Opdivo’s lung cancer latest: one fail, one shrug

Bristol-Myers Squibb’s byzantine Checkmate-227 study yields one flop and one technical success that, in reality, counts for nothing.

Most biopharma watchers have now accepted that Merck & Co’s Keytruda has an unassailable grip on first-line metastatic lung cancer. Yet the predictable failure of Opdivo in this setting initially sent Bristol-Myers Squibb down, suggesting that some investors had, after all, held out hope.

The result, revealed by Bristol after market close yesterday, concerns the extensively overhauled and confusing Checkmate-227 trial. To be fair, only one analysis was a clear fail, but that was the important readout; Bristol also claimed a hit in a separate part of '227, but this concerned an effectively irrelevant setting.

The '227 study had been overhauled to look initially at a combined subgroup cut by tumour mutation burden (TMB), and gave the company an arguable win backing an Opdivo/Yervoy combo. Still to read out was overall survival for Opdivo/Yervoy in PD-L1 ≥1% expressers (part 1a) and for Opdivo plus chemo in all non-squamous subjects (part 2).

Yesterday Bristol said part 2 had failed, yielding a non-significant median OS of 18.83 months versus 15.57 months for chemo alone. Had this succeeded it could have put Opdivo plus chemo on a par with Keytruda, which thanks to the stunning result of Keynote-189 has a stranglehold on the first-line chemo combo setting.

As for the supposed success, this related to '227 part 1a. But this is irrelevant because the PD-L1 ≥1% population can already get Keytruda monotherapy, based on the Keynote-042 trial, and there is no logic in adding on top of PD-L1 blockade a highly toxic drug like Yervoy.

On an analyst call today Bristol insisted that the latest result potentially differentiated its approach by virtue of chemo sparing. This extraordinary claim is presumably backed by an exploratory analysis of part 1a in PD-L1-negative subjects, though even so the rationale of replacing chemo with Yervoy is questionable.

Bristol lost 2% in the premarket, before climbing 3% today after second-quarter financials were released.


Of course, even the earlier purportedly positive TMB analysis (part 1 combined) is now shrouded in controversy. In January Bristol pulled its US filing for TMB-high NSCLC after an FDA data request, suggesting that the US regulator saw little relevance in TMB status (Bristol makes investors squirm a little more, January 24, 2019).

The data the FDA had asked for concerned '227 part 1a, and were not available at the time. They are available as of yesterday, but even though they are positive all they do is reiterate the importance of PD-L1 status.

An interesting point in the post-mortem over '227 part 2 is the extent to which Keytruda’s benefit in Merck’s corresponding chemo combo trial, Keynote-189, was driven by underperformance of the chemo comparator cohort.

Cross-study comparison in first-line metastatic NSCLC
Trial Active treatment Population mOS data Result
Checkmate-227 part 1 Opdivo + Yervoy TMB high (≥10mut/Mb) 23.0mth vs 16.4mth Unclear*
Neptune Imfinzi + tremelimumab TMB high (≥20mut/Mb) H2 2019 TBC
Keynote-042 Keytruda PD-L1 ≥1% 16.4mth vs 12.1mth Hit
Checkmate-227 part 1a Opdivo + Yervoy PD-L1 ≥1% Not disclosed Hit
Keynote-189 Keytruda + chemo Non-squamous all-comers 22.0mth vs 10.7mth Hit
Checkmate-227 part 2 Opdivo + chemo Non-squamous all-comers 18.8mth vs 15.6mth Miss
Keynote-407 Keytruda + chemo Squamous all-comers 15.9mth vs 13.2mth Hit
Checkmate-227 part 2 Opdivo + chemo Squamous all-comers 18.3mth vs 12.0mth Exploratory**
Poseidon Imfinzi + chemo +/- tremelimumab All-comers H2 2019 TBC
Checkmate-9LA Opdivo + Yervoy + chemo All-comers H1 2020 TBC
*Called into question by a similar result in TMB low subjects and the FDA's request for more data; **number of subjects in this non-prospective analysis not disclosed.

For its part, Bristol today bemoaned the fact that chemo alone in '227 outperformed expectations. Still, this was surely down to control arm subjects getting second-line anti-PD-(L)1 therapy, reflecting the real-world setting that all companies have to live with.

To the extent that it is still worth trying to compete against Keytruda in first-line metastatic NSCLC, Bristol investors now have the Checkmate-9LA study to look forward to. This tests Opdivo plus Yervoy and chemo, and is due to read out next year.

And Astrazeneca today confirmed that its Neptune and Poseidon trials were still on track to read out in the second half. On today’s second-quarter call Astra made no mention of the failed Mystic trial, which it had claimed showed a benefit if cut by TMB, suggesting that it was waiting for Neptune and/or Poseidon before deciding on a way forward.

Perhaps if you wait long enough a Keytruda competitor will deliver an unequivocally positive result.

Share This Article