I’m Matt Galsky, I’m a medical oncologist at the Icahn School of Medicine at Mount Sinai in New York. I wanted to talk a little bit about the IMvigor130 study. IMvigor130 is a randomised phase III study that enrolled patients with metastatic urothelial cancer and randomised patients to one of three arms – atezolizumab alone, atezolizumab plus chemotherapy or chemotherapy alone. This trial was really meant to answer two important questions simultaneously – is there still a role for single agent immune checkpoint blockade in metastatic urothelial cancer? Should we be giving all patients combination treatment? And is this really better than using chemotherapy first and using immune checkpoint blockade in the second line which is a standard strategy?
The trial had a couple of interesting turns. One of the interesting turns was that even before the trial completed enrolment the data safety monitoring committee of this trial, along with a parallel similarly designed trial, KEYNOTE-361, had noticed that patients who were randomised to the single agent atezolizumab arm who had low PD-L1 expression were dying at a faster rate than patients randomised to chemotherapy. So that led the FDA to review that data and before the trial was even completed it changed clinical practice, along with the KEYNOTE-361 study, in that in some regions of the world the labels for immune checkpoint blockade in the frontline cisplatin ineligible setting had changed to restrict use from an all-comer population to a PD-L1 high defined population. So that happened before the study was completed.
We now have the first analysis of the results of the study that was presented by Enrique Grande at ESMO in 2019. There was some updated data including biomarker data at ASCO this year. The initial analysis of IMvigor130 included a comparison of the chemotherapy plus atezolizumab arm versus the chemotherapy arm. That showed a significant improvement in progression free survival which was a co-primary endpoint. Overall survival showed a similar effect size but the p-value didn’t meet the specified interim threshold for significance and so the overall survival analysis is being carried forward until the final analysis.
The biomarker data that was presented at ASCO this year really sought to focus on the atezolizumab monotherapy arm, even though that hasn’t been fully evaluated formally yet, based on where it is in the hierarchical testing scheme. But it really asked the question whether or not some of the biomarkers for single agent immune checkpoint blockade that had previously been defined based on single arm datasets actually held up in a large randomised study, one. And, two, because of the nature of our randomised study we could help define whether or not these biomarkers were predictive or simply prognostic. These biomarkers included PD-L1 expression, of course, which was a pre-specified analysis as per the protocol. Also included tumour mutational burden and it also included a fibroblast TGF-β response gene signature that has been associated with resistance to single agent immune checkpoint blockade.
Each of these biomarkers was validated in this randomised dataset. Each of them was shown to likely be predictive rather than just prognostic as they favoured atezolizumab, TMB and PD-L1 compared to chemotherapy whereas the fibroblast TGF-β gene signature was associated with resistance compared to chemotherapy with atezolizumab alone. Interestingly, when you combined TMB and PD-L1 expression, so about 15% of the patient population harboured both a high TMB and increased PD-L1 expression, when you combined both of those the hazard ratio for overall survival with atezolizumab monotherapy was 0.22 compared to chemotherapy. So there’s probably a small patient population that can be biomarker defined that has tremendous benefit from single agent immune checkpoint blockade versus chemotherapy. This was a retrospective analysis but in the context of a randomised study still a retrospective analysis are not fully suitable for informing care, although we do obtain PD-L1 expression data routinely in the context of clinical care these days as we are increasingly determining TMB through commercial genomic sequencing providers. So this information is typically more commonly available these days and certainly it does give some thought in terms of who might benefit the most from single agent immune checkpoint blockade in metastatic urothelial cancer.