Portable MRI Scans Patients for Signs of Stroke | Hacker News Reader

Portable MRI Scans Patients for Signs of Stroke

spectrum.ieee.org

59 pointsrbanffy26 comments

I’m a neuroradiologist. The article is confused about the stroke workup—the most critical thing is to detect acute hemorrhage, which is much easier, faster, accurate, and cheaper with a CT than an MRI. We already have mobile CT units in deployment.

That’s not to say there isn’t value in mobile MR, but it’s not as the article states.

There are even a handful of ambulances equipped with CT scanners so that stroke patients can be diagnosed and treated in the ambulance. (By “handful” I mean that, when I toured one, I think there were a grand total of two in California.)

I imagine it's hard to think about investing in one when LUCAS machines are out there, too.

I'm curious if you think the premise that "medical imaging is essential". There are a few start-ups here in Sydney that are developing non-imaging technologies for measuring stroke via blood flow, EEG, etc.

These devices are tiny and inexpensive (basically EEG headbands). Nuroflux is one (https://nuroflux.com/), I forget the name of the other.

Came here to say this

Contrast CT is preferable over MRI in hyper-acute stroke because of much shorter acquisition times. MRI might have higher resolution, but there is little value to that when time is brain (i.e starting fibrinolytics).

Given these units are portable, I’m curious how acquisition times and/or resolution compares to gold-standard CT.

The time-savings seem to come from having to avoid transport to diagnostic imaging, but many EDs are already equipped with CT scanners for trauma cases.

Well it depends, France adopted widespread use of diffusion weighted MRI as first line modality for stroke because it's much more sensitive than cect, but yeah, most institutions do CT scan as a first line for several reasons including one you've provided.

Can you provide a citation for the France assertion? I think it’s wildly unlikely a protocol for acute stroke would favor mri over ct but could be wrong. It would take 20 minutes to transfer a pt to mri in a lot of stroke centers in the USA, as opposed to CT’s that are generally across the hall, where imaging should be read within 30 minutes of door time I believe.

Also I’m not sure what you increased “sensitivity” would get you. Acute stroke is a clinical diagnosis, the imaging determines the type of stroke and treatment.

> Can you provide a citation for the France assertion? I think it’s wildly unlikely a protocol for acute stroke would favor mri over ct but could be wrong.

https://www.sciencedirect.com/science/article/abs/pii/S00353... (there's free pdf available when you search for it): "The first-line brain imaging at WH was MRI in 69 SU (56.1%), CT in 6 (4.9%), and either MRI or CT depending on delay and severity in 48 (39.0%). The first-line brain imaging at NWH was MRI in 54 SU (43.9%), CT in 16 (13.0%) and either MRI or CT in 53 (43.1%). In practice, the proportion of patients who really underwent first-line MRI was higher than 90% in 46 SU (37.4%) at WH and in 36 SU (29.3%) at NWH"

> Also I’m not sure what you increased “sensitivity” would get you. Acute stroke is a clinical diagnosis, the imaging determines the type of stroke and treatment.

In clean and easy cases sure, not all cases are like that though and MRI is very useful then; by sensitivity I mean sensitivity - https://pmc.ncbi.nlm.nih.gov/articles/PMC1859855/

Reading that couldn't be more clear, CT is the primary modality for stroke, worldwide.

  > by sensitivity I mean sensitivity

You're a little confused. You're using "sensitivity" to mean sensitivity of detecting ischemic stroke. MRI is the obvious follow-up. When available, worldwide. But it doesn't guide emergency treatment.

> Reading that couldn't be more clear, CT is the primary modality for stroke, worldwide.

Well yes, it's primary modality for stroke worldwide and it's leading modality in France, just like I've said before.

> You're a little confused. You're using "sensitivity" to mean sensitivity of detecting ischemic stroke. MRI is the obvious follow-up. When available, worldwide. But it doesn't guide emergency treatment.

I would appreciate if you stopped using condescending tone. It does not guide emergency treatment decisions because in most cases it is not performed in emergency settings. When it is performed in this setting it is guiding treatment and MRI is included in stroke guidelines for cases where clinical diagnosis is not clear (and these cases are not that rare). Why is it not widely adopted? Mostly logistic reasons (which can be overcome - like they were in France) and because TOF-MRA is generally worse than CTA. It has others positives apart from higher sensitivity though, e.g. you can use FLAIR/DWI mismatch in wake-up strokes which are VERY common (obviously perfusion serves generally same purpose).

