Esmo 2022 – Lumakras’s lung cancer shocker
Having promised much, Esmo underdelivers with a late-breaker that appears to hand some advantage back to Mirati.
After a huge build-up, Amgen’s Esmo appearance with Lumakras has taken a rather disappointing turn. The company’s Codebreak-200 trial, intended to confirm Lumakras’s US lung cancer approval, has yielded no overall survival benefit, as well as giving a reminder about liver toxicity, a special late-breaking Esmo abstract has just revealed.
On the face of it this is positive for Mirati, whose stock had traded up when Amgen’s opaque topline Codebreak-200 disclosure hinted that all was not well. However, OS could be numerically positive, the Codebreak-200 dataset might yet be approvable, and Mirati still faces a tight deadline to get its competitor, adagrasib, across the US finish line.
The battle is coming down to the wire. Adagrasib faces a December 14 Pdufa date, but this accelerated approval might be under threat should the FDA earlier grant Lumakras a full green light. So has Amgen done enough in Codebreak-200 to secure full approval, and if it has how quickly will the US regulator act?
Today’s late-breaker, which Esmo had moved heaven and earth to include in the conference programme, helps with the first question. PFS, Codebreak-200’s primary endpoint, was known to be positive, and the benefit here has been revealed as a 34% reduction in risk of progression versus docetaxel, with high statistical significance of p=0.002.
But this is where the good news ends. Overall survival, a key secondary endpoint, was “not significantly different between treatment arms”, the late-breaker states, though it also cautions that Codebreak-200 had not been powered for OS. It is unclear what the statistical bar was for OS, and the abstract is silent on whether any numerical benefit was seen.
Further worry comes in the adverse events summary. 10% of patients experienced liver enzyme elevations, and the incidence of ALT and AST elevations at grade 3 or above was 7.7% and 5.3% respectively. Liver enzyme elevations became a major issue when Amgen recently reported data from Lumarkas’s Keytruda combo trial at World Lung.
Still, this is by no means a new concern for Lumakras, whose label already includes dose-modification instructions in the event of grade 3 or 4 liver enzyme elevations, or grade 2 elevations with symptoms; there is no suggestion that Hy’s law has been triggered. And the problem concerns adagrasib too, though (on a cross-trial basis) not to the same extent.
|Cross-trial liver enzyme elevation comparisons for Kras inhibitor monotherapy in NSCLC|
|Project||Study||Any grade||Gr3+||Any grade||Gr3+|
|ALT=alanine aminotransferase; AST=aspartate aminotransferase. Source: NEJM, Esmo & Asco.|
A separate question is how the market will perceive these findings, and whether the sellside had already done enough to rein in expectations.
Notably, initial hopes were that OS had been hit, albeit perhaps not convincingly, but once Amgen toplined the PFS result analysts started mooting the possibility of no OS benefit, citing study underpowering and extensive patient crossover. There was even some fear that Amgen would not report the OS number at all.
It might also come as a surprise that Codebreak-200, an open-label, second-line trial versus chemo, had PFS as primary endpoint. Typically the gold standard of OS is needed to confirm a remission rate benefit backing an accelerated approval, and using PFS to do this seems strange since PFS is itself a surrogate endpoint.
Still, the FDA will presumably have signed off on the acceptability of Codebreak-200’s primary endpoint as confirmation. Some analysts had suggested that a median two-month PFS benefit would be good, but the abstract only provides a one-year landmark analysis – 28% versus 10% – in addition to the hazard ratio.
Investors have to wait until Monday’s presidential session to see the curves for PFS and hopefully for OS too, to gauge the benefit fully and to look for positive signs as to the latter. The wait for the FDA’s verdict will be longer.