Good afternoon, my name is Dr Yelena Janjigian, I’m a physician scientist and medical oncologist at Memorial Sloan Kettering Cancer Center in New York city. It’s a pleasure to discuss with you the results of our biomarker analysis of CheckMate 649 clinical trial samples that we presented the data at the AACR meeting this year.

By way of background, CheckMate 649 was a large phase III study that already changed practice for countless individuals worldwide. A combination of FOLFOX plus nivolumab was FDA approved and approved in more than 50 regions based on the results of CheckMate 649 data. This study was also studying the combination of nivolumab plus ipilimumab in the first line setting. The pre-specified boundary for overall significance, or for overall survival significance, was not met and this arm did not meet the significance and was a negative arm. 

We presented the results from the biomarker of those patients treated on nivolumab/ipilimumab and compared with outcomes to chemotherapy. We know that in the intent to treat population 25% of patients do survive long term with long-term follow-up of combination of nivolumab/ipilimumab and that if patients respond the duration of response to nivolumab/ipilimumab is much more durable than with chemotherapy. So we were very interested in studying other subpopulations of patients that particularly benefit from nivolumab/ipilimumab.

We looked at whole exome sequencing and bulk RNA-Seq in the pre-treatment samples obtained prior to initiation of nivolumab/ipilimumab or FOLFOX or CAPOX chemotherapy. More than half of samples from baseline were available for testing for whole exome and bulk RNA-Seq and we had MSI testing done by PCR on all samples at baseline: more than 90% of samples were available for MSI testing. We looked at baseline patient characteristics, they were well balanced among the two arms of patients analysed, suggesting that this population, these analyses, will reflect on the rest of the study population and intent to treat population. 

Firstly, based on MSI status, not surprisingly, we demonstrated with longer follow-up, a minimum of four years of follow-up, there were sustained separation of the curves for the MSI high population. These patients clearly benefit from immune checkpoint blockade in first line setting compared to chemotherapy. The hazard ratio was 0.29, highlighting the importance of MSI testing in this patient population. 

We then looked at tumour mutational burden by whole exome sequencing using ≥199 mutations per exome to define a high TMB population, which equates to about ten mutations on a FoundationOne assay. We found, not surprisingly, that TMB high patients also benefitted to a greater degree from immunotherapy, immune checkpoint blockade, nivolumab plus ipilimumab, in first line setting. What’s important to highlight is the majority of TMB high patients were also MSI high. So in this population of 21 TMB high patients, of those 16 were MSI high. 

Then, going on to looking at gene expression signatures by bulk RNA-Seq obtained from the baseline samples, we looked at five different types of signatures and looked at both stromal and endothelial expression, proliferation rates in the tumour. Also, because CTLA-4 blockade is known to augment or to change Treg response, we looked at Treg signatures. So looking across these different gene signatures, we found that some signatures are predictive of survival benefit for nivolumab/ipilimumab. Particularly having high proliferation gene signature, high inflammation gene signature and high baseline Treg signature were predictive of benefit in overall survival from nivolumab/ipilimumab. The hazard ratio was close to 0.6. 

Then, conversely, low stromal signature – we looked at cancer-associated fibroblasts as one of the genes in that gene signature – low stroma and low endothelium or angiogenesis were predictive of overall survival benefit from nivolumab/ipilimumab. We know that the fibroblasts and angiogenesis can be immunosuppressive so that’s also not surprising but it was very nice to see.

In summary, we presented at AACR CheckMate 649 biomarker analysis from the nivolumab/ipilimumab versus chemotherapy arm and we looked, based on whole exome and RNA-Seq, on various predictors of outcome from nivolumab/ipilimumab. We found that the MSI high population benefits the most and certain gene expression signatures that highlight high inflammation, high proliferation, high Treg signature, low stroma and low angiogenesis signatures, those patients may benefit from nivolumab/ipilimumab in first line setting. 

These data are interesting and important to publish. They are not immediately translatable into clinical practice at this time. We need to do further validation in a prospective manner to understand which patients may benefit the most from immune checkpoint blockade in the first line setting without chemotherapy, particularly from combined immune checkpoint blockade such as nivolumab and ipilimumab.

Thank you for your attention, it was a great opportunity to study these samples. We are, as always, thankful for our patients, their families and, of course, the investigators who made this research possible.