EADO Congress 2015

Progress in targeted therapy of melanoma

Prof Keith Flaherty - Massachusetts General Hospital, Boston, USA

You’ve been looking at a phase I study; you’ve been looking at combining molecular drugs together with immune checkpoint inhibition. Can you tell me what it is that you’ve been doing in this new study that you’ve got here with PD-L1 blockade?

In previous years we’ve established the benefit of BRAF inhibitor monotherapy, BRAF MEK combination therapy, for the patients who harbour an activating BRAF mutation with metastatic melanoma. And we’ve been performing studies in patient tumour specimens over the years documenting immunologic effects of those treatments, pointing towards but not fully proving, but pointing towards the possibility that adding an immune agent, a drug designed to liberate immune function, if you will, which the PD-L1 blocking antibodies are amongst the class of such drugs, that adding those drugs might leverage some of the effects that are otherwise achieved even just with the so-called molecularly targeted drugs. So in this trial the first goal in melanoma patients only is to establish safety of the combination and we’re actually looking at two different combinations. For patients with a BRAF mutation we administer the BRAF inhibitor, the MEK inhibitor and a PD-L1 antibody; in patients who lack a BRAF mutation, which includes those who harbour NRAS mutations, NF1 mutations and others as yet somewhat undefined in terms of their quote-unquote oncogenic drivers, they also get a MEK inhibitor with PD-L1 but no BRAF.

Now you could use molecular drugs, targeted agents like dabrafenib, to add to regular immune checkpoint inhibitors, PD1 targeting drugs or CTLA-4 drugs. So why did you go for PD-L1 targeted agents here?

If you look at the available evidence, and in clinical practice we see the same, that CTLA-4 blocking antibodies promote the greatest wave of autoimmune toxicity, so have the highest rate of overall and severe autoimmune toxicity. PD1 antibodies less so and yet they have a higher response rate in melanoma which is an interesting dual observation. PD-L1 antibodies were theorised and clinically appeared to have comparable efficacy to PD1 blocking antibodies but less autoimmune toxicity risk. The available evidence has been generated not in head to head trials of PD1 versus PD-L1, but with hundreds of patients treated in comparable trials in cross-trial comparisons one can see that lower rate of autoimmune toxicity and overall toxicity. So if one were to simply enter this realm of combination molecularly targeted immune therapy with a perspective of wanting to have a priori the agents with the lowest overall risk profile of any type of toxicity these PD-L1 antibodies are currently the most attractive.

So what’s happening, then, in your phase I study?

So far we’ve observed, as predicted, that we’re able to give the drugs in combination at their so-called full doses, in other words at their individual doses. Dabrafenib trametinib as a doublet are able to be given at their full doses and are in the US approved for regulatory purposes and clinical practice. The MEK inhibitor has its own stand-alone dose and each of those full doses are able to be combined with the quote-unquote standard or full dose of the PD-L1 antibody that’s been evaluated in hundreds and hundreds of patients as monotherapy so has a fairly firm understanding in place in terms of what the so-called recommended phase II dose is. So we’ve already established that we can give full doses safely and having treated enough patients, more than twenty in each of the combination arms, at full dose we’ve seen in fact that the toxicity profile doesn’t seem to be in any way suggesting synergistic toxicity. At worst we have just each agent contributing their own, or so-called additive, toxicity which already points to the ability to then proceed to more dedicated efficacy studies as well as on-going molecular mechanistic studies to understand are these drugs really doing something together that they can’t do alone in terms of immunologic effects.

So what sort of efficacy signal, if any, is coming out of this study so far?

Really the primary efficacy hypothesis is that we will extend duration of response, either comparing to molecularly targeted therapy alone or the PD-L1 antibody alone. That’s where we’re in a luxurious position in the field where the average duration of efficacy, be it objective response or even stabilisation, is reasonably long with these therapies, which is to say patients who respond typically maintain response for about a year’s time on molecularly targeted therapy, up to 18-24 months on PD1, PD-L1 blocking antibodies including this agent. So when you’re then asking the question of whether you’re able to produce more durable effects at present we have about one year follow-up on the patients who were treated in the early portion of this study but we may need substantially longer follow-up times to have a sense of whether we’re able to maintain those remissions. So the early evidence is that, yes, we’re able to achieve responses, as we would have predicted, now the issue is can we maintain them longer than we would otherwise.

Of course, initially drugs like ipilimumab were giving a tail that was quite promising and then it seemed that the targeted agents might not be giving the same tail. Are you saying you could extend that tail of the curve?

Yes. In fact, you’re right that a few years ago there was a sense that molecularly targeted therapies didn’t produce a tail but in fact three year, four year follow-up data now from some of the larger studies from years past actually shows a tail, even in patients who have never gone on to receive immune therapy or other treatments. So, in fact, we have a sense that a tail exists with both approaches, an important point because if either molecularly targeted therapy or immune therapy can produce a tail on their own when we think about developing these combination strategies we’d like not to give them to patients who would otherwise do perfectly well on the available treatments. So that’s an important side-bar which is a maintained focus on predictive biomarkers that will designate who those patients are.

Obviously this is a moving field, a rapidly moving field, and a very fascinating one. So what’s the message right now, the clinical message, coming out of this for doctors?

Simply put, sequential therapy with molecularly targeted treatment and immunotherapy as separate therapies given until they’re no longer effective, that’s the standard. As these drugs are approved and utilised in practice I wouldn’t want to suggest that there’s any basis for throwing them together now. Even in a patient who has a high burden of disease that’s symptomatic where one would like to be able to essentially throw all therapies at the situation to try to gain control and ideally maintain control, thereafter while those motivations are understandable the fact is we haven’t firmly ruled out antagonism of these therapies at this early date. In other words, is it possible that by putting the drugs together you’d actually get less than an additive effect of the treatments. It is only with more time, more patients and more follow-up that we’ll be comfortable that that’s not the case. So I would say in these early days this is not ready for clinical adoption at the moment but we do have a green light from the early safety signals to proceed now into full scale efficacy-focussed investigations which is where we’re transitioning now.

So you can consider adding in PD-L1 blockade?

Exactly.