Posting here as well:
Two patients excluded from plot due to missing data. This is not aligned with your imputation theory. If this were an imputed waterfall, they would be, as protocol specifies, either imputed (other missingness) or BLOCF (no efficacy)
This is important. I emailed IR