Introduction & Hypothesis: Diffusion-weighted imaging (DWI), a technique sensitive to acute brain ischemia, may hold utility in predicting neurological outcome in comatose post-cardiac arrest patients. Outcome studies are biased by early withdrawal of life sustaining therapy (WLST), leading to a self-fulfilling prophecy. This creates a need to reassess the use of DWI as a neuroprognostic tool in patients who have not undergone WLST. We hypothesize that DWI abnormalities is a robust predictor of poor neurological outcome in our WLST - controlled cohort. Methods: We leveraged the MOCHA database, a registry of over 300 comatose post-cardiac arrest patients, to retrospectively examine neurological outcomes in a cohort of patients who did not undergo WLST. A good outcome was defined by a cerebral performance category (CPC) score at discharge of 1 - 3, while a poor outcome CPC 4 - 5 (n=43). We first examined the relationship between the number of brain regions with DWI abnormalities and CPC score using a linear regression. We then examined how DWI abnormalities in specific brain regions correlated with CPC score outcome groups using a fisher exact test. DWI abnormalities were qualitatively determined by two vascular neurologists. Results: We found a positive correlation between the number of brain regions with DWI abnormalities and CPC score ( linear regression , R 2 =0.572, p=2.670x10 -9 ). Interestingly, the association between DWI abnormalities and CPC score exhibited brain region-specific variability. DWI abnormalities exhibited the strongest association with poor neurological outcome in the occipital lobe ( fisher exact test , p=7.413x10 -10 ), parietal lobe (p=9.125x10 -9 ), frontal lobe (p=5.385x10 -9 ), temporal lobe (p=3.904x10 -8 ) and basal ganglia (p=2.342x10 -7 ); and the weakest association in the white matter (p=1.000) and brain stem (p=6.612x10 -2 ). Conclusion: Our preliminary results suggest that the region of ischemia is an important factor to consider in predicting neurological outcome. This warrants a larger scale WLST-controlled study examining region-specific DWI abnormalities and neurological outcome - the findings of which would improve our neuroprognostication capabilities in comatose post-cardiac arrest patients.

Introduction: Brain death occurs in 10-15% of successfully resuscitated out-of-hospital cardiac arrest (OHCA) patients. Early identification of potential organ donors is critical to prevent withdrawal of life sustaining therapy (WLST) and ensure adequate organ perfusion. To predict brain death after OHCA, we developed a novel brain death risk (BDR) score. Methods: The BDR score was developed from a retrospective, single center cohort of OHCA patients admitted from 2011-2020. After excluding patients with early WLST (defined as < 72 hours from OHCA), univariate regression models identified independent predictors of brain death, which were used to build the BDR score. We included the following variables: non-shockable rhythm (1 point), drug overdose as etiology of arrest (1 point), evidence of grey-white differentiation loss or sulcal effacement on head computed tomography (CT) within 24 hours of arrest (2 points), Full-Outline-Of-UnResponsiveness (FOUR) score of 0 (2 points), FOUR score 1-5 (1 point), and age < 45 years (1 point). Head CT findings were based on neuroradiology reports. We validated the BDR score in an independent single center OHCA cohort. The primary outcome was occurrence of brain death. Using the area under the receiving operator characteristic curve (AUC), we assessed the score’s prediction of brain death. Results: The development cohort included 256 OHCA patients; 15.6% (40/256) experienced brain death. The AUC (95% CI) of the BDR score was 0.921 (0.870-0.971). In the validation cohort, 24.4% (21/86) experienced brain death. The AUC (95% CI) of the BDR score was 0.830 (0.7266-0.9335). Table 1 shows the rate of brain death at each BDR score. In both cohorts, a BDR score ≥ 5 was the optimal cut off (sensitivity 0.9 and 0.714, specificity 0.843 and 0.831, respectively). Conclusion: Early scoring systems may be able to identify those at highest risk for brain death after OHCA. Our data suggest that a BDR score ≥ 5 could help predict progression to brain death.

Introduction: Previous studies have shown that lower gray:white matter ratio (GWR) is associated with poor outcomes in patients initially comatose after cardiac arrest. However, those studies focused on survival, a measure that may be confounded by withdrawal of life-sustaining therapy (WLST) decisions. We assessed the association of GWR with severe cerebral edema (SCE), a more objective imaging surrogate for neurologic injury. Methods: We retrospectively analyzed datasets acquired as part of a prospective observational study of 500 cardiac arrest patients with non-traumatic coma. Our analyses were limited to patients who underwent CT within 72 h of arrest, resulting in 151 patients. Some patients underwent repeat imaging, leading to a total of 175 imaging studies. Regions of interests (ROIs; 3x3) were drawn bilaterally in the caudate nuclei (CN) and posterior limbs of the internal capsule (PLIC) at the level of the basal ganglia on non-contrast axial CT scans. GWR was calculated as the ratio of mean x-ray attenuation in the CN to that in the PLIC. Scans whose radiology report described herniation or more than minimal ventricular effacement were classified as SCE. Poor outcome was defined as modified Rankin Scale Score > 4 at 6 months. Results: Patients with SCE were more likely to be younger (P<.0001), have a non-shockable rhythm (P=0.028), have a worse Glasgow Coma Scale (GCS) score on hospital admission (P=0.0047) and have a significantly lower GWR (P<.0001) compared to patients without SCE (see Table 1). Patients with SCE had a higher proportion of deaths at discharge due to brain death (P<0.0001). There was no significant difference in the proportion of patients with poor outcomes between the groups (P=0.13). Conclusion: For patients initially comatose after cardiac arrest, significant reductions in GWR are associated with SCE. Since this study utilized SCE as intermediate outcome measure for neurologic injury, these GWR results are not confounded by WLST decisions.