Our goal was to see if we could pick out a group of women who get no benefit from chemotherapy, even though they have very high risk disease. In fact, these were premenopausal women. To answer this question we had to go back to some quite old trials that were done when it was still an open question whether chemotherapy benefitted women at all. At that time only quite high risk women were being considered for chemotherapy and this study, the Danish Breast Cancer Group trial 77B it was called, the women who were enrolled were premenopausal, they were mostly node positive with high risk disease but not metastatic disease.
How was the study formatted?
The idea was to use a definition of low risk breast cancer called luminal A that my group and others have developed over the past decade. Because the material that was available from this older trial was in the format of tissue microarrays the only way to test for the luminal A status was using immunohistochemistry. So there’s a panel of six biomarkers, ER, PR, HER2, Ki-67 and a couple of basal markers that can be used to identify the different subtypes. For here what we were really concerned about was the group called luminal A and they are defined as ER positive, HER2 negative but also low proliferation by Ki-67 and high progesterone receptor, signifying that they’re still highly hormone dependent tumours most likely.
The idea would be that they would be low risk molecularly but in this particular trial, because all these women were node positive, they were premenopausal and in many cases they had large tumours clinically they would be considered a relatively high risk group.
What was the outcome?
We identified over 600 cases that we could give a subtype and about 170 of them were luminal A. We compared their response to chemotherapy with all the ones who were non-luminal A and the short version of it is the non-luminal A patients did very well, the chemotherapy improved their outcomes a great deal, the hazard ratio was 0.5. The luminal A patients, on the other hand, had completely overlapping Kaplan-Meier curves whether they received chemotherapy or not, hazard ratio was 1.0 so there was no sign that despite them being premenopausal with large tumours and node positivity, if they were luminal A there was no sign that the chemotherapy was benefitting them.
What was the chemotherapy used?
This was an older regimen, a first generation chemotherapy so it was classic CMF chemotherapy. Some of the women also got cyclophosphamide as a single agent but that drug is certainly still used today and is considered the most active drug in many regimens.
Since these are premenopausal women, would you feel confident in saying they could not have chemotherapy?
That is a question for the research community, isn’t it? Or the clinical community because this is a bit of a departure. It’s part of a spectrum of changes we’ve seen over the past few decades of realising that we can back off somewhat in certain situations on how aggressive our therapy needs to be. Many people accept that if you have other low risk characteristics, you have a small tumour, it’s ER positive and being treated with tamoxifen in older women and therefore your prognosis is very good, that it’s definitely safe to avoid chemotherapy in those situations. Even if the chemotherapy worked it would only be able to make a small difference in a really good prognosis situation. Here we’re dealing with something that in principle is a little more risky. These are women who do have a higher risk of relapse to begin with but we’re saying that if they’re luminal A subtype that risk won’t get any better. It’s a bit more of a paradigm shift than other studies have shown. So many people may not feel comfortable with this data on its own changing therapy. It would qualify as level 2 evidence because it was a formal prospective study of an existing clinical trial, so a prospective retrospective study, and many feel for that to be practice changing you may need to repeat it or have some other corroborating evidence.