Data on 2-year survival in the phase III CheckMate 067 trial testing nivolumab plus ipilimumab showed a slight survival benefit for the combination over nivolumab monotherapy—but the difference may not be dramatic enough to justify using the combination given the added side effects.

A phase III melanoma study testing two Bristol-Myers Squibb drugs—the PD-1 inhibitor nivolumab (Opdivo) and the CTLA-4 inhibitor ipilimumab (Yervoy)—has now run long enough to yield much-anticipated 2-year survival data, which were presented on April 3 at the American Association for Cancer Research Annual Meeting 2017 in Washington, DC.

According to findings from the 945-person CheckMate 067 trial, patients with advanced melanoma who received a combination of both drugs achieved a 2-year survival rate of 64%, compared with 59% for those who received nivolumab alone, and 45% for those who received ipilimumab alone. After 28 months, median overall survival (OS) had not been reached for either nivolumab-containing study arm; it was 20 months for those taking ipilimumab alone.

As expected, the toxicity of combination therapy was also the highest, “consistent with earlier experience,” reported James Larkin, MD, PhD, of the Royal Marsden Hospital in London, UK. For those taking the combination, 58% experienced serious adverse events, versus 21% and 28% for those who received single-agent nivolumab or ipilimumab, respectively.

Early results from the CheckMate 067 trial, published in 2015, showed that the response rate was greatest, and the progression-free survival (PFS) was longest, for those taking the combination. That finding led to accelerated approval from the FDA and the European Medicines Agency. These newer OS data confirm the clinical benefit of the combination, at least over ipilimumab monotherapy, which experts say should lead to full approvals.

“CheckMate 067 was a positive study that met both of its co-primary endpoints,” namely OS and PFS, said Larkin. “Descriptively, the combination showed better overall OS, PFS, and more durable response outcomes than nivolumab alone, with consistent results seen across clinically relevant subgroups, such as those with low PD-L1 expression, patients with elevated [liver enzymes], and those with tumors that were BRAF-mutated.”

Howard Kaufman, MD, of the Rutgers Cancer Institute of New Jersey in New Brunswick, said he was disappointed not to see more of a difference between nivolumab with and without ipilimumab. “I think we have to tell patients, ‘Well, this is what we know about it: So far, there is a small but real survival benefit if you get the combination, and it's going to come at a significantly increased risk of having side effects,’” he said.

However, the study was not powered to demonstrate superiority of the combination over nivolumab alone. That statistical quirk of the study design stems from the fact that when CheckMate 067 began enrolling patients in 2013, ipilimumab was the only checkpoint inhibitor on the market. That made it the putative standard against which other experimental regimens could be tested, and the idea of statistically pitting nivolumab alone against the combination was never written into the protocol.

“That's really suboptimal, as essentially all investigators in the field feel that ipilimumab was an inferior comparator arm for benchmarking this combination regimen,” said Jason Luke, MD, from The University of Chicago Medicine in Illinois.

Even so, Luke said, the descriptive comparisons from the trial will “clearly impact the standard of care.” He expects many patients may continue to receive anti–PD-1 and anti–CTLA-4 drugs sequentially, because other clinical trials have suggested that this timing produces a clinical benefit similar to that of combination therapy in patients who receive the drugs simultaneously—but with less toxicity.

That may change, however, with longer follow-up. According to Larkin, “the inflection point on a survival curve is probably around about 3 years,” and future analyses could show added benefit of the combination therapy. “We'll see where we are at 3 years and whether the curves separate more or not.”

Larkin said he hopes the field will find biomarkers that can help predict which patients would be most likely to benefit from this and other immunotherapies. Based on the CheckMate 067 results, Larkin described BRAF-mutation status and PD-L1 expression levels as “very intriguing” possibilities, but, he added, “the data are still premature.” –Elie Dolgin