What six years of bedside scans actually show
Researchers reviewed portable MRI utilization at an urban tertiary care center between July 2018 and December 2024, publishing the results in Neurology as "Portable MRI in the Hospital Setting: A Six-year Evaluation of Safety, Tolerability, and Utility." The team, based at Yale School of Medicine's departments of neurology and radiology, ran exams on a 0.064-T low-field portable MRI system in the emergency department, intensive care units and operating room.
Of 1,173 attempted examinations, 1,060 — or 90.4% — were completed successfully. No serious adverse events occurred over the entire six-year period. Among the 113 incomplete scans, the leading cause by far was the patient asking to stop (77 cases, 68.1%) — 27 for physical discomfort, 23 for claustrophobia, 27 unspecified. Only a small minority failed for reasons related to the equipment or the patient's clinical state: physical incompatibility with the device (10), machine failure (8), or clinical instability (7).
The neuro ICU accounted for the largest share of scans (523, or 48.7%), followed by the emergency department (346, or 32.2%) and the medical ICU (58, or 5.4%). By clinical indication, suspected acute ischemic stroke or stroke-code evaluations made up 32.4% of exams (371) and intracerebral hemorrhage 11.4% (131), with the remainder covering white matter hyperintensities, tumors, subarachnoid hemorrhage and post-cardiac-arrest evaluation. The authors' own conclusion, as reported by Radiology Business: the tool is "safe, well-tolerated and feasible across acute care settings."
Why hospitals want the scanner at the bedside
The case for point-of-care MRI has always been about the patient, not the pixels. Transporting a critically ill, ventilated or hemodynamically unstable patient to a fixed MRI suite means disconnecting and re-connecting monitoring equipment, coordinating staff, and briefly interrupting intensive care — all before the scan even starts. A portable, low-field system removes that step by bringing the magnet to the bed.
Earlier, smaller studies quantified that trade-off directly. A UC Irvine analysis of bedside point-of-care exams found ICU MRI turnaround time dropped from a median of 11.7 hours with a fixed scanner to 5.3 hours with a portable unit — though only 72% of those bedside scans reached full diagnostic image quality, mostly limited by patient motion, as Radiology Business reported. The 2026 Yale data extend that early signal to six years and over 1,100 scans, which is why the completion rate and safety profile matter: this is no longer a small pilot, it's a long-running operational track record.
What the data don't cover: the report
Every figure in the Yale study describes whether the scan itself could be safely acquired at the bedside. None of it describes what happens after the scan finishes — how long the images sit before a radiologist reads them, drafts findings, and a final report reaches the chart. That is a separate constraint, governed by radiologist staffing and coverage, not by where the magnet sits.
The mismatch matters here specifically because of who gets a portable scan. Stroke codes, intracerebral hemorrhage and post-cardiac-arrest evaluations don't wait for business hours, and neuro ICUs run reads around the clock. A bedside scanner that eliminates a transport delay still hands off to the same overnight or weekend reporting queue every other stroke CT or MRI depends on — the queue this data set never measured. It's the same relocation-not-resolution pattern seen when acquisition-speed AI cut one hospital system's MRI wait times by more than 60%: faster scans just moved the constraint downstream to reporting.
| Constraint | What portable MRI addresses | What still depends on the reporting queue |
|---|---|---|
| Transporting an unstable ICU patient | Removed — scanner comes to the bed | — |
| Waiting for a fixed-scanner slot | Reduced with a dedicated bedside unit | — |
| Time to image acquisition | Comparable, sometimes slower at low field strength | — |
| Time to a finalized, signed report | Not addressed by the hardware | Still gated by radiologist availability |
| Off-hours and weekend coverage | Not addressed | Same coverage gaps as any other modality |
Where xAID fits
xAID's own focus is CT, not MRI, but the gap this data set exposes is modality-agnostic: acquisition can get safer and more convenient without reporting getting any faster behind it. On CT, xAID's AI drafts a structured, ready-to-sign report within minutes of a study reaching the queue, xAID's in-house radiologist reviews every preliminary, and the client's reading radiologist signs the final — the same day or night the study was acquired, not whenever the next opening in a reporting queue appears. Whatever produces the images, someone still has to read them; that's a reporting-workflow problem, and it needs a reporting-workflow fix, not just a more portable magnet. See the fuller picture in CT report turnaround time benchmarks.
Frequently asked questions
What did the new portable MRI study find?
Researchers reviewed portable MRI use at an urban tertiary care center between July 2018 and December 2024, publishing the results in Neurology. Using a 0.064-T low-field system in the emergency department, ICUs and operating room, 1,060 of 1,173 attempted scans (90.4%) were completed successfully with no serious adverse events over the six-year period. Most incomplete scans stopped because the patient asked to stop, not because of equipment failure or instability.
Does portable MRI reduce risk for critically ill ICU patients?
That's the rationale for the technology: transporting an unstable, ventilated or monitored ICU patient to a fixed MRI suite requires coordination and carries its own clinical risk, which a bedside scan avoids. A separate, earlier analysis at UC Irvine found ICU MRI turnaround time fell from a median of 11.7 hours with a fixed scanner to 5.3 hours with a portable point-of-care unit, though 72% of those bedside scans met full diagnostic quality and image quality issues remained more common than with fixed MRI.
Does a faster portable MRI scan mean a faster final report?
Not automatically. The six-year Yale data measure whether the scan can be safely completed at the bedside — they say nothing about how quickly a radiologist reads and finalizes the resulting images. Portable MRI is disproportionately ordered for stroke codes, intracerebral hemorrhage and post-cardiac-arrest evaluations in ICUs, exactly the acute, often off-hours scenarios where a reporting queue is most likely to be the actual bottleneck, not the scanner. Moving the scanner to the bedside removes a transport problem; it doesn't add reporting capacity.
How does xAID fit into the reporting side of this problem?
xAID's own focus is CT, not MRI, but the mismatch this data exposes — faster or safer acquisition without faster reporting — is modality-agnostic. On CT, xAID's AI drafts a structured, ready-to-sign report within minutes of a study reaching the queue, xAID's in-house radiologist reviews every preliminary, and the client's reading radiologist signs the final. Whatever produces the images, someone still has to read them; closing that gap is a reporting-workflow problem, not an acquisition-hardware one.
Source: "Portable MRI in the Hospital Setting: A Six-year Evaluation of Safety, Tolerability, and Utility," Neurology (2026), doi.org/10.1212/WNL.0000000000216224, as reported by Radiology Business. UC Irvine turnaround comparison via Radiology Business. Figures are rounded as reported.