The prognostic value of quantitative diffusion-weighted MRI after pediatric cardiopulmonary arrest

Objectives: The prognostic value of quantitative diffusion-weighted magnetic resonance imaging (DWI MRI) in predicting neurologic outcomes after pediatric cardiopulmonary arrest (CPA) has not been determined. The aim of this study was to identify a DWI MRI threshold for brain volume percent that correlates with neurologic outcome in children who remain comatose or display significant neurologic deficits immediately after resuscitation from CPA. Methods: This single-center retrospective study analyzed DWI MRIs of pediatric patients who remained neurologically impaired after CPA. Any MRI obtained within 2 weeks after CPA was analyzed. The apparent diffusion coefficient (ADC) value of each voxel within the brain was determined. Percentage brain volume with voxels below each ADC threshold between 300 and 1200 x 10(-6) mm(2)/s with a step of 50 were calculated. Area under the receiver operating characteristics curve (AUC) was used to identify optimal DWI MRI thresholds for brain volume percent most predictive of poor neurologic outcome. The primary outcome measure was neurologic outcome 6-months after CPA based on Pediatric Cerebral Performance Category (PCPC) score. Poor neurologic outcome was defined as PCPC score of 3-6, or a worsening from baseline score >= 1 if baseline PCPC score was >= 3. Results: Twenty-six patients were included in this study. The median age was 8.5 years (2.2-14) and median time from CPA to MRI was 4 days (2-7). Two ADC thresholds for brain volume percent had the largest AUC for predicting poor neurologic outcome. An ADC threshold of <600 x 10(-6) mm(2)/s in >= 7% of brain volume; and <650 x 10(-6)mm(2)/s in >= 11% of brain volume both demonstrated a specificity of 1.0 (0.76-1.0, 95% CI) and a sensitivity of 0.8 (0.44-0.96, 95% CI) for poor outcome. Conclusions: In pediatric patients who remain comatose or have significant neurologic deficits after CPA, quantitative DWI MRI correlates with neurologic outcome. Both an ADC threshold of < 600 x 10(-6) mm(2)/s in >= 7% of brain volume and < 650 x 10(-6) mm(2)/s in >= 11% of brain volume are highly specific for predicting poor neurologic outcome. A prospective trial to validate these thresholds is needed.