Thank you Suzanne. It’s a great pleasure to present these data today on behalf of my co-authors. Atezolizumab in metastatic triple negative breast cancer; we provide long term data from this expanded phase I study and some of the biomarker data as well. For those of you who obviously don’t treat triple negative breast cancer it’s a more aggressive subtype of breast cancer characterised by the absence of hormone receptors and HER2 receptors which means we have no targeted therapy at the moment for these triple negative breast cancer cases. The treatment we have is chemotherapy; most patients develop resistance relatively quickly and unfortunately the median survival is still relatively short with 9-12 months. The data we are presenting today have to be seen in this context.
On the other hand, triple negative breast cancer is probably, of all breast cancer subtypes, the best subtype in terms of selecting patients for immune therapy and this is based on a high degree of genetic instability, a high rate of mutations, higher levels of PD-L1 expression and tumour infiltrating lymphocytes in the tumour. Atezolizumab is a humanised engineered monoclonal antibody that selectively targets PD-L1. We have seen established activity in several diseases and the treatment is currently licensed for bladder cancer and non-small cell lung cancer.
What we present today are data with atezolizumab in metastatic triple negative breast cancer. These are the data of the expansion phase of the initial phase I study and are the first data to provide long-term outcome for these patients. The study population was expansion of the phase I. Patients were selected by measurable disease, all received single agent treatment with atezolizumab for initially 16 cycles; later on the protocol was amended to allow patients to be treated until progression. Patients were initially selected by positive PD-L1 expression but, again later on, we also included patients with low or absent PD-L1 expression just to get an impression, a better feel for whether PD-L1 may be a useful marker in selecting these patients for immune checkpoint inhibitor therapy.
For the analysis today I want to focus on response data, duration of response data and survival data. If you look to the right side you see the patient characteristics: most patients were pre-treated, some of them heavily pre-treated, 80% second or third line therapy. Again, to put it into context, chemotherapy response rates are clearly below probably 20% and even lower when we go to third line therapy for this group of patients. Two-thirds of patients had PD-L1 expression of 5% or higher in the immune cells, a third of patients had low expression or absent PD-L1 expression.
If you look at the objective response rates according to conventional RECIST, first we had a response of 10% in all patients, a higher response in patients with higher expression of PD-L1. Really important, there’s a higher response rate for patients who received immune therapy early on with 26% in first line setting. If you look at immune-related RECIST criteria which includes patients who had pseudoprogression there were four patients in this study where initially, according to former RECIST criteria, we observed progression or stable disease. If you look at these data the response rate goes up to 13%, again still 26% response rate in first line setting, 11% response rate in second and third line setting.
We also had a percentage of between 13% and up to 30% of patients who had stable disease but stable disease was often relatively short lived. The disease control rate was 27% in patients with 5% or higher PD-L1 expression in immune cells and 16% in patients with low or absent PD-L1 expression.
If you look at the spider plots, and I appreciate they are slightly confusing, but if you look at the lines in red these are lines of patients who meet the criteria of disease progression according to conventional criteria. As you can see, there was a relatively large proportion of patients who progressed relatively quickly. On the other hand, if you look to the right side you see there’s a substantial number of patients with long disease control and if you look at the median duration of response it’s 21.1 months. If you take this into context what I showed you initially of the median overall survival normally observed in this group of 9-12 months then a median duration of response of 21.1 months is, in my opinion, significant.
If you look at the third point, overall survival, the median overall survival in this study was 9.3 months at a follow-up for 15 months. The [?? 4:24] overall survival rates were 41% in one year, 22% at two years and three years which again suggests that there is a form of tail as we have seen in other diseases.
What we analysed then was survival, depending on whether patients had a benefit of this treatment or not. As you can see in green, those patients who had a complete or partial response on the left according to RECIST, on the right to IR RECIST. All patients were still alive at the time of the analysis with 100% one year and 100% two year overall survival rates. If you look on the left side patients with stable disease one year overall survival was 69%, patients with progressive disease one year overall survival was 33%. Of interest, if you look on the left side you see two year overall survival 11% and three year overall survival 11%. These were patients who had pseudoprogression. Actually if you look at the right side of the slide, the immune-related RECIST criteria, these patients if they are taken out because they are classified now as responders then the one year overall survival goes unfortunately further down.
So the conclusion of our data today with an extended follow up we can show that atezolizumab has substantial activity in these patients. I haven’t shared with you the safety data but the safety data we saw in these studies are very much in keeping with what we know from atezolizumab in larger progress in other diseases and from other immune checkpoint inhibitors. The key messages for [?? 5:46] are higher response rate in patients in earlier lines with a response of 26% in the first line setting whereas the response goes slightly down in later lines. We also see a slightly higher response in patients with a higher expression of PD-L1 although patients with low or absent PD-L1 expression had also benefitted in this study.
The second key message from me is the duration of response with a median of 21 months, that is significant in this disease setting. The third key message is from me the overall survival was substantially longer than what we see with other therapies at the moment.
What I didn’t share in the interests of time was our exploratory biomarker analysis but just to quickly summarise we saw that a high response was associated with higher TILS, with higher CD8 cells and, to a lesser degree, with higher PD-L1 on immune cells. But all these markers weren’t black and white in the way that we can say we can select all the responders based on this or can deselect a certain group. What we’re doing with atezolizumab going forward is a randomised phase III study which is nearing completion and that’s looking at the combination of chemotherapy and atezolizumab in the first line setting in triple negative breast cancer. That will hopefully enable us to better define the role of immune checkpoint inhibitors in triple negative breast cancer. Thank you very much.
Thank you. So we’ve just heard about a new treatment option for triple negative breast cancer which accounts for 10-20% of all cases of breast cancer. Triple negative, as Dr Schmid pointed out, has evidence of an ongoing immune response and therefore it was a likely target to bring forward for trials of immune checkpoint blockers. This study that we just heard today is the largest of its kind, 115 patients, although there have been reports with other checkpoint blocking drugs over the past couple of years in triple negative breast cancer also showing response rates but today we heard the first overall survival report in this patient population.
I’d like to take the first question and just ask about patients who have the more common types of breast cancer. You mentioned combination therapy at the end and I wonder if there are combination therapies that are being tested or that you would like to test for the more common types?
If you look at the three major subtypes of breast cancer triple negative breast cancer obviously accounts for about 10-20%. I would still call this a common subtype but here we are moving into combination therapy as well. If you look at ER positive breast cancer, hormone receptor positive breast cancer, which accounts to about 60-70% of breast cancers, the tumours are less immunogenic and largely uninflamed tumours and single agent immune checkpoint inhibitor therapy has been relatively disappointing. There is a signal, there’s modest activity but the way forward is clearly combinations. There are combinations with chemotherapy being explored as well as combinations with other drugs that modify the tumour biology, whether it’s other immune drugs or small molecules, for example, that increase the penetration of tumour infiltrating lymphocytes.
In HER2 positive breast cancer which is a subtype of about 20% of breast cancer there is preclinically and in early clinical signs a good interaction between the HER2 targeted therapy, which again is usually monoclonal antibody based, and therefore there’s a clear rationale of combining immune checkpoint inhibitors with these agents and the studies are currently ongoing.
Thank you.
