Dear all, I am no statstical pro, and before anybody comes up with the obvious answer, no, I have no access to skilled statisticians to get the answers I want and I have to find them by myself. I really need to deepen my knowledge on clinical trial design (especially in cancer research). Specifically I need to calculate how many patients I would need if I want to assess the significance of a variable to predict survival in two groups of patients who receive different treatments, also keeping into account other well established predictive variables (so in this case I don't need ad interim analyses or stopping rules). Are there any free softwares (like some R package) or even commercial ones that would do the trick? I have tried IBM SamplePower, but it doesn't provide the answer for multivariate predictors. Neither does nQuery, as far as I can see. Finally, any suggestions on good and up-to-date texts on this topic? Thanks