According to Palmer, utilization is a function of two variables - 1) specific gravicty of the actual medium that the hops are in, and 2) time the hops spend in that mix.
Utilization = f(G) x f(T)
f(G) = 1.65 x 0.000125^(Gb - 1)
f(T) = [1 - e^(-0.04 x T)] / 4.15
So yes, it does seems that utilization would be maximized by adding NO extract until the very end of the boil. Of course, you have to account for this better utilization by adjusting the amount of hops and/or the time that they are added.
This would also suggest that it would be possible to do a boil (or partial boil) in less than an hour. There must be some other drawbacks. There could also be some benefits from minimizing the amount of time the extract boils.