What is point to continuing testing if it failed at OS and PFS? Why are they wasting $$ which they don't have?