Let's summarize what we know so far. Jill is taking independent random samples from her target population of mice. Then she is measuring the tumor size once in each mouse for both a group taking treatment A, a group taking treatment B, and a group taking both treatment A and treatment B. Now we have a clearer picture of how we should go about calculating sample size. Using a factorial design is a very efficient design if we have biological evidence to suggest that there isn't an interaction between the treatments. Hence the sample size formula will change depending on this knowledge. Both formula's are rather complicated and will not be given here. See Clinical Trials, A Methodologic Perspective, by Steven Piantadosi for a reference.
-- ErinEsp - 13 Mar 2010