As far as I know, no. You could probably set up an in vitro system where you have peripheral immune cells and you expose them to the antigen in a slow-release manner (maybe using adjuvants or liposomes or nanoparticles or something that can act as a depot and release the antigen slowly). Although this does not fully mimic all aspects of intramuscular immunization, it may be enough. The better approach would be to do IM immunizations in small animal models, if it is possible.
I am not sure what you mean by an ex vivo phantom method. I am not aware of any ex-vivo method that can mimic injection routes. Just imagine the complexity of modeling skin, muscles and vascular structures in an ex-vivo system to track the route that the inoculum takes. The studies that I have been involved with have used in vitro methods (to study the immune responses to antigen-adjuvant complexes) or in vivo methods (using animal models and stained antigen-adjuvant complexes so you can track its dissemination over time).