Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

The semi-quantitative cardiac arrest brain ischemia (CABI) score for magnetic resonance imaging predicts functional outcome after cardiac arrest

Arctaedius, Isabelle LU ; Wassélius, Johan LU ; Lang, Margareta LU ; Drake, Mattias LU ; Johnsson, Mikael ; Friberg, Hans LU ; Leithner, Christoph ; Kenda, Martin ; Lybeck, Anna LU orcid and Moseby-Knappe, Marion LU (2025) In Critical Care 29(1).
Abstract

Background: Magnetic resonance imaging (MRI) is recommended by guidelines to evaluate the severity of brain ischemia after cardiac arrest for functional outcome prediction. However, standardized assessment criteria are lacking. We have developed a semi-quantitative Cardiac Arrest Brain Ischemia (CABI) score to assess the extension of brain ischemia on MRI. Objectives: To evaluate the prognostic performance of a novel semi-quantitative CABI score in comparison to qualitative routine radiological MRI assessment after cardiac arrest. Methods: A retrospective, multicenter observational study of adults (≥ 18 years-old) admitted to intensive care after cardiac arrest at four Swedish hospitals from 2014 to 2018. Four radiologists, blinded to... (More)

Background: Magnetic resonance imaging (MRI) is recommended by guidelines to evaluate the severity of brain ischemia after cardiac arrest for functional outcome prediction. However, standardized assessment criteria are lacking. We have developed a semi-quantitative Cardiac Arrest Brain Ischemia (CABI) score to assess the extension of brain ischemia on MRI. Objectives: To evaluate the prognostic performance of a novel semi-quantitative CABI score in comparison to qualitative routine radiological MRI assessment after cardiac arrest. Methods: A retrospective, multicenter observational study of adults (≥ 18 years-old) admitted to intensive care after cardiac arrest at four Swedish hospitals from 2014 to 2018. Four radiologists, blinded to clinical information except patient age, assessed the MRI as per guideline recommendations and according to the CABI score. The CABI score evaluates extension of ischemia in 12 predefined territories (3 cortical and 2 deep grey nuclei territories bilaterally based on vascular supply; the cerebellum; and the brainstem) from 0–3 (0 = No acute ischemic lesion; 3 ≥ 50% of the territory). The CABI score ranges from 0–36. Poor functional outcome was defined as Cerebral Performance Category 3–5 assessed at 2–6 months post-arrest. Prognostic ability was evaluated with sensitivity and specificity and with the area under the receiver operating characteristics curve (AUC). Inter-rater reliability was calculated by Fleiss´ Kappa or intraclass correlation coefficient. Results: Ninety (n = 90) patients examined with MRI at median 5 days (IQR: 3.7−6.3) were included Median age was 63.7 years (IQR: 58.3−70.5), 74.4% (n = 67) were men and 85.6% (n = 77) had a poor functional outcome. Among four blinded raters, routine qualitative assessment according to guidelines showed 84.6–100% specificity and 61–76.6% sensitivity for poor outcome. Inter-rater reliability for the qualitative assessment was moderate (Fleiss κ = 0.64, 95% CI: 0.53−0.75). The CABI score achieved an AUC of 0.87–0.92 (p< 0.05 vs. routine assessment) and a sensitivity of 59.7−85.7% while maintaining high specificity (76.9–100%) at cut-off ≥ 8 points. The intraclass correlation coefficient (ICC) for the CABI score was 0.67 (95% CI: 0.58–0.75), reflecting moderate agreement among raters. Conclusion: A semi-quantitative approach to MRI evaluations after cardiac arrest may increase the prognostic accuracy compared to non-standardized routine assessment. Trial registration: SweCrit biobank, retrospectively registered at ClinicalTrials.gov no. NCT04974775 18th of June 2021.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Cardiac arrest, MRI, Neuroimaging, Prognostication
