AACR – Long-term melanoma data open another Opdivo weak spot

Bristol-Myers Squibb may have celebrated an overall survival win for Opdivo plus Yervoy in melanoma in two-year follow-up data from Checkmate-067, but an overlooked aspect is the study’s failure to provide convincing evidence that the two-drug combination beats single-agent Opdivo.

Some analysts argue that this could uncouple the two checkpoint inhibitors in this indication and encourage oncologists to prescribe Opdivo first followed by Yervoy on progression. This change could benefit Merck & Co’s rival Keytruda, and open up yet another weakness in Bristol's immuno-oncology strategy.

Bristol presented the long-term follow-up data from Checkmate-067 at AACR on Monday, having previously used PFS data to gain approval of the combo in 2015. Analysts were initially relieved that the OS data supported the combo, which is now standard of care, but there is a growing belief that specific comparisons might undermine Bristol’s strong competitive position in this indication.

The 945-patient phase III trial had three arms – the combination and both agents separately – but the statistical plan had only two comparisons in the primary analysis: combo versus Yervoy alone, and Opdivo versus Yervoy as single agents. The more interesting comparison of combo versus Opdivo alone was not included as a primary, since Yervoy was standard of care when the trial began.

Combo vs sequence

It is also important to note that patients in Checkmate-067's monotherapy arms could cross over to the other on progression, and most would have done so, so the study effectively compares the two drugs given concurrently with their use as a sequenced combination.

The data presented at AACR show that the combo conferred a 45%, and Opdivo alone a 37%, reduction in risk of death relative to Yervoy, in both cases to a high level of statistical significance. However, the hazard ratio for the combo relative to Opdivo alone was 0.88, not significant.

In terms of across-trial comparison, Merck’s Keynote-006 study showed a 37% reduction in risk of death for a two-week regimen of Keytruda and a 31% reduction for a three-week regimen, relative to Yervoy. Although the hazard ratio for the more effective regimen was identical to that of Opdivo in Checkmate-067, the comparison is tricky because Keynote-006 did not permit crossover.

Checkmate-067 also showed an unexpected result when analysed by PD-L1 expression, with the combo favoured relative to Opdivo only in the low-expressing subgroup, and possibly being disadvantageous in higher expressers.

Taken together the data provide little evidence of any benefit for the combo versus single-agent Opdivo, and this is before side effects are considered. Any subsequent shift in practice to use an anti-PD-1 first line would immediately remove the commercial disadvantage that Keytruda has by virtue of not being indicated for use in combination with Yervoy.

Indeed Cowen analysts said physicians they had spoken to believed that most clinicians would now switch to PD-1 monotherapy first line, a concerning development for Bristol-Myers.

Future combinations

If Keytruda does gain share from Opdivo in first-line melanoma it could pave the way for future combinations involving the drug. Merck has two phase III studies under way involving its anti-PD-1 therapy in melanoma – with Incyte’s epacadostat and Amgen’s Imlygic.

Anticipation of this could explain Bristol’s decision to add an anti-PD-1 relapsed/refractory melanoma cohort to the Echo-204 phase I/II study of the Opdivo/epacadostat combination, as part of a wider expansion of its Incyte collaboration (Incyte dramatically ups the immuno-oncology combo ante, April 3, 2017). However, Merck has an exclusivity period with Incyte in melanoma that would surely limit Bristol’s ability to test an Opdivo/epacadostat combination. 

After the dramatic shift in lung cancer in favour of Merck last year, it looks like this company could now make greater inroads into one of Bristol’s strongest Opdivo indications.

To contact the writer of this story email Robin Davison in London at [email protected] or follow @RobinDavison2 on Twitter