Best:
MRI/dwMRI
Useful:
CTA/CTP
Standard of care:
NCCT
It depends on what level of clinical skill you have and what treatment modalities you have available.
If you are in podunk community memorial hospital and all you have to offer for stroke is IV lysis, the only question you need answered is bleed/no-bleed and swelling/no-swelling. That is where the NCCT will give you all the answers you need.
If you are at a place with interventional stroke treatment, you want to have an answer (quick) whether this is one of the 5% of patients that might potentially benefit from agressive intervention. This is where CTA/CTP really shines, in a 5 minute scan you can answer:
- major vessel involvement
- approximate extent of ischemic territory
All this on equipment readily available and with technical staff that is routinely in house (CT techs). The only downsides are that CTP can be a bit finicky and you blow 100+ ccs of contrast into a patient that potentially will see a couple of 100cc during his intervention.
A dw MRI otoh will give you a 3D delineation of the ischemic territory, the most sensitive tool to look for hemorrhagic conversion and when combined with a 3D-TOF MRA will give you the answer for the major vessel involvement. All that without potentially nephrotoxic contrast. The downside is mainly logistics. Your MR techs are rarely in-house, the limited MRI/MRA will take longer than a CTA/CTP, the patient needs to be moved on a MRI compatible stretcher, EKG electrodes have to be switched for MRI compatible stock etc. Also, often stroke patients have pacemakers limiting your ability to do MRI (and if they are not awake or able to talk to you, you can't screen them for MR contraindications like pacers/pumps/clips).
So, if I had unlimited resources to set up a stroke protocol, I would have:
- an in-house MRI tech
- all patients with stroke symptoms in triage would be put in the 'stroke room' where only MR compatible equipment is available
- I would x-ray head/chest/1view l-spine to r/o hard contraindications to MRI on all comers.
- from the 'stroke room' where a trained nurse-specialist would do the NIH and whatever other stroke scales, the patient would be shipped right into the scanner next door and get his/her dwMRI/MRA.
Now, 'cost' in medicine is a very fluid concept. If you want to know the charges for the various modalities, check on the AMA website:
https://catalog.ama-assn.org/Catalog/cpt/cpt_search.jsp?checkXwho=done
(strictly speaking those charges which are given in medicare funny-money don't really apply to inpatient medicine as those are billed under a different system. but it gives you an idea as to the relative values)
After you already own a MRI scanner (which will depreciate whether you use it or not and boil off helium whether you scan or not), the proportional variable cost of doing another MRI is purely limited to the electricity it takes to run the scanner and the personnel cost of calling in a very qualified staff member.
In CT otoh, your proportional variable cost includes a tech that is already there (and as such costs you nothing) but you have to pay for every second the beam is on (tube-time) and you have expenses for IV contrast.
So the cost equation is different whether you are in a private hospital setting, a university, a different country like the UK etc. Any time someone publishes a paper about relative cost of different modalities you have to ask hard questions as to what agenda that person is trying to pursue. Especially if people base their fuzzy math on charges. Reimbursement for medical services is the most f-ed up system you can think of. An evil genious couldn't come up with a more ******ed setup. (imagine you bring your car to the shop and they tell you 'well, if you pay cash, that new oxygen sensor is going to be $600, if your boss pays for the repair its going to be $200 and if you give us a phantasy name and just drive off after we do the work you'll get it for free and we'll put it on the bill for the next guy').