Biotech Values

[poorgradstudent]

ARQL:

1. For clinical trials, it is vital that a control (placebo) arm and a drug-arm are well balanced. But in MARQUUE, how do we know that the both arms were well balanced regarding to c-Met? Any sizable arms imbalance will endanger the MARQUUE trial integrity.

In the age of targeted therapies, this is actually an interesting proposal. Hypothetically, if the trial is stat sig on an ITT basis but it is shown that there is a significant imbalance of c-Met patients in favour of the treatment arm, you're suggesting that the FDA could waffle on approval.

I understand your principal concern, but unlike more systemic / multi-factorial imbalances (ECOG status, metastases), I don't think we're at a point where the FDA can act in the way that your comment suggests. ECOG status can't be prospectively targeted by a drug the way c-Met can, so it makes sense that trials be balanced for that which we can't necessarily manipulate directly. But when talking about targeted therapies, the raison d'etre is to target. Therefore, I'd be more inclined to think that a trial with everything equal on both arms except the target (c-Met in this case) is closer to the ideal than a trial balanced for c-Met*.

Bottom line, I don't see the FDA turning sour towards a well controlled trial with a stat sig survival advantage even if the c-Met expression (or target X of your choice) is more prevalent on the treatment arm.

2. The same goes for EGFR and KRAS mutant patients control/drug arms balancing.

At the point we're at, I think this goes too far. There are going to be too many disease specific mutations / translocations for companies to prospectively consider. In other words, for proteins that are not the direct targets of the therapy, it should be at the discretion of the company to test or not test for mutations. Or the FDA has to declare it prospectively. Retrospectively mining a trial to ferret out imbalances in certain proteins of your choice will cause a lot of paralysis.


* provided, of course, that this is hypothetically possible. If a certain prognosis, such as worsening ECOG status, segregates with c-Met expression then obviously this would not be a feasible design.