The standard therapy for patients with stage 2 and 3 HER2-positive breast cancer is neoadjuvant therapy with multiagent chemotherapy plus two anti-HER2 antibodies: trastuzumab and pertuzumab. That results in pCR rates of about 60% and three-year disease-free survival of about 92%.
The purpose of ECOG-ACRIN, or EA1181, is to use a less intensive therapy – neoadjuvant, single-agent taxane plus trastuzumab and pertuzumab – for twelve weeks, or four cycles, and then to complete a year of HP, use radiation and endocrine therapy if indicated, and to see if, for the patients who get a pCR with that less intensive regimen, if the outcomes are equivalent to those who get a pCR with the standard, more intensive regimen. That primary aim of the trial needs longer follow-up. We have completed enrolment and patients are being followed, but we do need longer follow up for that primary aim.
What we’re reporting on at this meeting is the pathologic complete response rates, the pCR rates, for the entire group, as well as for the ER-negative and the ER-positive disease cohorts. But also importantly the clinical and the molecular predictors of who got a pCR. As we start to use these less intensive, de-escalated approaches, it will be really important to know how to optimally select the patients who are the best candidates for this kind of less intensive therapy, and that’s what we were reporting on at this meeting.
What was the methodology and what were the findings?
The study enrolled over 2,000 patients, and 2,141 patients who had, again, stage 2 or 3 HER2-positive breast cancer. 2,141 received at least one dose of the taxane plus the trastuzumab and pertuzumab, and that’s our population that’s evaluable for a response. We separated the two cohorts, the ER-negative disease and the ER-positive disease – about a third of the patients had ER-negative disease and about two thirds had ER-positive disease. 95% of those patients completed their four cycles of neoadjuvant therapy and went to surgery per protocol, and we are reporting on them.
What we found was that overall, pathologic complete response rate was 44%. It was about 64% for ER-negative disease and about 34% for the ER-positive disease. Importantly, what we did was looked at clinical and molecular predictors of pCR. For the clinical pathologic factors, we found that estrogen receptor-negative disease, as many have known, has a higher pCR rate, but among the ER positive disease about a third of the patients had a lower estrogen receptor expression level – 70% or lower – and those patients had a higher pCR rate. So if the ER level was 1-10%, the pCR rate was 63%, similar in ER negative disease, and if the ER was 11-70%, pCR rate was 52%, so very good.
Second, we found that if the tumour had IHC of 3+, then that was a significant independent predictor of pCR. Finally, surprisingly, we found that the choice of taxane made a difference. So about two thirds of the patients had received weekly paclitaxel as their taxane and about a third had received once-every-three-week docetaxel. We found a higher pCR rate with the use of the paclitaxel, and that wasn’t due to amount of expected drug received, so that was a bit of an unexpected finding. We found those three clinical pathologic factors were helpful in predicting who would have pCR: the ER level, the IHC3 plus, and the use of the paclitaxel.
We then used a molecular tool called HER2-DXPCR score. This is a genomic tool that combines gene expression of the tumour for various things like immune activation, luminal differentiation, HER2 expression proliferation, and it combines it with some clinical parameters – the tumour size, whether the nodes were involved. We used that HER2-DX score, the pCR likelihood score, and found that for the patients who had a high HER2-DX score versus a low HER2-DX score, there was about a 40% difference in the pCR rate, and that was true whether the tumour was ER negative or ER positive.
Then we actually did a multivariable analysis. We took the HER2-DX score and we took our three clinical pathologic factors that we had found and we analysed them all together and found that clinical pathologic factors remain significant, the HER2-DX score was significant. So these molecular and clinical factors each independently were helping to predict who would get a pCR with this less intensive regimen, the THP.
What are the clinical implications of these findings?
I would caution that we need longer follow-up to ensure the primary aim of EA1181. Namely that a pathologic complete response rate obtained with less intensive therapy is equivalent to one achieved with standard therapy. I think that until we have that primary aim, standard neoadjuvant therapy should be used. But I think that in the meantime we should be analysing, as we’ve been doing, which are the predictors, so if that is a successful approach and we adopt that, how can we optimally select patients who are going to do best with this less intensive therapy.
We’re also going to be looking at radiology – so radiomics, MRIs – there are some other molecular tools such as circulating tumour DNA that we’ll be looking at. So we’re sort of using this time to try to really get ready, that if that approach is greenlighted, we really can help select the patients who are going to be best for this approach.
Is there anything else you would like to add?
I think it was just a wonderful effort. Although it was led by ECOG-ACRIN, the Alliance, and SWOG, and NRG, and many of the collaborative groups in the United States participated. So in order to do this kind of prospective trial of over 2,000 patients it took a lot of cooperation and collaboration between all of the cooperative groups in the United States.
