I don't think we have any exisiting examples that go back and forth between CHARGE and FDTD. If you have something like a photodiode, then you do the simulation first in FDTD to simulate the absorption of light, then you import this to CHARGE to calculate the photocurrent under bias. If you have a modulator, then you calculate the charge injection first in CHARGE. Then the index perturbation due to either of these is calculated to determine other modal properties like effective index, losses etc.
So, I think this can also be done in Lumerical, this is more like a two-way simulation. Here are the links to the examples on: