Hyperfine MRI: Innovations in Portable Imaging

Hyperfine’s portable MRI scanner operates at 64 millitesla, roughly one-fiftieth the magnetic field strength of a conventional hospital MRI, and it can be wheeled directly to a patient’s bedside on standard hospital power. That combination of low field strength, small footprint, and genuine portability has opened up imaging scenarios that were previously impossible or impractical, from scanning stroke patients in the ICU to imaging newborns without transporting them out of intensive care. The technology is still young compared to the high-field systems it complements, but a growing body of clinical evidence and some clever engineering are rapidly defining what portable MRI can and cannot do.

What Makes a Portable MRI Possible

Conventional MRI scanners use superconducting magnets cooled by liquid helium to generate powerful magnetic fields, typically 1.5 or 3 tesla. Those magnets weigh several tons, require dedicated shielded rooms, and need specialized cooling infrastructure. Hyperfine’s Swoop system takes a fundamentally different approach: it uses a permanent magnet that needs no cryogenic cooling at all. Cryogen-free magnet designs have simplified installation and dramatically lowered operational costs for low-field systems in general.¹ The trade-off is signal strength. At 64 mT, the raw signal coming from hydrogen atoms in tissue is far weaker than at 1.5 or 3 T, which means the images are inherently noisier and lower in resolution. Overcoming that gap has become the central technical challenge, and much of the innovation around portable MRI is really about solving it.

The portability itself matters more than it might sound. A conventional MRI suite is a fixed installation with a radiofrequency-shielded room, specialized power supply, and a climate-controlled environment. Hyperfine’s system plugs into a standard wall outlet, weighs around 630 kilograms, and fits through a standard doorway. Progress in electromagnetic noise cancellation and machine learning reconstruction algorithms has been critical to making this work outside a shielded room, where interference from nearby electronics, power lines, and other hospital equipment would otherwise overwhelm the weak signal.²

Closing the Image Quality Gap with Deep Learning

The most significant recent innovation in portable MRI is not a hardware improvement but a software one. Multiple research groups have developed deep learning models that take noisy, low-resolution images from low-field scanners and reconstruct them into something much closer to what a high-field system would produce. One approach called AUTOMAP, an end-to-end deep neural network for image reconstruction, demonstrated signal-to-noise improvements of 1.5- to 4.5-fold over standard reconstruction methods when applied to human brain data acquired at 6.5 mT, outperforming two other contemporary denoising algorithms.³

Another approach, called LoHiResGAN, takes paired images acquired on both a 64 mT scanner and a 3 T scanner and trains a model to translate low-field images into synthetic high-field versions. The resulting synthetic 3 T images significantly improved image quality compared to several other deep learning methods and provided more consistent brain measurements across different brain regions when compared against actual 3 T scans.⁴ These AI-based reconstruction techniques are not just cosmetic. They are starting to make portable MRI clinically useful for diagnoses that require reasonably precise anatomical detail, not just gross abnormality detection.

Detecting Stroke and Hemorrhage at the Bedside

Stroke is the clinical scenario where portable MRI has seen the most testing, and for good reason. In acute stroke, every minute matters, but getting a critically ill patient to a conventional MRI suite often takes an hour or more when you account for scheduling, transport logistics, and safety screening. A scanner that rolls to the patient’s room eliminates all of that.

For intracerebral hemorrhage, a study of 144 portable MRI examinations at Yale New Haven Hospital found that neuroradiologists correctly identified bleeding in about 80% of cases (45 of 56), with high specificity of roughly 97% for ruling out hemorrhage in patients who did not have it.⁵ That sensitivity is not yet where it needs to be to replace CT as the gold standard for hemorrhage detection, but it is high enough to be clinically useful for monitoring known bleeds or screening patients who cannot be transported.

For ischemic stroke, the picture has improved with newer hardware and imaging sequences. Research on enhanced diffusion-weighted imaging found that the positive predictive value for distinguishing acute ischemic stroke from stroke mimics reached 95% using multi-directional diffusion imaging on a next-generation portable system. That same upgrade reduced scan time by about 30% and could detect lesions as small as 0.15 mL.⁶ Detecting tiny stroke lesions on a 64 mT system would have seemed implausible just a few years ago, which speaks to how fast both hardware and software are evolving.

