The KEYNOTE 427 study is a study of frontline pembrolizumab in untreated renal cell carcinoma patients. It has two cohorts: it has a clear cell cohort A and a non-clear cell cohort B population. The study presented at ASCO this year is looking at that study and using tissue samples to do biomarker analyses. The clinical endpoints for that study have previously been reported. The cohort A population of 110 clear cell patients has been shown to have an overall response rate of 37%.
So we’re looking at that patient population with an RNA-Seq based assay and defining coherent gene signatures, multi-gene expression profiles, that include a T-cell inflammation profile and ten other consensus gene signatures, and looking for an association with overall response rate in this cohort. The findings for the T-cell inflammation gene signature did associate with overall response rate but none of the other ten consensus signatures. There was not a strong association with other clinical endpoints, progression free survival or overall survival.
So, again, this study of 110 patients, it is a smaller study than some of the large frontline phase III trials, but there has been an effort to do biomarker analysis across many of the frontline clinical studies for renal cell carcinoma. Our data does correspond to observations that have been made from the combination of atezolizumab, an anti-PD-L1 antibody, plus bevacizumab and observations made from the JAVELIN study of avelumab plus axitinib that T-cell inflammation at baseline does seem to associate with superior clinical outcomes.
Now, that isn’t true across all of the analyses performed in the same way. At ASCO this year there was also biomarker data presented from the CheckMate-214 study, that’s frontline ipilimumab nivolumab. So pure immunotherapy, if you will, in the frontline setting did not show a strong association with any of the biomarker analyses that were tested in that cohort. There was also a very large dataset of PD-1 blockade monotherapy from pooled studies, primarily from previously treated patients receiving PD-1 monotherapy, that again had limited ability to define clear-cut biomarker observations. There is a T-cell subset within that overall analysis that does seem to associate again with improved clinical outcomes.
So the theme of T-cell inflammation at baseline within tumours that may predict a better overall outcome, whether it’s overall response, whether it’s progression free or overall survival, depending on how the study has been designed. But what to do with that information? Is this really poised to change our clinical practice? At this point the answer is no; the patients that do not have the biomarker still generally fare rather well with our checkpoint blockade. So although it’s interesting and we can enrich for the response phenotype it’s not a discrete enough separation of the population that it’s going to be a marker that is going to be used as a diagnostic to include patients for checkpoint based therapy.
So the field is still in search of more discrete or more powerful markers that might better select patients that are going to have uniquely good outcomes or select patients that are destined to have primary refractory disease where we might move in a different direction with their primary therapy. So although interesting in providing some context about how the drugs work, not yet ready to impact our routine clinical management.