We presented this morning at the AACR plenary the first results, the definitive results on the primary endpoint, of the EORTC 1325/KEYNOTE-054 trial. That is a trial where in the adjuvant setting pembrolizumab is being compared to placebo in lymph node positive stage 3 melanoma patients after resection of their regional lymph nodes. The trial has an unusual design in the sense that it is set up to be able to address two questions and not just the simple question – what is adjuvant pembrolizumab going to have in terms of impact on relapse free or recurrence free survival? So, is the benefit going to be great because anti-PD1 is a very important drug in advanced melanoma, so will we see fewer relapses and will there be later relapses, these types of observations. So we report on this primary endpoint where there is a very clear impact on relapse free survival – there is a 43% overall reduction in the intention to treat population of relapses. So a big difference.
The second question in the trial is a very crucial and important question which makes this trial so unique because it’s never been done in an adjuvant trial. If you end up in the placebo arm in this trial you may get a relapse or any patient who gets a relapse is actually unblinded at the time of relapse. If it turns out, if it is demonstrated, that you are a placebo patient you then have free access to two years of treatment with pembrolizumab for the relapse that you have incurred. So it’s a crossover trial design which is extremely unusual in the adjuvant setting, basically because you can compensate for what you may have won by the adjuvant use of pembrolizumab by giving an effective drug at the time of relapse for those who haven’t had pembrolizumab. Then it may make up for everything and then in the end there may be not an overall survival benefit, overall survival difference. But yet this is, of course, a fundamental question because in the end the question is should we offer adjuvant therapy to the whole population or should we just offer it to those who need it at the time of relapse. So this question has not been answered by our report today because it’s too early for that. The overall survival data will reach maturity in about 2½ years from now. So we will have to wait for it but we finally have now a trial that addresses the most crucial question of all.
Then there was a co-primary endpoint in this trial, namely the patient population who have PD-L1 positive tumours. Why? Because in advanced melanoma there is an indication that those patients with PD-L1 positive tumours do better on pembrolizumab treatment than PD-L1 negative tumours. There seems to be no indication in this trial that this is the case; it seems like all patients, regardless of sub-stage of stage 3, regardless of PD-L1 expression yes/no and regardless of having a BRAF mutation yes/no, the hazard ratios that we are observing in all these subgroups are very close to one another. If you want to better understand what it means in absolute benefit at eighteen months we have looked at where are the curves with the pembrolizumab treated patients and where with the observation patients. Roughly we were very close to a 20% difference in absolute terms; we were very close in the intention to treat population – 72% of patients having a relapse free survival at eighteen months in the pembrolizumab arm whereas it’s roughly 52% in the placebo arm. So this is a very substantial benefit. What we notice is that the curves are still diverging so the pembrolizumab curve is much flatter than the descending curve of the placebo arm. This gives us great hope that it will translate into very significant distant metastasis free survival differences and it increases only the likelihood that it indeed will translate into overall survival benefit. But at least now by the crossover design we are really testing that question because otherwise it would have already been demonstrated, basically. So that’s the importance of the clinical trial.
What is the toxicity like?
What makes anti-PD1 such a great drug is its therapeutic index, so significant efficacy and relatively low toxicity. So with the immune checkpoint inhibitors we are mostly looking at immune related adverse events and we know the frequency and the grade of immune related adverse events is a problem with ipilimumab. But for the anti-PD1s, nivolumab or pembrolizumab, it’s a completely different story. So we see much fewer immune related adverse events and lower grades of immune related adverse events so they are easily manageable and they are easily resolved. Out of the 514 patients that were randomised into the pembrolizumab arm there are only 39 patients who had a grade 3 or 4 immune related adverse event. So what everybody immediately asks is ‘How’s the colitis?’ because the colitis grade 3/4 is like 6%, 7%, 8% with ipilimumab, here it’s 2%. So we only saw ten patients with a grade 3/4 colitis. Another question is always, ‘How is the hypophysitis?’ so it knocks out your pituitary gland. In ipilimumab it’s between 4-5%, here it is 0.6% so we saw only three patients with a hypophysitis.
What you do see as a frequent endocrinopathy that is immune related is hyperthyroidism as an end stage of a short hyperthyroidism and then sliding into hypothyroidism situation and we see that in 20% of patients. But this is just grade 1/2; it’s basically just lab values, differences, and if need be it can be very easily treated with thyroxin.
So overall we have pretty much, relatively speaking, an ideal drug because of this therapeutic index.
Will it be expensive?
The great thing is that both nivolumab, because nivolumab was approved for this indication in December 2017 and pembrolizumab we expect to be approved in 2018, so we have a competition. Normally competition should drive down the cost because that’s how it should be so it’s very important that we have alternatives like this. The original approval for nivolumab was for more complex dosing, 3mg/kg, and that means that you actually also waste material, the strength, and it was given every two weeks. The strength of the pembrolizumab schedule is it’s every three weeks and it’s a flat dose so you don’t lose anything, you just have a vial to give to the patient. But nivolumab was just, only two weeks ago, also approved for a flat dosing every two or every four weeks. So that initial advantage is no longer there and therefore I foresee that it will be a pricing decision also, among other things.