Follow-up on Exelixis (EXEL) Failure to Secure Cabozantinib (XL184) in mCRPC Phase III Trial (306) SPA
On Oct. 31, 2011, EXEL announced decision to initiate Cabozantinib in metastatic castration-resistant prostate cancer (mCRPC) Phase III trial (306) after failure to secure Special Protocol Assessment (SPA) from FDA. On Nov. 29, 2011 at Piper Jaffray Health Care Conference Call, EXEL CEO Mike Morrissey gave much more details on the issues raised by FDA during SPA process than previously disclosed. Here is part of the transcript:
Mike Morrissey: Yeah. Well, I think a SPA would have been certainly preferred really to go relative to having a lock in agreement on a trial which we believe is a very important part of the process relating to how we hope to again build this commercially differentiated profile for cabo. That being said, we had a very good dialogue with the FDA. We’ve gone into great detail on the call we’ve had back in October and I think in all the different documentations. We have -- in that dialogue we had two-step process, if you will. We had a pre-end of Phase 2 meetings before ASCO. We’ve got some of the documentation after ASCO. The first set of feedback that we got in the August timeframe was actually very encouraging. A number of enhancements were suggested to put a protocol around inclusion criteria, response criteria, narcotic optimization, the nature of the bone scan response and primary vs exploratory endpoint. All things that we, I think we’re very happy to deal with and include in the protocol relative to the fact that we had very high level of confidence in the bone scan response data that came from the RDT trial, as well as what we were seeing from the non-randomized trial that was recently published now at EORTC. I think the difference came into play on the second round of feedback, where I think the magnitude of the response was being to be small or too small relative to the potential for inadvertent unblinding by the differential AE profile of cabo versus mitoxantrone. We had a 17% delta in the confirmed pain response rate between cabo and mito/prednisone. The trial was designed at 25% confirmed response rate for pain for cabo, 8% was for mitoxantrone/prednisone. And the feedback was that that’s probably too small to be able to get around the idea of inadvertent unblinding. And that was a judgment issue, I think at that point in time we had pretty clear picture of what the agency was looking for in terms of the overall protocol, importantly the magnitude of the effect of 25% versus 8% was reasonable. We think we could probably beat that certainly the -- if you look at the RDT data from EORTC that the number of patients that had about a 30% reduction was about half the patients that we studied, so using that as kind of a good metric for what we might see in 306, one could imagine we had a much higher confirm pain response. We weren’t prepared to commit to a higher level, why raise the bar for a trial that obviously is important one to us, I think the issue is the data will define the utility of that lead out relative to a label and we don’t really response for that, the data will speak for itself. We’re planning to run the trial very, very carefully and collected exhaustively and we’re confident in what we’ve seen so far from a data point of view relative to this impact on pain. Again, this is a different profile from what we’re seeing some other compounds to-date. Our goal ultimately is to have a survival trial i.e. 307 and the pain trial 306 lead out at approximately the same time where we’re modeling in the first half of 2014, if that is to be the case maybe a little bit sooner, we would be able to then again have a strong foundation to go forward with a label that would support both the survival we have, but also a reduction in pain, reduction in narcotics both in response. So ultimately it’s that differentiated commercial opportunity that will drive value and we, again, based upon the data that we have from the last 275 patients its going to be pretty good as how that’s going to go.
Thus, FDA wanted much higher delta in pain response rate between Cabozantinib arm and mitoxantrone/prednisone arm due to potentially inadvertent unblinding of the trial because of significantly different adverse event profile between the two arms. However, EXEL didn't want to risk failure of the trial by raising the delta significantly in 306 trial. This implies the following:
1. FDA potentially is willing to approve Cabozantinib if 306 trial eventually produces much larger differential in pain response rate between the two arms.
2. EXEL wants to get the pain label, but doesn't want to risk failure of 306 trial if higher delta is agreed upon for the SPA. This is consistent with the point I made in previous blog on this subject that EXEL wants to preserve the two options for 306 trial, one is to file NDA based primarily on pain endpoint if the trial produces spectacular result, the other is to file NDA along with 307 trial if 306 trial just meets endpoints to have pain as part of overall label, which is an important differentiation factor for Cabozantinib from other drugs in mCRPC. If EXEL raises the bar for 306 as stand alone trial via SPA that results in failure, Cabozantinib would not get pain label even if 307 trial succeeds.
Posted on Dec. 1, 2011