2024 American Academy of Neurology Abstract Website

Julia Zabinska¹, Vineetha Yadlapalli², Mercy Mazurek³, Nethra Parasuram¹, Dheeraj Lalwani⁴, Emma Peasley⁵, Emily Gilmore², Jennifer Kim², Sacit Bulent Omay⁶, Gordon Sze¹, Sam Payabvash¹, Annabel Sorby-Adams⁷, Juan Iglesias⁷, Matthew Rosen⁷, William Taylor Kimberly⁷, Kevin Sheth⁸
¹Yale School of Medicine, ²Yale University School of Medicine, ³Harvard Medical School, ⁴Yale University School of Medicine, Department of Neurology, ⁵Yale University, ⁶Yale Univ, ⁷Harvard University, ⁸Yale University Division of Neuro and Critical Care

Objective:

We deployed a low-field (0.064 T) magnetic resonance imaging (LF-MRI) device in patients with traumatic brain injury (TBI) and investigated the agreement between traumatic intraparenchymal hematoma (IPH) volumes demonstrated on LF-MRI and conventional neuroimaging (non-contrast computed tomography (CT) or 1.5/3T MRI).

Background:

CT is the gold standard neuroimaging modality for diagnosing and monitoring patients with TBI. TBI patients require frequent neurological assessments to inform care management in the acute setting of brain insult and after interventions such as decompressive craniectomy. The advent of LF-MRI allows for frequent, accessible neuromonitoring of TBI patients in these settings.

Design/Methods:

LF-MRI examinations were performed on patients with traumatic IPH in the neuroscience/surgical intensive care units (ICU) and emergency department (ED) from August 2020-April 2023. Fluid-attenuated inversion recovery (FLAIR) sequences were obtained at the bedside using a portable, LF-MRI device (Hyperfine, Inc). 15 exams were completed within 72 hours of conventional imaging, while one exam was completed 92 hours after and one exam was completed nine days after conventional imaging. Lesion volumes from each patient’s LF-MRI exam and closest conventional imaging study (CT or 1.5/3T MRI) were manually segmented using 3D Slicer (v5.4.0). Intraclass correlation coefficient was calculated to assess agreement between lesion volumes on LF-MRI and conventional imaging.

Results:

We examined 17 patients (ages 20-94, 19% female) with traumatic IPH. Of these, three patients received one neurosurgical intervention (mini craniotomy, craniectomy, and decompressive hemicraniectomy) and one patient received two neurosurgical interventions (burr hole and craniectomy) before their LF-MRI scan. The median [range] GCS at the time of the LF-MRI scan was 14 [3-15]. Lesion volumes manually segmented on LF-MRI exams strongly correlated with conventional imaging volumes (ICC = 0.981, 95% CI: [0.948-0.993], p<0.001).

Conclusions:

Portable, low-field MRI is a safe and practical neuroimaging tool that may complement existing conventional modalities and support management of critically ill TBI patients.