I want to give a quick introduction to the role of oestrogen in breast cancer. Oestrogen has long been known to be the key hormone driving normal mammary development so if you look here on the slide we’re just looking at a mouse mammary gland. Then if you take away the oestrogen receptor in the bottom you actually really constrict down the mammary development. So it’s no surprise then that given that essential role just of normal mammary gland development that oestrogen receptor and oestrogen is a key for 70-80% of breast cancers.

In terms of recent work, we and others found mutations in the oestrogen receptor in patients who had advanced cancer that had spread outside of the breast, so-called metastatic breast cancer, and in particular patients who had prior exposure to a class of drugs called aromatase inhibitors, drugs that lower circulating levels of oestrogen. So what do these mutations do? These mutations make the receptor no longer need oestrogen. The breast cancers seem to become oestrogen independent and that’s shown here in this preclinical work where cells can grow despite the absence of oestrogen being supplied to them.

So that led us to two major questions as we were thinking about this for the clinic. How often do we see these constitutively active oestrogen receptor mutations in patients and then what are the outcomes, how do patients fare who have these mutations present? Do they do better than patients who have no mutation, do they do worse that patients who have these mutations present?

To answer that question we went to an already reported clinical trial, the BOLERO-2 clinical trial, because it enrolled a large population of women who had the key features. They had oestrogen receptor expressing breast cancers that had spread outside of the breast, they were metastatic breast cancers. All of those patients had previously been treated with the drug, the aromatase inhibitor. We were able to analyse samples from 541 of the patients on the trials, those who consented to this type of analysis. What we did was we actually looked at the plasma, not the tumour itself but the plasma, that was taken at the time that they entered into the study. We looked for the two most common mutations in the oestrogen receptor that we had previously found, 538 and 537.

One of the key findings of the study was shown here: 29% of the patients had one of these two mutations; 15% had the 538 mutation, 8% had the 537 mutation and 5.5% actually had both mutations. So this is a common alteration in patients with metastatic breast cancer.

The second thing that we did, well what we typically do in genomically driven cancer medicine is we take the tumour and look for mutations in the tumour. To get the tumour we go back to maybe their initial breast cancer in the breast or we go to a biopsy from two or three years ago and we ask are the mutations present there. That’s kind of a standard that’s done for a lot of clinical trials and was, in fact, collected for this trial. So we used that part of the analysis to compare our test, which was a plasma based test, with the biopsy of the tumour based test. What we found was the mutations in the biopsy of the tumour were rare, only 1% of the tumour had it but 30% of the plasma at the time they entered the study had it. We can talk later about the reasons why that sort of discrepancy might be present but it did tell us that plasma was finding a lot more mutations for some reason.

Moving on to how did the patients fare. So we know the mutations are common, how did the patients who had these mutations fare compared to those who didn’t? Shown here are the survival curves looking at how long just women who had these alterations or not lived. The median overall survival was significantly lower for those who had either mutation present, and that’s just comparing the yellow versus black on these curves. If you look at each individual mutation, the 538, the 537 or both mutations, each of those groups had a shorter median survival than those who did not have any mutation present. All these were statistically significant. So immediately that told us that the patients who had these mutations had a worse prognosis than those who did not.

Finally, and it’s very hypothesis generating, an early look at what else we might learn by looking at these mutations. We looked at what happened on the trial itself. The trial itself was looking at giving those patients who had metastatic breast cancer previously treated with an aromatase inhibitor, we asked what happened if you gave them a different type of aromatase inhibitor or that aromatase inhibitor plus the mTOR inhibitor everolimus. The trial showed that the addition of everolimus significantly increased how long the patients were on therapy, median progression free survival, and led to the approval of everolimus by the FDA. So we looked at how did patients who had mutation do on these two specific regimens. Just shown here is one example of that  where we looked at the addition of the everolimus and asked how did the wildtype group do and how did the mutant group do. What you see is most patients benefitted by adding the everolimus, that’s the black to red on the right side. You see a large increase in the median progression free survival meaning how long they were on the therapy, a 4½ month increase by adding everolimus. That was also true for the group who had the D538 mutation but unexpectedly the group that had the Y537 mutation did not show this benefit. This was a small group of patients in the Y537 group, it’s not something that we would now say don’t go out and give today based on this result. But it does tell us the biology and what we need to do going forward in the clinic is look at these mutations, not just as a whole but individually, and ask how do they impact therapies that we’re developing going forward. So a hypothesis-generating result as we look at how these mutations might be impacting therapy in the future.

So with that I just want to reiterate a couple of the conclusions. First, we can detect mutations in the oestrogen receptor pretty readily using plasma analysis. These were actually plasma samples that were collected not with the intent of looking specifically for oestrogen receptor mutations but they were collected as standard plasma was and we were able to use them for this. Second, mutation was quite frequent in this population, 30% at least had mutations in the oestrogen receptor. Third, by using plasma we actually think we’re doing better at detecting mutation than if we go to old archival tumour materials. Fourth, patients who had mutation had a shorter median survival than those who did not. And fifth, the patients with different mutations might end up having different responses to therapies in the future and I think that’s an important area for future research.