#ESMO20: Pushing into frontline, Merck and Bristol Myers duke it out with new slate of GI cancer data
Having worked in parallel for years to move their respective PD-1 inhibitors up to the first-line treatment of gastrointestinal cancers, Merck and Bristol Myers Squibb finally have the data at ESMO for a showdown.
Comparing KEYNOTE-590 and CheckMate-649, of course, comes with the usual caveats. But a side-by-side look at the overall survival numbers also offer some perspective on a new frontier for the reigning checkpoint rivals, both of whom are claiming to have achieved a first.
Notably, Bristol tested Opdivo plus chemotherapy for three types of metastatic cancers — gastric, gastroesophageal as well as esophageal adenocarcinoma, while Merck enrolled patients with locally advanced or metastatic esophageal cancer and gastroesophageal junction cancer.
“So the population is much broader in CheckMate-649,” Bristol CMO Samit Hirawat notes.
Nonetheless Roy Baynes, his counterpart at Merck, considers KEYNOTE-590 broad in its own way.
“This was a really important trial for us,” he says, which “pretty much admitted all comers.” And Merck counts it as a major success.
“Basically,” Baynes adds, “every single group that was pre-specified hit at a highly significant level.”
PD-L1 status, though, defines a broader group of patients more likely to benefit.
In the ‘590 trial, investigators concluded that Keytruda combined with chemo reduced the risk of death compared to chemo alone by a significant margin of 27% among all comers. Median overall survival was 12.4 months for the combo compared to 9.8 months in the chemo control arm.
There was also a 35% reduction in the risk of disease progression or death. Looking at PFS versus chemotherapy in all randomized ESCC patients in the study, the combo hit a hazard ratio of 0.65, while those with tumors expressing PD-L1 fared better, with an HR of 0.51.
Narrow it down to ESCC patients whose tumors express PD-L1 (CPS ≥10) and the OS number jumps to 13.9 months compared to 8.8 months.
With Opdivo, median OS among the all-randomized population was 13.8 months versus 11.6 months for those on chemo, translating to a hazard ratio of 0.80. Investigators reported a PFS benefit for the broad group as well, although they didn’t spell out the numbers
Zoom into the PD-L1 positive group (CPS ≥5), and the HR is down to 0.71 with median OS of 14.4 months and 11.1 months, respectively. Within the cohort, adding Opdivo to chemotherapy reduced the risk of progression or death by 32% (PFS 7.7 months vs. 6.0 months on chemo alone).
“That 3.4 month-improvement is certainly very meaningful in a very hard, very challenging disease to treat,” Hirawat says.
The next step for both companies will be to bring the data to regulators. But while Bristol is likely seeking a label expansion that covers all upper GI cancers, Hirawat highlights that Merck only manages to prove that Keytruda is non-inferior, but not superior, to chemotherapy in advanced gastric or gastroesophageal junction cancer with the previously presented KEYNOTE-062 data.
There’s still plenty to prove. And ultimately physicians could have their own takes on the data.
“I think we’ve always taken the view that this is not a winner-take-all market,” Baynes says.