Pfizer and Merck go for an even longer throw of the Javelin

Pfizer and Merck KGaA’s tweaks to the Javelin Lung 100 study of their PD-L1 antibody avelumab, which will delay its readout by almost two years, look like a painful but necessary adjustment to commercial realities in this competitive cancer indication. The changes will more than double the size of the study, in first-line non-small cell lung cancer, and push data out to mid-2019.

The delay is, however, unlikely to concern either of the two PD-1 incumbents in NSCLC, Merck & Co and Bristol-Myers Squibb, or the two companies with near term readouts, Astrazeneca and Roche. Indeed, the changes to Javelin Lung 100 seem to be designed to preserve some sort of foothold for avelumab in this indication, which is seen as strategically vital, even if the product’s commercial value lies in other solid tumour types.

An update on Clinicaltrials.gov shows Pfizer and Merck to have boosted recruitment in the trial, which evaluates avelumab as monotherapy versus platinum doublet chemotherapy, from 420 to 1,095 patients. The study now has overall as well as progression-free survival as primary endpoints, which will be examined with three different cutoffs according to PD-L1 expression: high, intermediate and any. 

PD-L1 cutoffs

The previous protocol did not plan to evaluate the population by different levels of PD-L1 expression. The increase in recruitment presumably allows for the study to remain powered with these different cutoffs examined.

The companies have not disclosed the new cutoff for high PD-L1 expression, but this will presumably be informed by the experience of competitors. Merck & Co’s Keytruda is approved for first-line use as monotherapy at >50% PD-L1 expression, representing around 25% of all patients, while Bristol’s failed Checkmate-026 study of Opdivo used a 5% cutoff, which is now seen as too low.

The failure of Checkmate-026 of course has left open the possibility for another PD-1/PD-L1 agent to gain a monotherapy indication in first-line NSCLC at some intermediate level of PD-L1 expression; however, that opportunity for the moment belongs to Astra. It will obtain a PFS readout for durvalumab alone as well as in combination with the CTLA4 antibody tremelimumab from the Mystic study in the middle of this year. The study’s statistical analysis plan has an undisclosed cutoff for PD-L1 expression, widely thought to be 25%.

The recent change in Mystic's design, promoting PFS in PD-L1 expressors to co-primary status, was initially interpreted as suggesting a degree of caution over the durvalumab/tremelimumab combo. However, Astra has revealed that the change was motivated by the opportunity to capture a short term commercial advantage as a monotherapy.

Nevertheless, much rides on the outcome of Mystic, especially since it will provide the first opportunity to access how tremelimumab’s safety and efficacy profile compares against Bristol's rival CTLA4 antibody, Yervoy.

Second line

Pfizer and Merck KGaA are separately running a registration trial for avelumab in second-line NSCLC, Javelin Lung 200, which analysts believe could support filing this year. However, with the expectation that IO-IO and IO-chemo combinations will come to dominate first-line NSCLC, it seems unlikely that there is much of a long-term future in second-line disease.

That move towards combinations is already under way with Merck & Co having filed for Keytruda plus chemo based on the remission rates seen in a phase II study, Keynote-021. This filing for accelerated approval has a PDUFA date of May 10. Furthermore, the move prompted most competitors to revise the designs of their registration studies.

Roche should shortly have a readout from its Impower-150 study of Tecentriq plus chemo, and this will likely soon be followed by Astra’s Mystic. Analysts generally reckon that both studies will read out positively.

Meanwhile, some analysts believe Bristol’s Checkmate-227 study, which tests Opdivo with Yervoy, could read out as early as August, although it remains slated to do so in early 2018. Again, the consensus is that this will hit statistical significance on PFS in PD-L1-positive patients.

A potential $10bn annual market is projected for checkpoint inhibitors in NSCLC, likely split between the front-runners in anti-PD-1/PD-L1 therapy. How this split occurs will depend on the outcome of a handful of trials that read out this year.

Currently, most analysts believe that IO-IO combinations will be used in the higher-expressing PD-L1 positives, and IO-chemo combinations in the negatives and lower-expressors, although this remains open to debate. However, few analysts see much room for monotherapy in NSCLC in the long term.