Minimum sample size calculations for external validation of a clinical prediction model with a time‐to‐event outcome