EXEL is one IND and one fasttracker away form 3 Billion.
I think our disagreement about the FDA and the SPA process is more about how the FDA "agrees" to things. It's always a fascinating exercise to look at the Briefing documents prior to an ODAC meeting and look at a Sponsor's regulatory history then compare it to FDA's record. Sponsor's interactions with FDA basically consist of questions. As in a deposition or a legal case, what questions and how they are asked can dictate the potential response. In an SPA, the associated question might be : We believe this trial if positive will support a filing for drug X in the following indication...Does FDA agree? FDA might agree by saying that if this trial is perfectly conducted and the data are robust that it could support approval. The example of RR in melanoma is an interesting one because Roche/Plexxicon are facing it with their raf-inhibitor data in melanoma. The ORR is high but durability may be an issue. I'll also note that despite an SPA, ipilumumab required a randomized survival trial to likely get FDA approval. The single arm trial by itself would not do it.
The issue of compsite endpoints is complex. On the one hand, the idea is to capture more data in a shorter period of time. On the other hand, the statistical significance required gets divided among the various components of the composite score. So you might capture improvement in RBC counts, but if that only happens occasionally, it may dilute out a more positive endpoint. Another big unknown here is how Cabo will perform in a strictly defined post-taxane, post-abiraterone population. It may be that the positivity of these components of a composite endpoint may fall out if you have a population that truly has failed docetaxel. Of course, one of the reasonas for coming up with a composite, is to capture the bone scan effects since those appear to be a very strong signal and would help a less strong signal turn positive. How likely this is to work and whether this will help speed up an approval in HRPC depends upon more data. This data will need to be in the proposed population (I would think at least post-docetaxel) that cabo positively affects all the surrogate measures associated with disease and that overall, this correlates with clinical benefit.
Time will tell on this and I'd be happy for the data to be so strong that almost any endpoint will deomnstrate activity, but it's rarely that straight forward.