in
Critical Care
volume
29
issue
1
article number
373
publisher
BioMed Central (BMC)
external identifiers
  • pmid:40836345
  • scopus:105013641141
ISSN
1364-8535
DOI
10.1186/s13054-025-05595-1
language
English
LU publication?
yes
id
aa13f77e-a516-4e00-88f0-af1527e982e2
date added to LUP
2025-10-03 12:54:34
date last changed
2025-10-03 12:55:30
@article{aa13f77e-a516-4e00-88f0-af1527e982e2,
  abstract     = {{<p>Background: Magnetic resonance imaging (MRI) is recommended by guidelines to evaluate the severity of brain ischemia after cardiac arrest for functional outcome prediction. However, standardized assessment criteria are lacking. We have developed a semi-quantitative Cardiac Arrest Brain Ischemia (CABI) score to assess the extension of brain ischemia on MRI. Objectives: To evaluate the prognostic performance of a novel semi-quantitative CABI score in comparison to qualitative routine radiological MRI assessment after cardiac arrest. Methods: A retrospective, multicenter observational study of adults (≥ 18 years-old) admitted to intensive care after cardiac arrest at four Swedish hospitals from 2014 to 2018. Four radiologists, blinded to clinical information except patient age, assessed the MRI as per guideline recommendations and according to the CABI score. The CABI score evaluates extension of ischemia in 12 predefined territories (3 cortical and 2 deep grey nuclei territories bilaterally based on vascular supply; the cerebellum; and the brainstem) from 0–3 (0 = No acute ischemic lesion; 3 ≥ 50% of the territory). The CABI score ranges from 0–36. Poor functional outcome was defined as Cerebral Performance Category 3–5 assessed at 2–6 months post-arrest. Prognostic ability was evaluated with sensitivity and specificity and with the area under the receiver operating characteristics curve (AUC). Inter-rater reliability was calculated by Fleiss´ Kappa or intraclass correlation coefficient. Results: Ninety (n = 90) patients examined with MRI at median 5 days (IQR: 3.7−6.3) were included Median age was 63.7 years (IQR: 58.3−70.5), 74.4% (n = 67) were men and 85.6% (n = 77) had a poor functional outcome. Among four blinded raters, routine qualitative assessment according to guidelines showed 84.6–100% specificity and 61–76.6% sensitivity for poor outcome. Inter-rater reliability for the qualitative assessment was moderate (Fleiss κ = 0.64, 95% CI: 0.53−0.75). The CABI score achieved an AUC of 0.87–0.92 (p&lt; 0.05 vs. routine assessment) and a sensitivity of 59.7−85.7% while maintaining high specificity (76.9–100%) at cut-off ≥ 8 points. The intraclass correlation coefficient (ICC) for the CABI score was 0.67 (95% CI: 0.58–0.75), reflecting moderate agreement among raters. Conclusion: A semi-quantitative approach to MRI evaluations after cardiac arrest may increase the prognostic accuracy compared to non-standardized routine assessment. Trial registration: SweCrit biobank, retrospectively registered at ClinicalTrials.gov no. NCT04974775 18th of June 2021.</p>}},
  author       = {{Arctaedius, Isabelle and Wassélius, Johan and Lang, Margareta and Drake, Mattias and Johnsson, Mikael and Friberg, Hans and Leithner, Christoph and Kenda, Martin and Lybeck, Anna and Moseby-Knappe, Marion}},
  issn         = {{1364-8535}},
  keywords     = {{Cardiac arrest; MRI; Neuroimaging; Prognostication}},
  language     = {{eng}},
  number       = {{1}},
  publisher    = {{BioMed Central (BMC)}},
  series       = {{Critical Care}},
  title        = {{The semi-quantitative cardiac arrest brain ischemia (CABI) score for magnetic resonance imaging predicts functional outcome after cardiac arrest}},
  url          = {{http://dx.doi.org/10.1186/s13054-025-05595-1}},
  doi          = {{10.1186/s13054-025-05595-1}},
  volume       = {{29}},
  year         = {{2025}},
}