EXEL is one IND and one fasttracker away form 3 Billion.
<<A troubling recent trend has been for FDA to pretty much always demand a randomized trial. If there is a plan to try a single arm refractory trial population, typically FDA won't agree in advance (FDA regs demand adequate and controlled trials- getting by with a single trial is an oncology exception that can be given based on robust data) and will also require that patients have rigorous demonstration of refractoriness to all available therapies.>>
Doc, under Dr. Pazdur, the FDA has been remarkably consistent on single trial and uncontrolled trial requirements for several years. I also disagree that they won't agree in advance to an SPA for a single trial. Examples---the Ipilimumab single arm trial in refractory melanoma was conducted under an SPA that specified a minimum ORR necessary for approival. Ofatumumab was approved for refractory CLL based on a relatively small, single arm SPA approved phase 3 trial. Subpart H accelerated approvals are alive and well. Trendwise, I believe we see fewer of them because as more lines of treatment are approved, getting responses from fully refractory patients is becoming more and more problematic. Below is a link to a 3 year summary of all oncology drug approvals:
FDA can be quite dogmatic. Way back with the original Erbitux filing in colorectal cancer they underscored the principal of full refractoriness for subpart H approvals. It is a principal on which they do not want to compromise, even when common sense would appear to dictate otherwise.
Cabo has demonstrated a RECIST ORR of about 10% in CRPC, too low for a subpart H approval. Pain relief is not a good endpoint in an open label trial. Bone scan resolution is not a recognized endpoint, although bone scans are being used by Amgen in a Denosumab trial to document spread of disease to bone. I don't rule out a subpart H single arm trial, but I agree it is less likely than randomized controlled trial.
That again highlites the endpoint issue. Here is a very informative article on the subject of composite endpoints:
The premise is that in a particular indication, clinical trials are lengthy and expensive because the frequency of events to gauge efficacy is low. To compensate, an endpoint that is a sum of composite events, each of which may or may not be a validated endpoint onto itself is used speed the time to trial completion. Among the potential components for use to construct a composite endpoint for the first Cabo trial are:
RECIST responses Pain improvement/med reduction Markers for Osteolytic and Osteoblastic activity Hemoglobin improvement/med reduction Bone scan improvement/progression SRE's (skeletal related events) Deaths
I doubt if they will get all of these and it remains to be seen what form the final statistical analysis would/will take. It could be a "time to event" type analysis or a comparison at a milepost of good events vs bad events. My contention is that if this type trial is done in patients symptomatic for pain and refractory to Zometa/Xgeva, the presumption is that event totals should pile up quickly in Cabo's favor. That makes for a relatively small trial with a short duration.