Is Bevacizumab Useful in Early Breast Cancer?

Kathy D. Miller, MD; Gunter von Minckwitz, MD


December 17, 2010

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


Kathy D. Miller, MD: Hello. I'm Dr. Kathy Miller, Associate Professor of Medicine at the Indiana University School of Medicine in Indianapolis. This Medscape Oncology Insights in Breast Cancer is coming to you from the 2010 San Antonio Breast Cancer Symposium

Joining me today is Dr. Gunter von Minckwitz, Chairman of the German Breast Group (GBG) Institute, Alternate Director of the Luisenkrankenhaus in Düsseldorf and Associate Professor of Gynecology at the University of Frankfurt, Germany.

GeparQuinto (Neoadjuvant) Trial

Today I would like to focus our discussions on neoadjuvant chemotherapy and particularly the important results from the 2 arms of the GeparQuinto study.[1]The GBG group has a long history of doing important neoadjuvant studies and this most recent effort really divided the group into the HER2-amplified patients and the HER2-normal patients. So take me through the design for the HER2-positive patients first.

Trial description: HER2+ arm.

Gunter von Minckwitz, MD: We assessed patients according to HER2 status and if they were HER2-positive by local pathology, patients were given chemotherapy with epirubicin, cyclophosphamide (EC) 4 cycles, followed by docetaxel 4 cycles either in combination with trastuzumab or in combination with lapatinib.

They were given the anti-HER2 agents all through the complete chemotherapy in combination with the anthracycline. Patients then got surgery and the pathologic complete response rate was assessed.

Dr. Miller: You had the confidence to give anthracycline and anti-HER2 therapy together, and I think that is something about which many of us might have been a little surprised.

Dr. von Minckwitz: Yes, but this is something that we have already explored in the GeparQuattro study, in which we treated 450 patients with this exact combination of trastuzumab with EC followed by docetaxel, so the new approach was to give lapatinib in combination with epirubicin. Otherwise we knew that this approach was safe and there are other studies, such as the NOAH [NeOAdjuvant Herceptin] study or the Houston study. There is also a metastatic trial where trastuzumab was given in combination with epirubicin. If you select patients carefully, meaning they need to have a left ventricular ejection fraction (LVEF) of at least 55% (so not a just normal LVEF but an LVEF just above normal) and if you exclude patients with cardiac comorbidities, then this strategy appears to be quite safe. We did not see any of the cardiac toxicity reported in the initial pivotal metastatic trials.

Dr. Miller: In the neoadjuvant setting, you can look at clinical and pathologic response as a way of getting early indications of efficacy. This is one of the first studies to directly compare trastuzumab and lapatinib. Did you see differences?

Dr. von Minckwitz: Yes. From the first HER2-positive patients we are getting quite convincing data that pathologic complete response (PCR) is really an important and reliable surrogate marker. We were able to show this in the TECHNO [Taxol Epirubicin Cyclophosphamide Herceptin Neoadjuvant] Study. This was a phase 2 study in which patients with PCR had significantly better outcome than those without PCR, and this was also shown in the NOAH study. In this population, PCR rate is really relevant and therefore we have chosen this as the primary endpoint of our trial.

We have to be very precise. We have chosen the most conservative definition of PCR, which means that only those patients with no noninvasive residual tumor in the breast and in the nodes were considered as having a PCR.

The aim was to show that lapatinib was better than trastuzumab, but it ended up that trastuzumab was better than lapatinib, the reverse of what we expected, so this result was significant. In this part of the GeparQuinto we included in the study 620 patients, so it was well-powered, and lapatinib did not reach the results of trastuzumab.


Dr. Miller: There were some toxicity differences, as well.

Dr. von Minckwitz: Yes.

Dr. Miller: I remember seeing more patients in the lapatinib arm who discontinued therapy because of toxicity.

Dr. von Minckwitz: Yes. We had a run-in phase during which we analyzed the the early data, and in combination with the information from the Neo-ALTTO [Neoadjuvant Lapatinib and/or Trastuzumab Treatment Optimisation combining lapatinib and trastuzumab] study, we decided to lower the dose of lapatinib. During the run-in phase, one third of the patients discontinued lapatinib early so we reduced the dose from 1250 mg to 1000 mg and we implemented a standby medication (loperamide) to improve compliance. In the end, 23% of the patients discontinued lapatinib early compared with trastuzumab, which was not complete in 30%. There is about a 10% difference with respect to discontinuations between the 2 arms, which might have had at least some effect on why lapatinib was not as effective but is probably not sufficient to explain the whole difference.

Dr. Miller: That was my question looking at the data. Although it would be improper to look just at the results in the patients who completed therapy, it is a bit tempting to see to what extent toxicity and early discontinuations influenced the results. There was about a 10% difference in PCR, and only about a 10% difference in discontinuation, so you would have to assume that all of those patients would have had a PCR for that to be the entire explanation.

