As I understand it, the partitioning coefficient will have to be determined experimentally. So under controlled conditions, determine how much is extracted out of a defined (i.e. prepared in the lab) sample of contaminated water, or a certain volume into a certain volume of toluene. Once you know that, it should scale, percentage-wise, with a real world sample of much great volume, proportionally.
That's a lot of caveats, but when you get to planning it out, you should see how it all fits.
N.B. your treated effluent still has organic in it. You've compensated for the toluene that has now dissolved in your treated effluent after extraction, right?