity.The neoadjuvant setting is just a means to an end; I think it’s largely to do clinical trials. If tumors are large, the neoadjuvant approach helps you with surgical procedures. However, outside of improving surgical endpoints, there is no benefit of giving the drugs in the neoadjuvant versus adjuvant setting. They are essentially the same.
The future needs to involve sorting out some of these questions about what actually improves relapse and survival, since pCR is not clinically relevant. The key thing for us is, in the HER2-positive space, these are ridiculously expensive drugs. We can’t give them all to every HER2-positive patient.
My hope and expectation is, over the next 5 years, that we get real ideas about the tumor genetic and microenvironmental influences on response for the adjuvant setting. Unless we can prospectively account for those, it is going to be very hard to optimize the care and treatment. HER2-positive disease is incredibly heterogeneous from the standpoint of immune cell activation. There are studies [demonstrating that] tumor-infiltrating lymphocytes show response to therapy.
I personally believe that the future lies in developing effective ways of identifying tumor biology that is, more or less, indolent for which we need to be less aggressive or more aggressive. The neoadjuvant setting is the right place to develop those techniques and then test them in the larger setting. It’s going to require collaboration.