Magnetic Resonance Microscopy. Группа авторов

one hemisphere).

      While CT (including portable CT) is effective at identifying a mass effect, it is less useful for answering the second most important clinical question: “Is it getting worse?” Because the mass effect and thus increasing pressure could be developing for hours post injury (e.g. in the case of a hemorrhage after trauma), a single imaging time point is insufficient. Multiple time points are desirable and would ideally be acquired with a wearable monitor. CT is still relatively cumbersome to use repeatedly in a POC setting and the ionizing radiation makes it ill-suited for anything resembling continuous monitoring. Instead, compact, single-sided “MR monitors” could fill this gap. This is a potential new role for MR technology in medicine: a real-time monitor of ventricular/CSF asymmetry to provide an early warning sign of impending herniation, particularly in patients where clinical exam is difficult (e.g. sedated patients).

In a true monitoring scenario, the patient will have the MR monitor on for hours, perhaps days, likely precluding traditional architectures where the patient goes “inside” the bore (of even a small magnet). Instead, the magnet is “worn” or simply adjacent to the head. Such a device will not provide a homogeneous field, or even allow imaging over the whole head. Instead, the clinical information must be extracted from a 3D image of only a region of brain, or perhaps a 2D or 1D image. Nonetheless, this line of instrumentation can follow the extensive development effort previously applied to single-sided MR devices for spectroscopy and materials evaluation [49,71,72], and relaxometry and diffusion measurements in oil well logging [73].

      3.4.4 Neonatal Intensive Care Unit (NICU)

A scanner situated in an NICU or capable of being brought to the NICU (beside incubator/isolette) could have many uses. Here we focus on one: HIE to provide a general picture of the uses and issues of MR in a NICU.

HIE and resulting hypoxic-ischemic injury (HII) is the leading cause of neonatal mortality and morbidity with an occurrence of 1 to 8 out of 1000 live births in the United States and up to 26 out of 1000 births in developing countries [74,75]. In survivors, HIE is a predictor of permanent neurodevelopmental disability. The significance of early diagnosis and intervention in HIE is profound, particularly when considering the long-term consequences in remaining years of life.

Once HIE is identified, immediate supportive care is applied to encourage brain perfusion, including intubation and mechanical ventilation for more severely affected neonates, titration of intravenous fluids, and blood pressure support [76]. Continuous electroencephalogram (EEG) monitoring is used to monitor background brain electrical activity and seizures. Expedient intervention with hypothermia neural rescue therapy within the first few hours of life is vital, improving survival rates and reducing future disabilities [77]. Thus, moderate to severe cases undergo neuroprotective interventions with therapeutic hypothermia (also known as hypothermia neural rescue therapy or “cooling”) where a cooling blanket and temperature monitor maintain a body temperature decrease of ~3°C for 72 h. Successful treatment of neonatal brain injury could eliminate many decades of disease burden and healthcare needs.

As the gold standard for brain imaging, MRI is the best imaging technique for diagnosing and staging HIE, particularly with diffusion-weighted contrast [78]. CT is less suitable due to its limited soft tissue contrast and the risks of exposing neonates to ionizing radiation. Although the preferred imaging method, MRI currently requires transport of the fragile patients to a radiology suite. This introduces safety concerns, since it requires a full hand-off of patient care responsibilities and exposes the neonate to stress from transport and acoustic noise. In particular, HIE patients are often not stable enough to transfer to the scanner for several days after birth. Although safety is the number one concern, we must also consider the burden on the hospital workflow. Transfer of critical care patients requires tremendous teamwork and coordination from multiple sources. For example, a neonatologist, nurse, and respiratory therapist must accompany the infant to the scanner and remain with the patient for the duration of the study. The time required for the scan and transfer to/from the NICU is around three hours and NICU staffing coverage must be implemented during their absence. Currently, bedside cranial imaging is available with ultrasound, but provides low sensitivity for early abnormalities associated with HIE [78].

Siting a small footprint MRI in the NICU would provide tangible benefits and a bedside (portable) MRI could offer additional benefits. Figure 3.5 illustrates some existing MRI systems and components that have been developed or adapted for MRI in the NICU. The benefits of bedside MRI accrue since even the trip from one side of the NICU to the other is difficult due to the removal/replacement of leads, tubes, ventilation, and temperature maintenance equipment. Furthermore, intravenous (IV) pumps are not MR compatible so extension tubing must be placed, which can interrupt the administration of critical medications (e.g. dopamine). Even if “MR-safe,” noncompatible or metallic equipment can cause susceptibility artifacts, which are more severe at higher field strengths [79]. All of these issues point to an extremely compact magnetic footprint magnet; likely a mid- to low-field magnet design with minimal fringe field.

      Figure 3.5 MR devices designed to support imaging in neonatal intensive care unit (NICU). (a) LMT MRI compatible incubator (LMT Medical Systems GmbH). (b) GE Healthcare investigational 3-T neonatal MR system installed at Sheffield Teaching Hospital in a dedicated NICU scan room (EMAP Publishing Limited). (c) The Hyperfine 64-mT portable MRI scanner. (d) Aspect Imaging’s Embrace neonatal scanner, a dedicated 1-T permanent magnet scanner designed for installation in the NICU without a shielded room. (e) Modified 1.5-T GE orthopedic scanner adapted for the NICU at Cincinnati Children’s Hospital.

      3.5 Technological Approaches to POC and/or Portable MRI

The goal of reducing cost, siting needs, and even enabling portability has resurrected interest in low-field MRI technology, including new magnet and encoding methods. Low-field MRI, loosely defined as scanners operating between 10 mT and 500 mT, has a long history in MRI, starting with the first human imaging systems [80–83]. Although high-field superconducting magnet-equipped scanners have largely displaced them, thousands of low-field MRI systems (mainly in the 0.2–0.35-T range) are in daily operation and their clinical capabilities are well known. Additionally, commercial systems operating below 100 mT have an established clinical history; for example, the Toshiba Access 64-mT system introduced in 1991. These older generation low-field systems are often called “open MRIs” due to the two-pole-piece dipole magnet geometry used. They are still big and heavy and have most of the siting issues of their high-field superconducting cousins. This review focuses on the modern generation of low-field scanners specifically aimed at the three POC uses outlined earlier. The “low” vs. “ultralow” tradeoffs have been recently reviewed in depth [84].

In addition to low-field architectures, considerable research has also focused on ultralow field (ULF), with B₀ below 10 mT [85]. However, several barriers remain to using ULF approaches to achieve low-cost or POC and/or portable medical imaging. The sensitivity of MR as a function of field strength is well known [18], and the severe penalty at ULF requires either additional technology such as prepolarization [86], cryogenic detectors [87–90], hyperpolarized media [91], or efficient but nonstandard brain-imaging sequence schemes [92]. This additional technology can be a barrier to achieving reduced costs or portability. Additional issues arise for image encoding at ULF due to unwanted encoding by the gradient coil’s concomitant field terms [93,94].

Because of the signal-to-noise ratio limitations of ULF MRI, this review focuses mainly on compact magnets with B₀ of above 50 mT or compact superconducting magnets in the 0.5–1.0-T range. We focus on magnet and system architecture and component technology; the constraints of MR physics at low field are well understood and are discussed in several recent reviews [15,18,19,84,95,96].

      3.5.1 Magnet Designs

      3.5.1.1 Advances in Cryogenics for Supercon