Bristol's CheckMate 067 Revives Debate On Rationing Yervoy/Opdivo Combo

Yervoy/Opdivo combo proves significant survival benefit over Yervoy alone in Phase III frontline melanoma study, but is on par with Opdivo alone in patients with higher PD-L1 expression.

Survival data for Bristol-Myers Squibb Co.'s Yervoy/Opdivo combination in the CheckMate 067 study in frontline melanoma reinforce the regimen's potency, but also revive a debate about whether the regimen should be reserved for select subsets of patients, with others getting enough benefit from Opdivo alone or other treatment options.

The combination of Yervoy (ipilimumab), a CTLA-4 inhibitor, and Opdivo (nivolumab), a PD-1 inhibitor, received accelerated approval from FDA in frontline metastatic melanoma in October 2015 based on an improvement in the trial's co-primary endpoint of progression-fee survival. (Also see "Opdivo/Yervoy Data Show Role For Checkpoint Combos; ASCO Debates Biomarkers" - Pink Sheet, 1 Jun, 2015.) Approval was conditional on demonstrating a survival benefit in the study.

Combination use of Opdivo/Yervoy for melanoma has not been a major commercial focus, as the opportunity is dwarfed by the potential for combinations in lung cancer, but Bristol has cited the stabilization of Yervoy sales as a sign of steady use of the combination in melanoma. (Also see "Under Fire, Bristol Plans To Pick Up The IO Pieces" - Scrip, 26 Jan, 2017.)

CheckMate 067 randomized patients to three arms: Yervoy/Opdivo, Opdivo or Yervoy. But the study was only powered to compare Yervoy/Opdivo to Yervoy and Opdivo to Yervoy. It was not powered to test the combination against Opdivo alone, though exploratory analyses were performed for this comparison. Patients in the monotherapy arms were permitted to crossover to combination treatment, making it difficult to show a benefit in overall survival.

The combination met the co-primary endpoint of improved overall survival (OS) compared to Yervoy in the study, Bristol reported April 3 at the American Association for Cancer Research (AACR) annual meeting in Washington D.C.

In an April 3 note, Barclays analyst Geoff Meacham noted that the survival improvement in CheckMate 067 was highly statistically significant and concluded that the positive overall survival data from the study should offset concerns about the Opdivo/Yervoy regimen’s accelerated approval.

"We are encouraged that the Opdivo and Yervoy combination is showing a meaningful OS benefit, particularly given the potential for crossovers to muddle the results," Meacham said.

The positive outcome is clearly good news, but the data prompted questions about whether the difference between the combination and Opdivo alone was big enough to justify combination treatment, at great expense, and whether patients could be more appropriately selected based on subset data.

Survival results were statistically similar for the combination and Opdivo monotherapy in BRAF wild-type patients and in patients with more than 5% of PD-L1 expression.

Questions about whether the benefits of the combination will be worth the costs and toxicities compared to Opdivo alone came up in past data releases. (Also see "Yervoy/Nivolumab May Need Stronger Survival Data To Be First Choice In Melanoma" - Pink Sheet, 2 Jun, 2014.)

"Modestly Higher" Survival

Results reflecting follow-up of 28 months in CheckMate 067 were presented at the AACR meeting by James Larkin, a consultant medical oncologist at The Royal Marsden Hospital in London.

Median overall survival was not reached in the Yervoy/Opdivo and Opdivo arms, versus 20 months for Yervoy. The combination and Opdivo monotherapy arms both demonstrated a statistically significant reduction in the risk of death, by 45% and 37%, respectively. In an exploratory analysis, the combo also had a numerical improvement in OS compared to Opdivo, with a 12% reduction in risk.

At the two-year mark, median OS was 64% for the combination versus 59% for Opdivo monotherapy and 45% for Yervoy alone (see table below).

The two-year OS rate for the combination was only "modestly higher" than Opdivo monotherapy, though both were well above Yervoy monotherapy, Meacham noted.

"Consequently, given that the regimen had a Grade 3/4 adverse event rate of 58% vs. 21% for Opdivo monotherapy (and 28% for Yervoy), we suspect the debate will move to the relative risk/reward profile. However, we would also note though that as physician experience with the regimen increases and the side effect profile becomes better characterized, we anticipate that adoption of the regimen will continue," the analyst concluded.

Larkin noted that the vast majority of events resolved within three to four weeks. In the combination arm, the adverse event rate caused 40% of patients to drop out, primarily in the combination phase, but even drop-outs benefited from treatment in terms of response.

