How can the intention to treat analysis avoid loss to follow up bias?

As the subjects dropped out form the group, the outcome of the subject would not be able to obtained. How can this lead to the prevention of loss to follow up bias?|||Huh?