Nicholas Hornstein, Assistant Professor at Northwell Health, shared a post on LinkedIn:
“This one stopped me in my tracks
We now have multiple reference points for single-agent IO in MSI-H mCRC. KEYNOTE-177 with pembrolizumab set the first-line bar.
CheckMate 8HW with nivolumab showed very durable disease control.
And now we have COMMIT.
ASCO GI 2026, Abstract 14
COMMIT phase 3 in 1L dMMR/MSI-H mCRC
FOLFOX + bev + atezolizumab vs atezolizumab alone
PFS 30.0 vs 4.3 months, HR 0.44
ORR 80.6 vs 46 percent
DCR at 12 months 62.9 vs 32.4 percent
That disease control rate in the atezolizumab arm is abysmal for this biology.
So what’s going on?
- Is this finally closing the PD-1 vs PD-L1 debate in MSI-H colorectal cancer?
- Is this a molecule issue, a trial design issue, or patient selection? COMMIT was amended post KEYNOTE-177 and ended up much smaller than planned, so the details matter.
- What does this mean for ATOMIC? If PD-L1 monotherapy underperforms this badly in metastatic disease, the implications spill into earlier-stage strategies.
This needs a very close look when the full dataset is presented. Something here does not add up.”
More posts featuring Nicholas Hornstein.