Diagnostic and prognostic MRI is recommended for infants for a range of conditions. These include gestational age below 30 weeks, in premature infants suspected of metabolic disease, and in term infants who might have sustained perinatal brain injuries or who show Stage 2 or 3 hypoxic-ischemic encephalopathy.

MRI preferred imaging solution for numerous conditions
Although ultrasound (US) is used as first-line imaging in certain cases like intracranial hemorrhage, MRI is indicated for most other infant brain and head neuroimaging. This has been the case for some time. One example is a report published in 1990 in the French-language journal ‘Pediatrie’ by a team from the CHU Hautepierre hospital in Strasbourg, which discusses the advantages of MRI over ultrasound in areas such as brain injury. The report, nevertheless, also points out the problems with neonatal MRI, such as the need for immobilization and lack of accessibility.  Such difficulties have persisted over the years.
Indeed, in the early 1990s, Britain’s Hammersmith Hospital installed a 1T MRI scanner in the NICU. However, it had a limited field of view and was replaced with a conventional adult-sized 3T system. In fairly short order, the 3T system was found not only challenging to use in the NICU due to its long bore and problems of access to infants, but also expensive to operate.

Guidelines for infant MRI imaging
At present, a multitude of guidelines recommend that MRI is used to follow up ultrasound diagnosis of parenchymal brain injury, post-hemorrhage ventricular dilatation as well as US (or clinical) suspicion of abnormalities in the posterior fossa and at the brain’s convexity. Other conditions in infants that indicate MRI imaging include brain inflammation (meningitis, encephalitis, brain abscess etc.) and seizures, abnormal consciousness and/or asymmetry which cannot be satisfactorily explained by US findings.
The case for MRI after ultrasound has also been studied extensively. One report from the Medical University of Vienna in 2010 stated that among infants undergoing cranial ultrasounds after clinical seizure, MRI was able to identify a causative pathology in 42% of cases where US findings were unspecific.

Conventional MRI “not designed” for infants
As mentioned in an ‘Advances in Neonatal Care’ analysis in 2005, it takes a single look at a typical MRI scanner to know that “it was not designed for an infant.”
Technically, a baby’s head size poses one of the first challenges. Standard MR head coils lead to sub-optimal picture quality and adult knee coils are often used instead.
Cooperation between neonatal team and radiologists
Given the very small size of a neonate brain, it is especially important to have high signal-to-noise ratios (SNR) for delineation of anatomical details. This was one of the major limitations of smaller, customized low-field MRIs designed for NICUs. At Royal Hallamshire Hospital in Sheffield, for example, a 0.17T system with 15mT/m gradients was installed in the early 2000s, but its low SNR made it impossible to use emerging  MRI techniques such as diffusion tensor imaging and MR spectroscopy in neonates.

The best way forward has instead been seen in tailoring MR protocols to the neonatal brain. This is however a complex task. MR protocols involve a wide range of technical factors: echo time, repetition time, flip angle, slice numbers, slice thickness, scan duration, field of view etc. Achieving this “requires close cooperation between the neonatal team, radiographers and radiologists,” according to a study at Ireland’s University of Cork, published in 2012 in the ‘British Journal of Radiology’.

The challenge of transfers
The transfer of infants from a continuously-monitored NICU to MRI suites has been one of the most vexatious problems. As discussed in the 2005 edition of ‘Advances in Neonatal Care’ cited above, MRI scanners “are often situated far away from the NICU.”
The move of infants to an MRI room involves multiple transfers – from NICU bed to incubator to scanning table, and then backwards. These have to be made in a relatively short period of time, which can add dramatically to physiological stress.
Specific problems during transfer include the chance of extubation and arterial or venous decannulation. Excessive movement in a premature infant is also known to adversely affect cerebral blood flow. This, in turn, can defeat the very purpose of an MRI, by altering results.

Sedation and hypothermia
The question of whether or not to sedate infants before transfer is also a major challenge. Sedation has risks. Moreover, a sedated neonate requires continuous monitoring during an MRI.
There are problems after the transfer, too. Once in the MRI room, infants must be removed from the warmth of the incubator to a cooler scanning table. Towards this, they are usually swaddled in blankets, accompanied sometimes by neonatal thermal packs to prevent heat loss. The American College of Radiology (ACR) also recommends use of temperature probes for infants to take an auxiliary temperature before and after the examination.
Even as the MRI begins, NICU staff need to be on alert to decide if an examination must be halted. This may be due to the impact of the transport, cold, stress, sedation etc..