Dr. von Minckwitz: Exactly. It is part of it, but it's not the whole explanation. The other issue was that although we were looking at diarrhea, because this was reported, it's not the only toxicity that led to discontinuation. We saw a full range of serious adverse effects (SAEs), but they are difficult to measure. The reduction of the dose was the most important, but that of course, affected every patient.

Dr. Miller: And certainly may have affected the efficacy, as well.

Trial description: HER2– arm.

Dr. Miller: Let's shift to the other side of the GeparQuinto study. The HER2-normal patients were randomized to bevacizumab (or not) and although there are several other studies looking at bevacizumab in the neoadjuvant setting, the GeparQuinto study is the first to have efficacy results available.

Dr. von Minckwitz: We recruited the same number of HER2-negative patients, and for the same time duration, and we ended up with almost 2000 patients in that arm. They received the same chemotherapy backbone, EC and docetaxel, and were randomized to either no bevacizumab or bevacizumab.

We also did a response assessment after the 4 cycles of EC and patients who did not respond clinically were taken off the study and re-randomized to weekly paclitaxel with or without the mTOR inhibitor everolimus. This was based on our experience that if you don't get a response after 4 cycles, your chance for a PCR is only 3% or 5%, so we thought that these patients needed an alternative approach and we continued only with responding patients.

Results: bevacizumab.

For the statistical analysis, these nonresponding patients were counted as having no PCR, to be certain that if the arms were not balanced, this did not bias our results. Actually, this is something that we have seen. There were 8% more nonresponding patients in the chemotherapy-alone arm compared with the bevacizumab arm, so respectively, 24% vs 16% nonresponders. This is still lower than what we expected from our previous trials, and it might be because of efficacy or because this was an unplanned assessment and investigators preferred to stay with the bevacizumab in some patients, and therefore said the patient had a marginal response.

Dr. Miller: Yes, there certainly is a gray area in assessing patients in the neoadjuvant setting.

Dr. von Minckwitz: So that is probably the explanation for this difference because at the end, the PCR rates were not significantly different. We had a 15% PCR rate in the chemotherapy arm and 17.5% PCR rate in the bevacizumab arm, according to this very conservative definition. However, if you look at the Houston definition, allowing residual ductal carcinoma in-situ or the National Surgical Adjuvant Breast and Bowel definition, not considering nodes, this 2%-3% difference is maintained for all of these analyses.

Dr. Miller: So the definitions changed the absolute number but not the relative difference or whether they were really significant?

Dr. von Minckwitz: Yes.


Dr. Miller: Now, there have been big concerns about toxicities that patients might have at the time of surgery. Did that turn out to be a real problem?

Dr. von Minckwitz: So far, we have only analyzed SAEs, because the study just ended last Friday. So far, in combination with epirubicin and cyclophosphamide, bevacizumab did not really increase the number of SAEs. In combination with docetaxel there was a doubling of SAEs, from 125 to 220 or so, so there was more toxicity. We mainly saw more neutropenias and mucositis, especially in the docetaxel patients. With respect to compliance, both arms were more or less the same. Four percent of patients discontinued bevacizumab and continued chemotherapy. In the control arm, 12% vs 14% respectively, discontinued chemotherapy and bevacizumab or discontinued only chemotherapy. So there was more toxicity. We didn't see a lot of hypertension, but we have learned a lot in the last couple of years.

Dr. Miller: And surgical complications?

Dr. von Minckwitz: We only had 5 SAEs for wound-healing complications in the bevacizumab arm and none in the control arm. In the protocol we said patients couldn't be operated on until 4-5 weeks after the last bevacizumab infusion. This interval was probably long enough.

Dr. Miller: Fewer problems with surgery than we might have anticipated.

Dr. von Minckwitz: We will analyze this in detail. We have collected a lot of data for that because it is of interest, but this is something that we have to report later.

Dr. Miller: Gunter, thank you for sharing these results with us and we will be looking forward to all of the correlative work you will be doing with the tumor samples.

Dr. von Minckwitz: Of course we did subset analyses. They were preplanned and patients were stratified according to hormone receptor status. In the 600-650 triple-negative patients, bevacizumab significantly increased the PCR rate, so there was a 40% increase. This is interesting and very important for the adjuvant study because it might have been a good strategy to include only triple-negative disease in this trial.

It seems that this 2.5% absolute difference is restricted to the triple negatives because the PCR rates in the hormone receptor positives were identical. There is a big opportunity to do a lot of biomarker research. We have collected several thousand specimens already, so that we can really do a lot of work there. That is most important in the neoadjuvant setting, that we have access to the tumor before and after treatment and can correlate the efficacy to these markers.

Closing remarks.

Dr. Miller: We look forward to those results in future years. Thank you for coming in today and thanks to our audience for joining us for this edition of Medscape Oncology Insights. This is Kathy Miller at the 2010 San Antonio meeting in 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.