Abstract 296: Comparison of Prognostic Performance Using Quantitative Analysis of Head Computed Tomography and Apparent Diffusion Coefficient Imaging Before Target Temperature Management in Out-of-Hospital Cardiac Arrest Survivors

Abstract

Aim: We compared head computed tomography (HCT) and diffusion-weighted magnetic resonance imaging (DW-MRI) findings before target temperature management (TTM) to predict neurological outcomes in out-of-hospital cardiac arrest survivors. Methods: Using prospectively collected data from a TTM registry, we included patients who underwent HCT and DW-MRI within 6 h of return of spontaneous circulation (ROSC). We examined the gray-to-white matter ratio (GWR) on HCT and apparent diffusion coefficient (ADC) values obtained through voxel-based analysis of DW-MRI. We analyzed the mean ADC value and the % voxels of ADC thresholds (percentage of voxels below ADC thresholds per total voxel) at 50-step intervals ranging from 200 to 1200 х 10 -6 mm/s, identifying thresholds that differentiate between good and poor neurological outcomes. Additionally, we analyzed the combination of pupillary light reflex (PLR) and neuron-specific enolase (NSE) values acquired at the same time points to predict the neurological outcomes. The primary outcome measure was the dichotomized cerebral performance category (CPC) at six months, defined as good (CPC 1-2) or poor (CPC 3-5). Results: Of 131 patients (26% female), 74 (57%) had poor outcomes. The time from ROSC to HCT and DW-MRI scans was 1.7h and 3.2h, respectively. The group with a good outcome had higher GWR (1.26 vs. 1.20) and mean ADC (787.10 vs. 697.49) values, but lower values in all ranges (250-1150) of ADC values (all P<0.001). Mean ADC and ADC values in the range of 350-650 demonstrated better predictive performance for poor outcomes than the GWR (P < 0.05). The mean ADC value had the highest sensitivity of 51.3% (95% CI 39.4-63.1; cutoff value ≤ 739.2 х 10 –6 mm 2 /s) when the false positive rate (FPR) was 0%. Furthermore, when combined with PLR and NSE values, the prognostic performance was significantly improved compared to using Mean ADC alone (0.83 vs 0.92; P<0.01). Conclusions: In this cohort study, early voxel-based quantitative ADC analysis after ROSC (i.e., before TTM) was associated with poor neurological outcomes 6 months after cardiac arrest. We found that the Mean ADC value showed the highest sensitivity when the FPR was 0% and that combining the NSE and PLR values resulted in an even better prognostic performance.