MR-compatible Incubator for Imaging Pre- and Term Neonates

INTRODUCTION

Magnetic Resonance (MR) is a versatile tool well suited for non-invasive examination of neonates [1]. Though there is no risk associated with MR scans, its application in newborns is limited due to the associated logistical challenges of patient transport, comfort, life support, and physiological monitoring.
A modular prototype incubator with air temperature and humidity regulation and integrated high signal to noise ratio (S/N) radio frequency (RF) coils was developed [2]. We assessed the performance of the incubator by studying 13 neonates already scheduled for MR examinations of the head or body.

METHODS and MATERIALS

Patients
All phantom and patient examinations were carried out on a GE Signa CNV/I 1.5T MRI at Children's Hospital Los Angeles (CHLA), Los Angeles, USA. Thirteen patients, 4 female, 9 male were studied. The gestational age at birth ranged between 24 -41 weeks, age at the time of examination was 4 weeks pre-term to 12 wks. Body weight was between 1.2 – 4.6 kg and head circumferences were 29.5 – 37 cm. Patients were recruited from the Neonatal Intensive Care Unit (NICU) and the Pediatric Intensive Care Unit (PICU). Patients were transported to the MRI suite and setup for MRI using routine CHLA protocols but were placed in the MR incubator. Patients were sedated as requested for the clinical scan. Studies were approved by the CHLA Internal Review Board (IRB). Informed consent was obtained from all parents before the examinations.

MR Incubator
The incubator is shown in Fig. 1. It features custom RF coils (head, body) sufficient to cover 95th percentile of the newborn patient population specifically designed for use in combination with the incubator. The double-walled patient section has several hand ports for easy access from either side of the MR table. For details of the incubator, please see reference [2].

Patient Set-up
The incubator was placed on the MR table. Patient temperature was measured at the axilla before and after the MRI study using standard disposable temperature probes. Patients were placed inside the incubator and connected to life sustaining equipment such as oxygen lines, ventilator, and infusion pumps and with vital signs monitoring equipment (pulse, heart rate, Spo2). The air temperature of the incubator was set (28.5 - 36 °C) as requested by NICU staff. Studies were run at moderate levels of humidity (< 55% rH). A commercial fluoro-optic temperature measuring equipment (Luxtron Corporation, Santa Clara, CA), was used to monitor patient skin temperature during the MR examination. The probe was placed on the head or abdomen depending on whether a brain or abdominal examination was attempted. Accuracy of the Luxtron equipment and the chemical thermometer was within ±0.1 °C. Patient vital signs, skin temperature (recorded with Luxtron) and incubator settings (temperature, humidity) were monitored and recorded during the course of the MR study. Patient set up inside the pediatric body coil is shown in Figure 2. Please see the corresponding images for details of the respective study and diagnosis. Our coils fitted over the patients without obstructing the ventilator, nasal cannula and endo tracheal tubes or monitoring leads (ECG, Spo2). All were routed through the portals on the incubator (see Figure 1b) and connected to the respective devices located inside the MRI room.

Figure 2. Body coil/patient set up for the incubator assembly on the GE 1.5T MR patient table (a). Coronal images are shown to illustrate pediatric body coil coverage (from the pelvis to the knee), the axial SPGR T1 fatsat images are shown pre and post (gadolinium 0.9cc) contrast to show enhancement over the soft tissue mass consistent with lymphangioma (b). From the rest of the images (not shown), there was no evidence of bone or muscle involvement.

Phantom Experiments
The performance of the incubator in respect to maintaining temperature, oxygen level, and humidity was checked inside the MR before any MR tests were conducted. Thereafter, several MR experiments with model solution were carried-out to test the
a) MR performance of radiofrequency coils without incubator,
b) MR performance of radiofrequency coils with incubator – but incubator functions switched off,
c) MR performance of radiofrequency coils with incubator – with incubator on.

The signal-to-noise ratio (S/N) was obtained from the images and compared with S/N obtained from images acquired with the standard GE head coil. With the head coil for newborns a S/N improvement of a factor of ˜ 3 was achieved when compared with the standard imaging coil. Before any patient studies were carried, RF coils designated for the incubator were tuned and matched on a few patients to study loading considerations and optimize performance over the wide range of patient sizes. These adjustments were done at the CHLA inside the NICU and were approved by the CHLA IRB.

MR Sequences
The standard adult head coil configuration file with additional 9 dB attenuation was used, since lower powers were required for our experiments. This configuration file was used for the pediatric head and body examinations to ensure the RF power deposition stayed below FDA guidelines for the specific absorption rate (SAR).


MR imaging of the brain:

• axial T2w fast spin echo (FSE)
• sagital T1w, axial fluid attenuation inversed recovery (FLAIR)
• axial T1w, FLAIR
• axial, FLAIR
• single voxel PRESS MR spectroscopy

MR imaging of pelvis and abdomen

• coronal T2w FSE
• axial and coronal fat sat T2w FSE
• coronal, axial and sagital T1w SE
• sagital T2w single shot FSE
• axial and coronal T1w fat sat spoiled gradient echo pre and post contrast

MR imaging of the heart

• axial and coronal gradient echo segmented k-space CINE
• axial and coronal T1w gated spin echo
• contrast –enhanced MR angiography

RESULTS

Patient Studies
There where no adverse effects. MR images of excellent quality were acquired in all studies. Brain MRI/MRS data from a 9-weeks-old male patient with hydrocephalus are shown in Figure 3. Performance of the incubator remained was unchanged inside the MR room. Skin temperature increase recorded with the Luxtron equipment was under 0.5 °C for all patients except one where it approached 1.0 °C.

Figure 3. MRI/MRS of a 9 week old male patient with hydrocephalus. The sagittal T1 and axial FSE T2 images display uniform coverage over the neonate brain. Axial FLAIR and diffusion images show superior contrast-to-noise in addition to the high image S/N. The proton spectrum as suspected had low levels of NAA which could be related to the low brain mass.

Phantom Studies
S/N of coils were within 1-2% with the incubator ON and OFF which was under the manufacturer's daily tolerance for MRI system operation. There were no detectable differences in the quality of the images acquired with the incubator turned on or off, respectively. This demonstrated the compatibility of the incubator with MR.

DISCUSSIONS and CONCLUSIONS

No problems were encountered with placing the patients into the incubator. The increase in skin temperature in one patient was could be due to the prolonged MRI study (~65 minutes) and temperature auto-regulation of the body rather than RF power deposition. The temperature measured by the nurse before and after the scan were within 0.5 °C for all but one subject where a drop of 0.7 °C likely due to anesthesia slowing down metabolism was observed. This drop could have been avoided by setting a higher temperature for the incubator. The image resolution and quality is far superior to images obtained with standard equipment. With this study we demonstrated that MR studies can be successfully carried out using a incubator system that provides controlled temperature, oxygen, and humidity levels without compromising the quality. This may in the future expand the patient population for MR studies to those that are classified unstable for a MR examination at present.

REFERENCES
[1] Pediatric Neuroimaging, MRI Clinics of North America, February 2001 and references therein.
[2] Srinivasan, R. et al. 10th ISMR, Book of Abstracts, p 799, 2002
[3] Dumoulin, C.L. et al. 10th ISMR, Book of Abstracts, p 2558, 2002