MRI-compatible incubators
Since the early 2000s, attention has focused on MRI-compatible incubators. These are equipped with an integrated head coil and accompanied by auditory shielding, temperature and humidity regulators, a ventilation support system and monitors specifically certified for the massive magnetic environment of the MRI.
In February 2004, ‘Pediatrics’ published a report on the imaging of seven non-sedated neonates via the use of an MRI-compatible incubator. The authors noted that the “constant environment reduces the risk of adverse events occurring during the transport and imaging of the neonate.”
Not all problems, however, were mastered by the incubator. For instance, the infant was not easily visible from the control room and required the presence of a staff member in the vicinity. In addition, in spite of temperature and humidity controls, additional monitoring was required for electrocardiography and oxygen saturation.
Nevertheless, interest in MRI-compatible neonatal incubators has continued.
In September 2010, the ‘European Journal of Paediatric Neurology’ published results of a study which found that MRI-compatible incubators reduced the mean gestational age of patients from 44 to 39.7 weeks, and in parallel, more than doubled incubator use from 14.8% to 36% for ventilated neonates.
Advantages of the MRI-compatible neonatal incubator also included halving the time required for handling the infant, a reduction of total procedure time by an average of 20 minutes, and in imaging time by four minutes. Such time savings arose from the fact that there was no need to stabilize the infant. Furthermore, no MRI procedure was terminated due to insufficient sedation or infant instability; previously, one in 10 infants had required additional sedation during the procedure.
Equipment compatibility and safety
In May 2013, researchers from Australia’s Royal Brisbane and Women’s Hospital published results of a three-year review on MRI-compatible incubators in the ‘Journal of Paediatrics and Child Health’. Although the overall conclusions were positive, with no adverse incident reported over the period, the authors drew attention to several “practical issues”.
The first was a 30-45 minute pre-warming period required to reach an appropriate temperature setting for babies. The second consisted of difficulties in reading the incubator’s patient monitor interface, including key data such as cot temperature, pulse rate and oximetry readings. Once again, as with the February 2004 ‘Pediatrics’ study mentioned previously, the Royal Brisbane researchers recommended “that staff remain in the scan room throughout the procedure to monitor the well-being of the baby.”
The biggest challenge, however, concerned compatibility of equipment connected to the incubator. For instance, though the ventilator was MRI-compatible, it was not designed to provide humidified or preheated gas. The researchers also noted the need to improvise very specific procedures, for example, in extending infusion lines from pumps located outside the imaging room, which were not MRI-compatible.
Indeed, the need to use MRI-compatible or MRI-safe accessories, ranging from thermal packs and temperature probes to noise protectors, remains one of the biggest drawbacks with MRI-compatible incubators outside the NICU. The authors of the Royal Brisbane study point to “difficulties in sourcing a gas supplier to refill the portable MRI-compatible air and oxygen cylinders because of their special status outside the usual medical gas cylinder refilling programme.”
The scale of such problems becomes dramatic when intubation or resuscitation is required. In such cases, the infants need to be rapidly removed from the MR system and its magnetic fringe. The only alternative is to ensure that, rather than just accessories, the entire range of medical equipment – from syringes and infusion pumps to laryngoscopes and suction equipment – is MRI-compatible.

More research needed
In February 2015, ‘Advances in Neonatal Care’ published results from a systematic review of 13 research studies, two quality improvement projects, as well as practice guidelines and articles on neonatal MRI imaging by the Norwegian Neonatal Network and Oslo University Hospital.
The authors concluded that although results seemed promising and increasingly consistent, “more research is needed before conclusive recommendations” could be established about MRI-compatible incubators and associated techniques.

Alternatives emerge
Recently, a system from Aspect Imaging known as Embrace Neonatal MRI has sought to close the gap between NICU imaging requirements and the capabilities of current MRI-compatible incubators.  Embrace received authorization from United States Food and Drug Administration (FDA) in July 2017, and in November obtained a CE marking for European Union sales.
Unlike conventional MRI machines, the new system does not require a safety zone or a radio-frequency shielded room. Since it is fully enclosed, medical device implants or equipment in the NICU in close proximity are not required to be MRI-compatible.  Other advantages include an always-on permanent magnet; it therefore requires no electrical, cryogenic or water cooling (click here for more details on this product).

Other approaches to neonate imaging are also under evaluation.
Cincinnati Children’s Hospital in the US, for example, has installed a commercial 1.5-T MRI system in its NICU, based on an orthopedic system coupled to custom-built components – most significantly, a high-end scanner. The unit’s gradient coil is about 2.5 times shorter than a conventional adult-sized system. In January 2014, the ‘American Journal of Roentgenology’ published results of a study at the hospital on imaging neonates. Although its scope was small (15 infants), the authors concluded that the system was capable of producing “high quality” images of neonates, not only of the brain but also the abdomen and chest.
As with other efforts to date, the modified system also attained several collateral objectives, such as ease of installation and operation in an NICU, improved visual contact and physical access to the infant, along with the use of advanced imaging techniques, ECG and respiratory gating and triggering.  One of “the most important benefits”, according to the authors, consisted of “the reduction of risk associated with transport of the neonate to and from the NICU.” As discussed previously, this has been the single biggest challenge for neonate imaging and a driver of most design and technology development for over 25 years.