Portable MRI has also been validated for detecting midline shift, a dangerous displacement of brain structures that occurs after large strokes or hemorrhages and often determines whether emergency surgery is needed. A study of 102 patients (48 with ischemic stroke, 54 with intracranial hemorrhage) found strong agreement between portable MRI and standard imaging, with sensitivity of 93% and specificity of 96% for identifying midline shift.³

Scanning Critically Ill Patients in Place

One of the most compelling use cases for portable MRI is in the intensive care unit, where patients are often too unstable to transport. Moving a ventilated patient through hospital corridors to a radiology suite carries real risks: lines can disconnect, ventilator settings may need adjustment, and hemodynamic instability can worsen. An early feasibility study of bedside portable MRI in ICU patients found no adverse events across all scans. Patients remained in their rooms connected to all necessary lines and mechanical ventilation throughout the scanning process, with no inadvertent line disconnections or extubations.⁷

The low field strength is what makes this possible. At 64 mT, the magnetic field is weak enough that standard ICU equipment, infusion pumps, monitoring leads, and ventilator tubing can stay in place. In a conventional MRI suite, any ferromagnetic object in the room is a projectile hazard, and even non-ferromagnetic devices can malfunction from the strong field. The portable system’s low field eliminates most of those concerns, though it does not eliminate them entirely, a point worth returning to.

Pediatric and Neonatal Imaging

Getting brain imaging for a sick newborn usually means transporting the infant from the neonatal intensive care unit to a radiology suite, often to a different floor. For fragile premature babies, that transport introduces risk. Portable MRI can be brought directly to the NICU bedside, and several studies have now tested this in practice.

A study of 102 paired scans in 87 newborns, including some as early as 31 weeks postmenstrual age, found no adverse events during portable MRI scanning. The optimized imaging sequences could visualize key brain anatomy and brain abnormalities, though the infants had a range of intensive care requirements during the scans.⁸ A separate feasibility study found that 94% of neonatal examinations were completed without image-impairing artifacts, though about half showed some radiofrequency interference that did not prevent interpretation.⁹

For older children, portable MRI has shown strong results in monitoring hydrocephalus, a condition where fluid builds up in the brain’s ventricles. A study of 153 pediatric hydrocephalus patients found near-congruent agreement between portable MRI and standard imaging for measuring ventricular size. The concordance correlation coefficients were above 0.92 for both of the standard indices used to track ventricular enlargement.¹⁰ For a condition where children may need dozens of follow-up scans over their lifetime, often requiring sedation for conventional MRI, a quick bedside scan is a meaningful quality-of-life improvement.

Beyond the Brain

Most published research on Hyperfine’s portable system focuses on neuroimaging, but the same hardware has been tested for musculoskeletal imaging. A feasibility study of knee imaging at 64 mT demonstrated that the portable system could adequately visualize knee joint structures, with no significant difference in readers’ evaluations compared to diagnostic standards. Researchers noted the potential for rapid, point-of-care diagnosis of acute knee injuries.¹¹

Earlier-generation portable MRI systems had already demonstrated advantages in joint imaging for rheumatology. In a study of 131 patients with inflammatory arthritis, a portable MRI system identified bony erosions in 95% of patients, compared to just 59% detected by conventional radiographs.¹² The open design of many portable systems also helps with patients who are claustrophobic or cannot fit into a conventional bore, and because low-field scans are quieter, they tend to be more tolerable overall.¹³

Cost and Operational Impact

Portable MRI is dramatically cheaper than conventional systems. A high-field MRI scanner can cost several million dollars to purchase and hundreds of thousands annually to operate, plus the cost of building and maintaining a shielded suite. Hyperfine’s system costs a fraction of that, and the operational savings extend beyond the purchase price.

An analysis of portable MRI’s economic and operational impact in an acute care hospital setting estimated total care savings of about $82,000 from avoided patient transports, plus an additional $66,000 saved by eliminating the need for MRI-compatible ICU supplies that are normally required when transporting patients to the radiology suite. The study found that portable MRI improved imaging access and optimized labor resources without compromising patient safety.¹⁴ Those savings represent a single institution over a limited period, so the numbers should not be extrapolated too broadly, but they illustrate the operational advantages of bringing the scanner to the patient rather than the reverse.

