@Phaedrus, I agree, some of this is off-topic for the thread... but we're here. Starting with your, "That's what we're dealing with.." comment and working backwards on your post... I presume you're referring to the selumetinib ORR of 66% (or 44% according to an independent reviewer) and wondering why not rem-L? Only 50 patients treated in an open-label study. I'm not here to defend the FDA, in fact I've stated repeatedly (
@DV) that I grant there may be... bureaucratic inefficiencies... "confusion"... unfairness... undue big pharma influence and the like. It's not an excuse for 2 CRL's in this setting. And there is some plausible explanation for the approval of selumetinib, just like we discussed with Lantidra. It all comes down to the nature of the illness and response to the proposed treatment. Some key words from the selumetinib announcement: "Allpatients had a partial response, and 82% of responders had sustainedresponses lasting at least 12 months."
That's in an illness where there is predictable tumor progression and essentially a 100% adverse outcome with no good alternative treatments. 100% partial response vs 100% indolent progression. Same with Lantidra... there's a predictable 100% chance of death in brittle Type 1 diabetics who don't get insulin and a 70% response to the new treatment lasting over 1 year. Approval only for patients who can't take insulin. Limited alternative treatment options available.
The situation ismore complicated with GvHD. Dr Kurtzberg's in vitro T-cellactivation poster that you linked gives a clue: “.... patientswith the most severe form of the disease and at highest risk fordeath: Minnesota high risk (D180 OS 89% vs 50%, p=0.01), MAP >0.29(D180 OS 100% vs. 17%, p=0.003) or Grade D disease (D180 OS 91% vs.50%, p=0.03).” In short, an alternate way to look at it: clinical evaluationby the doctor is able to predict mortality with only 50% accuracy. Meaning some patients do respond to other treatments. That's very different than the case with the two meds above. What is needed is a test that gives early categorization of responders/survivors vs non-responders/non-survivors. And the MAP score basedon biomarkers does better than the doctor doing an exam and labtesting. Still not 100% accurate, but much better that clinical grading. Yet some patients getbetter...meaning control patients get better... which is great but it makes evaluating meds to treat theillness statistically more difficult, because that's not what would happen with selumetinib and Lantidra controls. They ALL get worse. The illness itself leads tothe requirement for a higher level of testing... and testing in morepatients... to prove efficacy. Reviewing slide 13 from March, yes 50+ patients were in study 001, but apparently only 11 or 12 were in the MAP>0.29 category. That's not enough to draw statistical conclusions. And I have to say I'm surprised they didn't submit data on a larger cohort.
I think your comments on testing in adults agree with mine - that's the route followed for most medications if adult illness predominates. You're playing with semantics on the issue of the FDA "recommending" vs "requiring" randomization and control. In any event I consider it beside the point. The key issue is what indication should have been advanced by the company first... childhood SR-aGvHD with no RCT data or one of the others with RCT data. I've stated my views on the matter and will not repeat here.
Youcan argue for RWE and open label trials if you will. I think sufficient data should have been submitted with BLA#2 to make it a moot point. I can see the possibility for orphan indications in a subcategory with predictable mortality that approaches 100%. If we have that via serial MAP scores, great. Let's see what comes out of the Type A meeting.
Regarding confounders, many of the meds on the list I posted are continued indefinitely and in combination. There's no washout period. My understanding is that the course of treatment goes something like this: All patients start with an anti-rejection prophylaxis regimen (which is evolving - some patients are now getting ECP pre-graft). They then receive the allogenic transplant. Some do well, but GvHD develops in a significant percentage. The anti-rejection meds are continued, methylprednisolone is started, sometimes other meds are added... all as "primary" GvHD therapy. About half of patients respond (meaning it's unlikely steroids will be replaced or "banned" anytime soon, though the theory could be tested with a large enough cohort,
@DV). Steroids are tapered in responders. In non-responders "steroid resistance" is diagnosed at some point, a clinical severity classification is assigned and MAP biomarkers are measured. Secondary therapy is then begun using rux and/or other meds from the list I posted above. Standard for clinical trials is then to measure ORR at 28 days and more recently to repeat biomarkers. No response to secondary therapy and/or rising MAP score at 28 days is associated with highest mortality and might be an entry point for a tertiary Rx. So,
@Phaedrus, do you see rem-L pitted against rux as secondary therapy? I believe the original intent of Study 001 was to use rem-L as a secondary Rx. Is SI now moving it to a tertiary "medicine of last resort" comparable to what we saw with the ARDS trial? Smaller addressable market but perhaps an easier path to AA. I would think that decision would have more impact on trial design than any other. Also any thoughts on why additional data weren't submitted for a larger cohort of patients with MAP scores above 0.29? imo, Left-e
(20min delay)