I might be the only person on earth who got this one wrong, but regardless:

"ITT analysis includes every subject who is randomized according to randomized treatment assignment. It ignores noncompliance, protocol deviations, withdrawal, and anything that happens after randomization."[1]