Magnetic Resonance Imaging of the Codman Microsensor Transducer Used for Intraspinal Pressure Monitoring: Findings From the Injured Spinal Cord Pressure Evaluation Study.

STUDY DESIGN: Laboratory and human study. OBJECTIVE: To test the Codman Microsensor Transducer (CMT) in a cervical gel phantom. To test the CMT inserted to monitor intraspinal pressure in a patient with spinal cord injury. SUMMARY OF BACKGROUND DATA: We recently introduced the technique of intraspinal pressure monitoring using the CMT to guide management of traumatic spinal cord injury [Werndle et al. Crit Care Med 2014;42:646]. This is analogous to intracranial pressure monitoring to guide management of patients with traumatic brain injury. It is unclear whether magnetic resonance imaging (MRI) of patients with spinal cord injury is safe with the intraspinal pressure CMT in situ. METHODS: We measured the heating produced by the CMT placed in a gel phantom in various configurations. A 3-T MRI system was used with the body transmit coil and the spine array receive coil. A CMT was then inserted subdurally at the injury site in a patient who had traumatic spinal cord injury and MRI was performed at 1.5 T. RESULTS: In the gel phantom, heating of up to 5°C occurred with the transducer wire placed straight through the magnet bore. The heating was abolished when the CMT wire was coiled and passed away from the bore. We then tested the CMT in a patient with an American Spinal Injuries Association grade C cervical cord injury. The CMT wire was placed in the configuration that abolished heating in the gel phantom. Good-quality T1 and T2 images of the cord were obtained without neurological deterioration. The transducer remained functional after the MRI. CONCLUSION: Our data suggest that the CMT is MR conditional when used in the spinal configuration in humans. Data from a large patient group are required to confirm these findings. LEVEL OF EVIDENCE: N/A.

Combining MRI with PET for partial volume correction improves image-derived input functions in mice.

Accurate kinetic modelling using dynamic PET requires knowledge of the tracer concentration in plasma, known as the arterial input function (AIF). AIFs are usually determined by invasive blood sampling, but this is prohibitive in murine studies due to low total blood volumes. As a result of the low spatial resolution of PET, image-derived input functions (IDIFs) must be extracted from left ventricular blood pool (LVBP) ROIs of the mouse heart. This is challenging because of partial volume and spillover effects between the LVBP and myocardium, contaminating IDIFs with tissue signal. We have applied the geometric transfer matrix (GTM) method of partial volume correction (PVC) to 12 mice injected with 18F-FDG affected by a Myocardial Infarction (MI), of which 6 were treated with a drug which reduced infarction size [1]. We utilised high resolution MRI to assist in segmenting mouse hearts into 5 classes: LVBP, infarcted myocardium, healthy myocardium, lungs/body and background. The signal contribution from these 5 classes was convolved with the point spread function (PSF) of the Cambridge split magnet PET scanner and a non-linear fit was performed on the 5 measured signal components. The corrected IDIF was taken as the fitted LVBP component. It was found that the GTM PVC method could recover an IDIF with less contamination from spillover than an IDIF extracted from PET data alone. More realistic values of Ki were achieved using GTM IDIFs, which were shown to be significantly different (p<0.05) between the treated and untreated groups.