>when time is brain (i.e starting fibrinolytics)

Why not take fibrinolytics before the scan, just to be on the safe side? Ambulances could be equipped with fibrinolytics, for instance.

Is it because fibrinolytics could actually be harmful, depending on the stroke type?

The fibrinolytics have roughly 6% risk of causing a cerebral hemorrhage even if the diagnosis is correct and there is no bleeding to start with. Unfortunately their benefit is... usually less dramatic, according to the literature (with the initial NINDS study showing no benefit at all at 24 hours, with a change in the primary outcome to highlight the small benefit found at 30d).

A CNS hemorrhage can present similarly, but giving these meds would be a death sentence in this case.

I'm not sure I follow, does this imply fibrinolytics are dangerous drugs by default for the elderly population that tends to get strokes? Are they just dangerous in general?

>with a change in the primary outcome to highlight the small benefit found at 30d

So does that mean "time is brain" is a bit exaggerated then, if the benefits are small?

Fibrinolytics are indeed risky agents to use as they are non-selective in where clots are lysed. To give you a perspective, an injury from a fall from the initial stroke event is likely to re-bleed once fibrinolytics are administered. Patients sometimes bleed around the IV site when the med is running.

These meds need be given within 3-6 hours of stroke onset, depending on who you ask and what heroics you want to accomplish. After the six hour mark, you’re facing the same risks of disability from an iatrogenic bleed versus little gain from therapy because hypoxic tissue is dead at that point.

However, before that window closes, there is significant benefit in doing something versus nothing in stroke. It’s great when the use of the agent results in no deficits. The same can be said when the results are only limited deficits (i.e. use of a cane) versus a wheelchair or admission to long-term care.

> I'm not sure I follow, does this imply fibrinolytics are dangerous drugs by default for the elderly population that tends to get strokes? Are they just dangerous in general?

Dangerous in general, especially for patients with acute stroke. Thankfully the risk of an intracranial bleed is much lower for patients without a stroke (~1% IIRC), so if you get the diagnosis wrong (as there are many mimics) at least the risk isn't as high.

> So does that mean "time is brain" is a bit exaggerated then, if the benefits are small?

¯\_(ツ)_/¯ -- my impression is that the "time is brain" is emphasizing that "earlier is better," which is absolutely the case. Unfortunately, better is not always "good."

Thankfully, for massive strokes, we now have much more effective options (which are generally performed along with tPA). Unfortunately these options require highly trained subspecialists that my be practically unavailable to rural hospitals.

If the patient is having a hemorrhagic stroke, fibrinolytics could cause catastrophic bleeding. As soon as you’ve excluded hemorrhagic stroke (and other hard contraindications), you give fibrinolytics ASAP if you suspect an ischemic stroke.

The same symptoms can be caused by either a blood clot OR by hemorrhage in the brain. If you give a patient with a hemorrhage fibrinolytics, you killed him. That’s why you need the CT first: to rule out bleeding.

CT is not much faster than a quick T1 and T2 from a portable MRI these days. By the time you transport to the CT suite, set them up, run the scout, etc., you could have acquired the MR bedside.

Edit: To add some clarity, portable MR systems are low field, so that means the scans will be lower SNR and resolution, which are much faster to acquire. I think it would also rule out diffusion scans, since those require high gradients. MR acquisitions are also much faster nowadays due to compressed sensing and deep learning reconstructions.

There's no portable bedside MR available though.

Even if you have an MR machine with a very low strength primary fields you'll still need to consider safety issues related to the gradients and coils.

Portable low field bedside MRI has been on the market in the US for a few years. See Hyperfine Swoop.

Contrast CT? Do you mean like an angio? Or is this a typo? (Noncontrast is virtually always first for possible stroke presentations.)

Apologies, I was typing late at night. Thank you for catching this :)

TFA observes how the current resolution is fairly low - compare this to was available for live view on the frame buffers of the first EMI scanners, barely 80x80. Even that was revolutionary, and I think to a layperson illustrates how little resolution you can get away with if you have a third dimension.

https://www.jira-net.or.jp/vm/pdf/ct_catalog/EMI-SCANNER.pdf

Portable MRIs of various types have existed for a while. They even make 1.5T units on trucks. This one is itself preceded by the Hyperfine Swoop:

https://hyperfine.io/

And yes a CT is preferable for stroke but you can't use a CT in a room with other people. A low field MRI might be acceptable, but there is still the noise issue and obviously field safety (though this can be less when you operate at 0.1T or something).

That's a really interesting piece of hardware. Product design in the medical field is like nothing else on Earth.

Reply on news.ycombinator.com