Sample size estimation is important in planning clinical trials. The purpose of this paper is to describe features and use of SIZE, a comprehensive computer program for calculating sample size, power, and duration of study in clinical trials with time-dependent rates of event, crossover, and loss to follow-up. SIZE covers a wide range of complexities commonly occurring in clinical trials, such as nonproportional hazards, lag in treatment effect, and uncertainties in treatment benefit. The use of SIZE is illustrated by several hypothetical examples as well as applications to real study designs, each featuring a statistical issue.