The refrain from sceptics has long been that the FDA won't allow a comparison between an investigational drug and a historical control for approval purposes.
Well clearly, in certain circumstances, and on certain occasions they have done, and will likely increasingly do so. And yes, maybe a somewhat lower bar for medical devices may sometimes be in play here.
The door on this front is slightly ajar rather than wide open, I would suggest. But the FDA pretty much have the discretion to accept whatever evidence they see fit, particularly in the case of a rare disease without an effective treatment. And GBM is (currently), a rare disease. And certainly without an effective treatment.
That doesn't mean that if L has OS60 of 20% and SOC is around 5%, it is a shoe-in, because even historical controls have to be well-matched for all the variables. And you cannot simply compare outcomes for this trial population with outcomes for all GBM patients. Not even those in the US, the UK, and Germany.
But you can identify a closely matched control population, and make statistical adjustment for difference between covariates.
And yes ideally, you should have prospectively identified your historical control or controls.
If the FDA thinks that a valid comparison can be made between the L survival data and maybe the corresponding Optune survival data, there is nothing stopping them from doing so! Particularly if L demonstrates clearly superior milestone survival at all points beyond 2yrs. And if they recognise such a comparison as valid, then they may consider it as supplementary evidence on top of any in-trial comparison.
So the call of; 'the FDA never accepts historical controls' is clearly not the case.
