They will come up with some endpoint for their randomized trial based on bone scan reduction/pain reduction with secondary endpoint OS. That primary endpoint is a clinical endpoint (unlike tumor progression) and so is enough for accelerated approval.
Are there any examples of a cancer SPA that is neither a standard (e.g. RECIST criteria) nor a clinical endpoint in and of itself. E.g.
1)OGX-011 has one for pain (but that is an unambiguously clinically significant endpoint).
2)GTCB claimed to have one for Satraplatin but was 'confused'.
3)DCTH has one for hepatic PFS - this is probably the closest in that it is modification to general PFS. But still it is a long way from a composite endpoint. And it doesn't seem to be going too well in any case -g-.