Brain Diffusion Imaging Findings May Predict Clinical Outcome after Cardiac Arrest

Background and purpose: We aim to correlate the patterns of brain diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) map in post cardiac arrest (PCA) patients with clinical outcomes.

Methods: Thirty-eight adult patients with PCA (mean age, 52.8 years; range 18-87 years) whose DWI obtained within 5 days of PCA were retrospectively reviewed. The visual DWI/ADC map categories include: Group 1: Normal; Group 2a: Mild [restricted diffusion (RD) < 1/3 cortical involvement (CI)]); Group 2b: Moderate (RD 1/3 > and < 2/3 CI); Group 2c: Severe (RD > 2/3 CI); and Group 3: Embolic (scattered, discrete foci of RD). Clinical outcomes were categorized according to cerebral performance categories (CPC) and modified Rankin scale (mRS).

Results: The most common DWI/ADC map pattern was Group 1 (28.9%, n = 11). The incidence of other DWI patterns such as Group 2a, 2b, 2c, and 3 were 21% (n = 8), 10.5% (n = 4), 21% (n = 8), and 18.4% (n = 7), respectively. Twenty-seven patients (71%) were CPC-5/mRS-6 and died or were category CPC-4/mRS-5, and 4 patients were CPC-1/mRS 0-1 (10.5%). Interobserver agreement for visual classification of DWI/ADC map patterns was excellent (kappa = .8795). There was moderate positive correlation between clinical outcomes and visual DWI classification (r = .461, P = .00358). The positive predictive value of this qualitative classification on DWI/ADC in predicting a poor clinical outcome (CPC-4/mRS-5 and CPC-5/mRS-6) was 81.4 % in the presence of restricted diffusion.

Conclusion: Simple visual categorization system using DWI/ADC map may be helpful and practical in estimating the clinical outcome of PCA patients.