The Checkmate 227 study was, again, presented here at the ESMO ’19 meeting as an important study. It’s an important study that we’ve been waiting a long time to hear the final overall survival results for. We’ve heard a number of secondary endpoints and a co-primary endpoint of PFS by TMB status. The study is a complex study, it took patients in the front line setting and randomised them contingent on their PD-L1 status. In the PD-L1 positive group it randomised them to either nivolumab ipilimumab, nivolumab monotherapy or chemotherapy.

When we look at the PD-L1 positive group, that group is positive but that is a key secondary endpoint and not a primary endpoint. So nivolumab ipilimumab is positive versus chemo alone but, of course, we have to look at what is the influence of PD-L1 status. When we look at that further that outcome is very much driven by the PD-L1 50% positive group. So, to me, this drug has a role but it has a role in the PD-L1 50% positive group. But in that group we already have available pembrolizumab, we will shortly have atezolizumab, I hope, on the basis of the IMpower110 data and we are already using chemo pembrolizumab triplets. The hazard ratio for benefit was very similar to what we already have and the adverse event profile showed a higher rate of adverse events for double immunotherapy compared to single immunotherapy. So that needs to be borne in mind in decision making.

The second aspect of that study was the PD-L1 negative population. In the PD-L1 negative population patients were randomised to either standard chemotherapy, chemo plus nivolumab or nivolumab ipilimumab. Now, that part of the study was positive for overall survival for nivolumab ipilimumab but it wasn’t a primary endpoint, it wasn’t a pre-specified secondary endpoint, it was an exploratory endpoint. So we need to bear that in mind when interpreting the data.

When we look at that data I think it is a potentially important piece of data because in this setting the other option that we have is a chemo pembrolizumab triplet. One needs to think about the appropriate toxicities that we’ll have with either nivolumab ipilimumab or chemo pembrolizumab and the hazard ratio was very similar for nivolumab ipilimumab versus chemo to chemo triplet versus chemo. So we do have options in that scenario.

What was very clear, though, is that tumour mutational burden was not predictive of benefit for nivolumab ipilimumab, it was prognostic. Patients with high TMB did better with chemo and did better with nivolumab ipilimumab and patients with low TMB did slightly worse with nivolumab ipilimumab and slightly worse with chemotherapy. It wasn’t predictive of benefit for nivolumab ipilimumab.