Medicine Matters oncology

The CheckMate 227 data, I am sure you know the design, is a trial which has met its primary endpoint of improving overall survival versus platinum-based chemotherapy in the subgroup of patients which was analyzed in this primary endpoint, which is the PD-L1 positive population. It was 1,200 patients, a large trial, showing a sustained benefit with major overall survival curve. Even this benefit growing over time, like this usual shape of immunotherapy curve.

So basically, the message is this is a new option. You have other options. You can give in high PD-L1 and very high PD-L1 only pembro, maybe only atezo in the future. And you can give in all patients chemo and IO. But this is a new opportunity that the patient needs to be aware of. And we can discuss it among ourselves, but also with the patients every time you have this informed discussion about decision-making process.

What was interesting in 227 is there was another big subgroup of 550 patients with PD-L1 negative tumors, which was considered separately, like a separate trial, a separate cohort. Which also received ipi/nivo versus chemo. And the same benefit was seen.

So it was not primarily the end point, but the exact same benefit. And I would say similar ipi/nivo result were seen, with the hazard ratio even lower than 0.7. So it means that, despite the trial design, it looks like the strategy works irrespective of PD-L1.

Toxicity? Well, very different from chemo. Very different. But statistic in terms of number, same number of all grade and grades 3 and 4 toxicity in chemo versus nivo/ipi. Of course, if you consider chemo IO, then the toxicity level is higher than nivo/ipi, and it's important to stress it.

Last but not least, what I prefer in this trial is the duration of response. For example, in the primary cohort of positive PD-L1, almost 50%-- half of these patients-- are still in response after two years of treatment. So it means that you start ipi/nivo. If you respond, you have one chance out of two to still be in response at two years.

These peaks suggest and potentially let us assume that there will be a long-term benefit of the component, the contribution of the component of CTLA-4 in the disease. Which was seen in renal, in melanoma. It's being now trying to be observed in ovarian. So all this supports this durability, thanks to CTLA-4.

What are the next steps?

The next step is-- I freely agree with my colleagues that it's one option. And it's wonderful, because five years ago, we didn't have anything other than chemotherapy. So we shouldn't complain. We have many options.

But the next step is to try to understand really which patient should receive which treatment. And of course, listen to the patient, because the patient preference perspective will be very important. But we need, in terms of biomarkers, because unfortunately, in 227, we couldn't identify any biomarker predicting for the benefit of ipi/nivo versus chemo.

So we need to move ahead to try to have an algorithm with logical flow of patients with preferred treatment. And we still are not there. So that's why we discuss the data differently. It's because some say, I'm in favor, not in favor. I don't care. We have many treatment options.

The future? We're in the future. We need to sit together and find data, create the data, build the data set, in order to know what is the best treatment as much as possible. And if possible, all patients coming to our office.