Fair enough - I shouldn't say lose money at the physician level. I'll say takes up more physician, dosimetry, physics, AND therapist time than a standard palliation plan. The downstream effects of which will be felt by the whole department. Imagine if every bone met patient you had now was done with IMRT. We do daily CBCT for all IMRT.
How much longer would your machines run on a daily basis? How much more expense would the department have in therapist expenses? How much more physicist support would be required for the increased volume of QA required? Time for dosimetrists to contour every OAR on every spine met and do inverse optimization?
Again, if it's something that is worth it then OK. IMO I'm not going to routinely do IMRT for bone mets regardless of the APM situation. Same with SBRT outside of oligomet/radioresistant histology, especially outside of the spine.
* Efficient & quick contouring should be the modern rad onc's goal. If you can't get a spine SBRT all beautifully contoured up in <30 min, that's bad. And you can do the whole plan (including inverse opt) start to finish in ~30 min or there is also a problem. QA time commitment, at most, 15 min or so I guess.
* I do IMRT e.g. for every C-spine palliation case e.g. I put the gantry at about 90 & 270, and do a couch kick away from gantry of about 25-30 degrees. This way you can come in with "non-coplanar laterals" to avoid the shoulders but be very much more off the esophagus. I make things pretty tight in terms of anterior (into esoph) spillover, keep the CTV->PTV expansion tight, etc. This is a 2-field plan which can essentially only be done with IMRT to keep the target dose homogenous and the plan Dmax ~110% or less. All that said, this is a very easy plan to plan and implement and takes no longer than the "average" spine palliation by any measure.
* Herman Suit used to say, to justify proton therapy, "There is no clinical rationale for even one picogray of dose outside the target volume." This is a good ethos IMHO. And if you want to apply it for bone mets, and do IMRT vs AP/PA e.g., why not? I, personally, would want to be treated that way.
* In APM, I predict routine IMRT QA will go away. This, like "one above and one below," is a relic of the past. We do electronic wedges e.g., which are really just IMRT (that is not billed as IMRT). We never QA that. We never see sliding wedge dose disasters clincally due to lack of QAing the plan to make sure the sliding MLCs or jaws are applying the "wedge." I know some academic centers that do not do routine on-machine IMRT QA e.g. although they would never admit it. There is no clinical advantage to QAing EVERY IMRT case. Nor is there any advantage in "easy" SBRT or SRS IMHO. QAing is now technically a billing requirement for IMRT.
* You can do a very conformal treatment with two VMAT arcs which allows a tx to be easily still fit within a 15 min time slot.
* The 3D-> IMRT transition era did not see an increase in therapist expenses. No reason a IMRT->SBRT transition would either, especially since it implies much less OVERALL course-of-treatment treatment time per patient.
So, at least in my mind, in a very near future or today, SBRT for most things which used not to be SBRT will not see increased burdens on a dept. It will be less burdens, and equal or less expenses. It may very well be less reimbursement though. A 10-fraction 3D and IGRT can get more reimbursement than a single fraction SBRT. And a 10 fraction IMRT+IGRT sure will.