This is about adjuvant denosumab. Here are my disclosures, I have a number of successful industry collaborations, none of them is in relation to this randomised clinical trial.
The background of ABCSG-18 is that we now finally have evidence that bone-targeted treatment actually is not only successful in treating bony metastases and preventing treatment induced bone loss but also reduces recurrence and improves survival, at least in postmenopausal breast cancer patients. Actually it was two years ago when Dr Rob Coleman from Sheffield, on behalf of a global collaboration, presented this data exactly at this symposium and that was published in peer reviewed form in The Lancet early this year.
So the next generation of bone targeted agents is the anti-RANK ligand antibody denosumab and this is what we tested in ABCSG-18. Earlier this year at the occasion of the ASCO meeting I was able to show that that adjuvant treatment of low dose denosumab, which is 60mg twice a year, in fact cuts clinical fractures in half so that’s quite a dramatic difference. There is not really a lot… actually there is no measurable toxicity in this double blind trial. Now, the question remaining is preventing fractures, well that’s fine, but are we also impacting on outcomes on disease free survival and eventually overall survival? There are two trials testing adjuvant denosumab with these questions, this is D-CARE, which is ongoing, and ABCSG-18 which I am presenting today.
So, very briefly, this is the trial design: 3,400 postmenopausal patients all on AI. This is a placebo controlled double blind multi-centre trial, 58 centres in Austria and Sweden, and the primary endpoint set was time to first clinical fracture, presented at ASCO, and secondary endpoints include other bone factors but also disease free survival.
Just to remind you, this is the primary endpoint result: hazard ratio 0.5, basically cutting fractures in half and demonstrating overall, in this trial focussing on bone health, that the overall frequency of fractures in these patients who are at low risk for disease recurrence but at high risk for fracture is higher than we previously assumed and we have previously reported in other trials.
So following these primary endpoint results the IDMC recommended to us that we should offer patients the option of unblinding. So what they said was you cannot keep going with the placebo group and having around patients into clinical fractures which are relevant. So this is not some back pain, this is a fracture of your wrist, your hip, any clinical fractures. So what we are going to do, we are going to offer during the year 2016 all trial patients the option of being unblinded. If they, after advice from their physician, want to be unblinded and it turns out that they were on placebo they will be offered the antibody for three years.
However, since this would somewhat compromise the integrity of the blinded outcome data the recommendation was do a DFS analysis now based on a time-driven analysis and this is exactly what we did. So disease free survival in this trial is defined as time to any evidence of local or distant metastases, contralateral breast cancers, secondary cancer or death from any cause and this is a quite common definition. Briefly demographics: 3,425 patients, quite a common postmenopausal breast cancer population we see in the Western world; median age 64 years; about three out of four tumours smaller than 2cm; 71% node negative disease and one out of four patients had adjuvant chemotherapy or neoadjuvant chemotherapy and all patients had adjuvant AI. Again we did not find a measurable difference between those two subcue injections with the antibody and the placebo injections in this double blind trial.
So this is the intention to treat disease free survival result. As you can see, there were 203 DFS events in the placebo group and 167 in the denosumab group. This gives a hazard ratio of 0.816 or a 19% relative improvement which, because we are somewhat low on events, is of borderline statistical significance. In absolute figures the benefit is about 1% after three years, 2% after five years and 3% at seven years of follow-up. We did quite a number according to an intensive statistical analysis plan; sensitivity analysis because there were a couple of patients, so called fast losers for bone marrow density, who started active treatment, 52 patients actually in there. If you censor them then you see that the numbers change a little bit but in general the message is quite clear. This is exploratory, just obtaining signals for which subgroups may benefit more or not at all from adjuvant denosumab. There is some indication that early start of the treatment together with AI would be good. Large patients with larger tumours would benefit a little bit more - ductal histology and high luminality, so to speak, in modern terms. Just to give you an example, this is 950 patients with tumours of 2cm or more and when you look at the absolute differences this goes to about 4% at three years, 7% at five years and more than 10% at seven years, which indicates that that’s quite a relevant difference.
Just to put it in perspective or caution, indirect comparison with the bisphosphonate meta-analysis of the Oxford group, you can see that it’s 14% relative difference over there, 3% benefit on the recurrence after ten years. So I believe it’s fair to say that adjuvant denosumab does at least as much as adjuvant bisphosphonates do.
In summary, this time driven DFS analysis indicates that adjuvant denosumab actually improves DFS. In the ITT it’s of borderline statistical significance but sensitivity analysis shows that that may be a conservative estimate. To stress again, this is a very safe treatment and my clinical conclusion is that adjuvant denosumab reduces the risk of disease recurrence or death. This benefit is similar to what bisphosphonates can do. It comes in addition to highly significantly reducing clinical and vertebral fractures and I believe that we should actually offer this treatment to postmenopausal breast cancer patients on adjuvant AI. Thank you very much.