Looking at this again, my question now is about induced flow possibilities.
Could you do a sidepod that looks like the first 300mm stub of an airplane wing, have a spanwise radiator seperating the upper and lower chambers and the top having forward facing gills in the rearmost 25% of the legal surface, and the bottom having the same rear facing gills on the forward edge of the bottom side of the stub. The idea being that by blowing across the leading edge it creates a suction that draws in from the top/rear gills, and through the radiator.
If that works, I think you could then drop a vertical panel from the outer edge, to the floor, and shape that as a venturi as well. Thus, the bottom gills would be in the throat of the sidepod, drawing through air in at a boundary layer separation area, heating it up, and accelerating it out the rear "cannon". I think JaF has mentioned Merideth sidepods before... Something like this might be neat.