They got the most they possibly could out of that data set IMO. Twenty weeks at least had something that looked like an argument while twenty-four never did.
It won't affect real-world use given current practice, but it will allow them to make 20 week claims in advertising (with all the fine print.)