• Wang Pong Chan, Christine Nguyen, Noah Kim, Yorghos Tripodis, UEmily J Gilmore, David M Greer, and Rachel Beekman.
• Department of Neurology, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, USA; Department of Biostatistics, Boston University School of Public Health, Boston, MA, USA. Electronic address: wpchan@bu.edu.
• Resuscitation. 2024 Aug 21: 110370110370.

AimMagnetic Resonance Imaging (MRI) is an important prognostic tool in cardiac arrest (CA) survivors given its sensitivity for detecting hypoxic-ischemic brain injury (HIBI), however, it is limited by poorly defined objective thresholds. To address this limitation, we evaluated a qualitative MRI score for predicting neurological outcome in CA survivors.MethodsAdult comatose CA survivors who underwent MRI were retrospectively identified at a single academic medical center. Two blinded neurointensivists qualitatively scored HIBI amongst 12 MRI brain regions. Scores were summated to form four distinct score groups: cortex, deep grey nuclei (DGN), cortex-DGN combined, and total (cortex, DGN, brainstem, and cerebellum). Poor neurological outcome was defined as Cerebral Performance Category (CPC) score 3-5 at hospital discharge. Inter-rater reliability was tested using intra-class correlation (ICC) and discrimination of poor neurological outcome assessed using area under the receiver operating curve (AUC).ResultsOur cohort included 219 patients with median time to MRI of 96 (IQR 81-110) hours. ICC (95% CI) was good to excellent across all MRI scores: cortex 0.92 (0.89-0.94), DGN 0.88 (0.80-0.92), cortex-DGN 0.94 (0.92-0.95), and total 0.93 (0.91-0.95). AUC (95% CI) for poor outcome was good across all MRI scores: cortex 0.84 (0.78-0.90), DGN 0.83 (0.77-0.89), cortex-DGN 0.83 (0.77-0.89), and total 0.83 (0.77-0.88).ConclusionA simplified, qualitative MRI score had excellent reliability and good discrimination for poor neurologic outcome. Further work is necessary to externally validate our findings in an independent, ideally prospective, cohort.Copyright © 2024 Elsevier B.V. All rights reserved.

4 keywords...

hide…