From my perspective, APBI is pretty mainstream and I see it in the community quite frequently. One potential problem is that (like many areas in our field) the technology is rapidly outpacing evidence-based medicine. For instance, APBI randomized trials were done primarily with three technologies:
1. Catheter-based balloon systems (Mammosite)
2. kV energy photons (Intrabeam)
3. Electrons w/ lead shielding (mainly used in Europe)
All three of these technologies are easiest to use in the academic setting. For #1, HDR is required so you need an afterloader, Iridium sources, have to meet federal regulations for radioactive isotopes, etc. #2 and #3 are intraoperative approaches and generally only academic places have the volume and expertise to use them, not to mention the capital costs.
In the community, electronic brachytherapy (Zoft) is more commonly used, but this has not really been validated in randomized trials like the above technologies. Of course, one could argue that the physical principles are the same and therefore the efficacy must also be (like was done for IMRT and protons).
ASTRO has published guidelines for who is a good ABPI candidate (i.e. > 50 yrs old, no pure DCIS, no node positive disease, no positive margins, etc).