Karen Kelly, MD: Right, complicated indeed. You’ll remember that CheckMate-227 is really 2 cohorts—those who don’t have, or have less than 1%, expression; and those with greater than 1% expression. Each of those cohorts were randomized: for >1%, to nivolumab alone, nivolumab plus ipilimumab, or chemotherapy; and for <1%, to nivolumab plus ipilimumab, nivolumab plus chemotherapy, or chemotherapy. And so, this trial was going along. We saw the results of CheckMate-012, which was a negative trial. [Note: CheckMate-012 is a multi-arm Phase 1b trial; positive results have been reported for frontline nivolumab plus ipilimumab (Hellmann, Lancet Oncol, 2017) and for nivolumab plus chemotherapy (Laurie, WCLC 2017)]. And then we saw the TMB results come in, showing efficacy in those patients with high TMB. TMB has now emerged as a potential biomarker for efficacy for immune therapy. Certainly, we do need more biomarkers. We all criticize PD-L1, but it is really the best biomarker that we have, to date. We’re looking for better biomarkers.

What happened was that this trial was amended to include this co-primary endpoint of progression-free survival in the high TMB patients. They looked at nivolumab/ipilimumab versus chemotherapy. This was a subset of this very large 1200-patient [Note: N = 1739] trial. The other thing to point out is that you have to remember that not all patients could get TMB done. Only about 58% of the patients had TMB. We can talk about the issues with TMB in a minute. But in this group of patients, about 300 patients, we saw that the progression-free survival favored nivolumab/ipilimumab in the high TMB group, which was defined as greater than 10 mutations per megabase, versus the chemotherapy, with a hazard ratio of about 0.59. Here, again, we saw that objective response rates were doubled as well. That’s a consistent theme that we’re seeing with immunotherapy.

However, we don’t have overall survival data here. Until we see some overall survival data, I think the data is provocative. [Note: Early descriptive analysis showed encouraging overall survival for nivolumab plus ipilimumab versus chemotherapy in patients with high TMB ≥10 mut/Mb (HR 0.79; 95% CI: 0.56 to 1.10)] We can talk about what we think of TMB as a marker. I also want to point out that even though they didn’t put this in the primary manuscript—it was in the appendix—we did see the toxicity profile for this particular group of patients. The grade 3/4 toxicity for nivolumab/ipilimumab versus chemotherapy was about the same—at 37% and 36%.

Transcript Edited for Clarity