I want to return to this again because I think it is important. Failure to reach agreement on a 306 SPA was a rejection of a specific protocol, not a wholesale rejection of a pain endpoint.
First, let's talk about the SPA process. In it the FDA provides its current thinking on what it will take to get a drug approved by means of a specific trial protocol. By agreeing to a SPA they are saying that if you dutifully conduct this trial in compliance with the protocol and it acheives its primary endpoint, we would approve the drug today and in the absence of any developments that would invalidate the present landscape, we would expect to approve the drug at trial completion. Obtaining a SPA is not mandatory and in fact most phase 3 trials are not conducted under the auspices of a SPA. The advantage for the sponsor is obvious, it essentially removes the risk of a future FDA criticising the trial design and denying approval on that basis. With this in mind, FDA steers a very conservative approach in its SPA negotiations as it does not want to put limits on itself that it may regret later. By agreeing to a SPA they are also acknowledging that the efficacy required for approval is attaining the minimum efficacy needed to attain the primary endpoint.
Pain endpoints in oncology are controversial because the means of demonstration is highly subjective. Statisitically measuring subjective outcomes is prone to influences that can skew the result. Compared to objectively determining survival duration, PFS or response rate, it is intuitive that patient reported outcomes can be skewed by influences other than drug efficacy. Placebo effect due to side effect unmasking, differences in language, inconsistencies in data collection and treatment of missing data have all occurred in prior pain measurement data and contributed to disapproval or failure to grant a pain label. Despite this, FDA still encourages sponsors to measure pain and still lists pain improvement as an endpoint meriting full approval. Yet I still think they cringe whenever they see a pivotal trial with a primary endpoint with a pain component. So when approached with a SPA proposal for a trial with a pain outcome, my thinking is that it is not a surprise that it would be a protracted and difficult negotiation. Yet if confronted with an NDA based on a pain outcome, it is a different set of circumstances. All of the "what ifs" are gone and replaced with hard data.
Now lets take the general and apply it to the specifics of the 306 (Comet 2) proposal. Negotiating with FDA is time consuming. EXEL thought it was close to an agreement and then FDA reportedly backtracked in October 2011 and wanted 2 trials instead of one, expressed concern over the potential of SAE's causing inadvertant unmasking, and proposed a survival endpoint as an alternative. Remember this is within the context of Comet demonstrating minimal efficacy, approximately doubling the Mito pain response rate of 8% to 16%. Rather than reenter negotiations and submit a counterproposal, EXEL chose to proceed without a SPA. If history is a guide, the 306 actual outcome will be a treatment arm pain response rate of 50% vs control 8%. That level of efficacy will be the yardstick FDA, ODAC, and the EMEA will consider, not the theoretical minimum efficacy embodied in the SPA. That data will be supported by reduced narcotic use, secondary markers of bone met activity, CTC count reductions and bone scan improvements.