Expanding Global Access to MRI

The starkest promise of portable MRI is in parts of the world where conventional MRI simply does not exist. Across much of sub-Saharan Africa and South Asia, MRI scanners are scarce. Many countries have fewer than one machine per million people, compared to dozens per million in high-income countries. Ultra-low-field MRI systems, with their low weight, shielding-free design, and the ability to run on battery power, are fundamentally more suited to the installation conditions in low- and middle-income countries than anything that came before.¹⁵

At least one company (Hyperfine) has received FDA clearance for its portable system, and the Bill and Melinda Gates Foundation has committed to installing more than 45 of these units across Africa.¹⁵ Low-field and portable MRI technology has been specifically highlighted for its potential in expanding global access to high-quality radiological care in point-of-care, remote, and intraoperative settings.¹⁶

But deploying hardware is the easier part. A qualitative study of low-field MRI implementation in low- and middle-income countries found that practical challenges included infant positioning difficulties, unstable power supply, and unreliable internet connectivity, which matters because some of these systems rely on cloud-based image processing. Users recommended developing locally tailored reference materials, conducting refresher training, and providing ongoing technical and maintenance support.¹⁷ There are also ethical questions. When scans are performed in remote locations, who reads them? Teleradiology, outsourcing scan interpretation to experts elsewhere, is one solution, but it raises concerns about reliability, data security, and the challenge of communicating results in culturally appropriate ways to patients in remote communities.¹⁸

Practical Barriers in Emergency Settings

An even more ambitious idea is putting portable MRI into ambulances, so stroke patients could be scanned during transport before they even reach a hospital. Research into the feasibility of this concept identified several barriers: the need for specialized MRI operation skills among paramedics, space constraints inside ambulances, procedural complexity of patient handling around the scanner, and a lack of existing training programs or policies for prehospital MRI.¹⁹ Ambulance MRI remains more concept than reality for now, but the fact that it is being seriously studied reflects how far the technology has come.

Implant Safety Is Real but Not Automatic

One frequently mentioned advantage of low-field MRI is that it can scan patients with metallic implants who would be excluded from conventional high-field systems. Cardiac pacemakers, cochlear implants, certain joint replacements, and other devices can pose serious risks in a 1.5 or 3 T scanner because the strong magnetic field can heat metal, cause it to move, or interfere with electronic function. At 64 mT, those forces are dramatically reduced.

A case report described using portable MRI to image a patient with a cardiac implantable electronic device who presented with acute ischemic stroke and was excluded from conventional high-field imaging. The authors noted that continued technological refinement would improve diffusion-weighted imaging quality, and that larger studies would be needed to generalize portable MRI-based imaging for patients with such devices.²⁰ The evidence base for implant safety at low field is growing, but a comprehensive review from the International Society for Magnetic Resonance in Medicine warned that the assumption that lower fields are universally safe may foster a false sense of security and potentially lead to hazardous situations.²¹ The risks are smaller at low field, but they are not zero, and device-specific testing still matters.

Contrast Agents Designed for Low Field

Standard MRI contrast agents, typically gadolinium-based compounds, were developed and optimized for high-field systems. At low field strengths, the physics of how contrast agents interact with surrounding water molecules changes in ways that can actually be exploited. Superparamagnetic iron oxide nanoparticles, or SPIONs, have emerged as a promising contrast agent for low-field MRI. At 64 mT, certain SPIONs achieved longitudinal relaxivity values more than eight times higher than a commercial gadolinium-based agent at the same field strength, and more than an order of magnitude higher than corresponding values at 3 T.²²

The ratio of transverse to longitudinal relaxivity for these particles was close to one at 64 mT, which indicates a favorable balance for producing bright signal enhancement on standard imaging sequences.²² Further preclinical work has confirmed that the performance of these particles depends primarily on their core size, with both relaxivity and susceptibility increasing as the particles get larger, though coatings and solvents also play a role.²³ Iron oxide-based agents also sidestep some safety concerns associated with gadolinium, which can accumulate in the brain and has been linked to a rare kidney condition in patients with severe renal impairment. If iron oxide nanoparticles prove safe and effective in clinical trials, they could become the standard contrast agent for portable MRI, specifically engineered for the low-field environment rather than borrowed and compromised from the high-field world.