I had to re-derive it for a perfect cone to be sure. You are correct; at a distance where the size of the cross section of the beam is significantly larger than the source, deviation from inverse square becomes negligible.