Surprise! There’s a tumour in which Keytruda doesn’t work
The Merck & Co drug’s failure in small-cell lung cancer prompts a theory and leaves the way clear for Roche and Astrazeneca.
Keytruda’s domination of the anti-PD-(L)1 market sometimes seems so comprehensive that it is surprising to see it flunk a clinical trial. Yesterday this is what happened, with the Keynote-604 study failing to show an overall survival benefit in first-line small-cell lung cancer.
The setback seems particularly galling because two competitors that are in desperate pursuit of Keytruda – Roche’s Tecentriq and Astrazeneca’s Imfinzi – have already succeeded in this tumour type. And the disparity has already seen one bank turn to scientific papers for an answer.
Citing a retrospective analysis published in Jama, Leerink has argued that in tumours characterised by poor PD-L1 expression and less fit patients – of which SCLC is one – there is evidence that PD-L1 inhibitors like Tecentriq and Imfinzi work better than PD-1-targeting drugs like Keytruda.
Yesterday’s disclosure that Keytruda had failed Keynote-604, and the earlier failure of Bristol-Myers Squibb’s Opdivo in Checkmate-451, appear to back such a hypothesis. A less sophisticated view holds that PD-1 drugs overall are better than PD-L1s, though this is massively skewed by the timing of study readouts.
|Failed studies of anti-PD-(L)1 antibodies across various cancer types|
|Merck & Co||Bristol-Myers Squibb||Roche||Astrazeneca||Merck KGaA/Pfizer|
|Melanoma||Keynote-252 (1L)**||Imspire-170 (1L)*|
|Javelin Gastric 300 (3L)
Javelin Gastric 100 (1L)
Checkmate-227 pt 2 (1L)
|Javelin Lung 200 (2L)|
|SCLC||Keynote-604 (1L)||Checkmate-331 (2L)
|Head & neck||Keynote-040 (2L)||Eagle (2L)^^|
|Ovarian||Javelin Ovarian 100 (1L)|
|Hepatocellular||Keynote-240 (2L)||Checkmate-459 (1L)|
|Multiple myeloma||Keynote-183 (3L)
|*Cotellic combo; **epacadostat combo; ^data inconclusive; ^^CTLA-4 combo; 1L=1st line; 2L=2nd line; 3L=3rd line.|
Tecentriq is already approved for first-line SCLC on the basis of the Impower-133 trial, and last year Imfinzi matched its result in the Caspian study (World Lung 2019 – Astra’s achievement in SCLC unlikely to be a game changer, September 9, 2019).
Both these anti-PD-L1 drugs showed 12 to 13 months of median overall survival, yielding around a 30% reduction in risk of death versus chemo alone. In Keynote-604 the absolute survival benefit has not been disclosed, but Merck said the 20% relative reduction in risk of death was insufficient to hit statistical significance.
There was a statistical benefit in progression-free survival, Keynote-604’s co-primary endpoint, but this is unlikely to be enough for approval given the availability of Tecentriq and Imfinzi’s Caspian result. Merck stock was off 2% this morning.
Irrespective of the first-line failure, Keytruda retains its third-line SCLC label with the backing of remission rates seen in the Keynote-028 and 159 studies. And the front-line developments likely have little relevance for Pharmamar’s Zepsyre, which was recently filed for second-line SCLC and licensed to Jazz.
It is hard to gauge how much Keytruda could lose in sales; EvaluatePharma’s sales by indication consensus forecasts shows $748m of 2024 revenue coming from SCLC, but it is not clear how much of this comprises first-line use. Keytruda’s next significant test is the Keynote-355 readout in first-line triple-negative breast cancer.
|Selected first-line SCLC trials|
|Opdivo||Bristol-Myers Squibb||Checkmate-451*||Monotherapy, vs placebo||10.4mth vs 9.6mth (HR=0.84, failed)|
|+ Yervoy, vs placebo||9.2mth vs 9.6mth (HR=0.92, failed)|
|Tecentriq||Roche||Impower-133||On top of chemo, vs chemo||12.3mth vs 10.3mth (HR=0.70, p=0.007)|
|Imfinzi||Astrazeneca||Caspian||On top of chemo, vs chemo||13.0mth vs 10.3mth (HR=0.73, p=0.005)|
|+ tremelimumab + chemo, vs chemo||Due 2020|
|Keytruda||Merck & Co||Keynote-604||On top of chemo, vs chemo||(HR=0.80, failed)|
|*First-line maintenance setting.|