Looking For Guidance From Subsets

After the presentation of results at the AACR meeting, Larkin was asked whether the subset results should guide treatment.

The data are not definitive regarding use in patients with BRAF mutations, who can benefit from targeted therapies. In patients who were BRAF-wild type, OS was not reached in the combination and nivo groups and was 18.5 months for Yervoy. In an exploratory analysis of the combination compared to Opdivo in this subset, the hazard ratio was 0.97 in the combination's favor, not a statistically significant result.

In BRAF-mutant patients, the median OS was not reached for Yervoy/Opdivo or Opdivo alone and was 24.6 months for Yervoy. The hazard ratio in the exploratory analysis comparing the combination to Opdivo was numerically better at 0.71 but was not statistically significant. In the real world, these patients have the option of taking a BRAF inhibitor with a MEK inhibitor in the frontline setting; optimal sequencing of targeted vs. immunotherapy in this population has still not been determined.

In other cancer settings, most notably lung cancer, PD-L1 expression levels are used to identify patients most likely to respond to treatment – although this has not been relied on in melanoma.

In the quarter of trial participants who had PD-L1 expression over 5%, the combination looked very similar to Opdivo (hazard ratio 1.05) though the response rate was numerically higher. The combination looked better than Opdivo monotherapy for those with less than 5% expression (hazard ratio 0.84).

Larkin said that the subset data are intriguing, but was reluctant to say they could be used to guide treatment. The role of PD-L1 as a biomarker is still not yet really understood, he noted during a press briefing.

The role of the Yervoy/Opdivo combination in BRAF-mutant patients is also complicated by the question of whether this subset should instead be receiving a BRAF inhibitor and MEK inhibitor upfront rather than immunotherapy.

Larkin explained to Scrip that in his opinion, there are still not enough data to steer BRAF-mutant patients to the targeted BRAF/MEK combination instead of the immunotherapy combination. Studies are ongoing of different sequencing approaches and could provide answers in the future.

Larkin said that in his practice, he reviews aspects of treatments with patients and lets them make a decision. BRAF and MEK inhibitors are oral, whereas the PD-1 and CTLA-4 immunotherapies are given by intravenous infusion. The BRAF/MEK combination may be given continuously, whereas the Yervoy/Opdivo combination often causes severe side effects, resulting in early discontinuation in about one-third of patients.

The clinician said that once he reviews all of these details, patients usually are ready to make a choice.

"It's a sequencing question. It's not really a question of which is better… it's a question what is the strategy going to be in that patient to maximize control of disease, survival and quality of life," Larkin said.

Going forward, a major research focus will be developing biologically rational combinations for individual patients and trying to reduce toxicity, while maintaining efficacy, he added.

Michael Atkins, deputy director of the Georgetown-Lombardi Comprehensive Cancer Center in Washington, D.C., commented to Scrip that in his opinion the results in CheckMate 067 are good enough to support recommending the combination over Opdivo alone for many patients, particularly those who are PD-L1-negative or BRAF-positive.

The toxicities are manageable for experienced clinicians and the treatment is over faster, Atkins said.

The CheckMate 067 study doesn't address the question of whether an immunotherapy combination is better than BRAF/MEK upfront in BRAF-mutant patients – the National Cancer Institute-sponsored EA6134 study of sequential combinations will help answer that question, but CheckMate 067 supports the use of the combination in BRAF-mutant patients, he said.

CheckMate 067: Key Results
CheckMate 067 is a three-arm study that compared the Yervoy/Opdivo combo to Yervoy monotherapy andOpdivo monotherapy in the frontline treatment of 945 patients with metastatic melanoma. The study was powered to compare Yervoy/Opdivo to Yervoy and Opdivo against Yervoy. For patients in the monotherapy arms, crossover to treatment with the combination was permitted.
Endpoint	Yervoy + Opdivo	Yervoy	Opdivo
Median PFS	11.7 months	2.9 months	6.9 months
ORR	58.9%	19%	44.6%
Median duration of response	Not reached	18.2 months	31.1 months
Median OS, months	Not reached	20 months	Not reached
OS in BRAF Wild-type, months	Not reached	18.5 months	Not reached
OS in BRAF mutant, months	Not reached	24.6 months	Not reached
OS in ≥5% PDL1+	Not reached	18.5 months	Not reached
Two-year survival	64%	45%	59%
Rate of treatment-related Grade 3/4 events	58.5%	27.7%	21%