Replies to post #144031 on Biotech Values

[jq1234]

ONXX: Carfilzomib

I read the critical briefing doc differently. I think almost all of the questions raised by FDA are in the context of single arm trial which couldn't accurately assess the risk/benefit analysis because all of the risks from single arm trial assigned to Carfilzomib.

I also think point #4 exemplifies the mission creep into the accelerated approval process. If the FDA wants randomized trials for accelerated approval, then we should probably just remove the accelerated approval pathway.

ODAC already said even for accelerated approval, they would like to see randomized trial. I don't agree to ask for randomized trial, it means it ends accelerated approval. Carfilzomib 266-patient trial certainly could be run as randomized trial, either with 1:1 or 2:1 vs investigator's choice, with ORR as primary endpoint. It allows better assessment of both efficacy and safety signal. I think sponsors who intend to file accelerated approval should keep this in mind. Otherwise, people can argue all day about whether Carfilzomib is worthy of accelerated approval, and won't agree.

But I don't think your comments are all that convincing considering the pathway that Velcade followed to approval. It seems to me that you would have been equally disappointed with Velcade's initial NDA application, because it was essentially identical to this carfilzomib application. I guess where you see a poorly run single arm phase 2 trial, I see a trial that is following in the footsteps of a very successful example of accelerated approval.

There is no point to compare to Velcade accelerated approval. The treatment regimen and options have changed, so has the regulatory process. Same trial same result doesn't mean approval 10 years later. I am somewhere in between of you two, I think Carfilzomib data is borderline for accelerated approval. I have no problem if it goes either way.

[pcrutch]

FDA disagrees with your take and Onyx's on this. Never said it was therapeutic, but confounding.

Steroid Use
As noted previously, an uncertainty in this study is whether the regular use of dexamethasone has confounded the interpretation of the treatment effect of carfilzomib. The applicant has reported that dexamethasone was necessary in order to reduce infusion toxicities associated with the carfilzomib infusion. Patients were required to receive dexamethasone (4 mg orally or by IV) prior to all carfilzomib doses during the first cycle and prior to all carfilzomib doses during the first dose escalation (27 mg/m2) cycle. After that, dexamethasone use was optional and was restarted in patients with continued toxicities associated with carfilzomib infusion. Dexamethasone is routinely given either alone or with other therapies to treat patients with multiple myeloma. These doses typically are 160 mg/cycle (considered low-dose dexamethasone) to 480 mg/cycle (considered high-dose dexamethasone) range.

Although the maximum dose of dexamethasone a patient would receive per cycle in the carfilzomib efficacy study, Study 3, is lower (24 mg/cycle), a therapeutic effect cannot be ruled out in a single arm trial. In Study 3, the actual treatment effect of carfilzomib is confounded by the concomitant use of dexamethasone in the study and the response rates may be lower in the absence of steroids.

My comment on Velcade was only in response to some claims by Onyx that there is an unment need because patients are intolerant to Velcade. SubQ Velcade has drastic reductions in PN and other AEs, so Onyx claim doesnt seem all that strong. I wasn't making any claim about the cardiac related AEs, which Onyx seems to have plenty of as well.

Seeing such a high rate of protocol violations makes me wonder whether Onyx can properly run the ASPIRE and FOCUS Phase 3 studies.

The passage about the accelerated approval pathway was meant to illustrate that Onyx's trial does not seem to meet it for a couple of reasons. It's unclear if the ORR in their trial really qualifies as "providing meaningful therapeutic benefit to patients over existing treatments". I think the confounding of the data with Dex makes it a tough judgement call.

Recent article on treating r/r mm
http://asheducationbook.hematologylibrary.org/content/2010/1/303.full