Correlation Between Computed Tomography-Based Tissue Net Water Uptake and Volumetric Measures of Cerebral Edema After Reperfusion Therapy

Background: Cerebral edema after large hemispheric infarction is associated with poor functional outcome and mortality. Net water uptake (NWU) quantifies the degree of hypoattenuation on unenhanced-computed tomography (CT) and is increasingly used to measure cerebral edema in stroke research. Hemorrhagic transformation and parenchymal contrast staining after thrombectomy may confound NWU measurements. We investigated the correlation of NWU measured postthrombectomy with volumetric markers of cerebral edema and association with functional outcomes. Methods: In a pooled individual patient level analysis of patients presenting with anterior circulation large hemispheric infarction (core 80–300 mL or Alberta Stroke Program Early CT Score ≤5) in the HERMES (Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke trials) data set, cerebral edema was defined as the volumetric expansion of the ischemic hemisphere expressed as a ratio to the contralateral hemisphere(rHV). NWU and midline-shift were compared with rHV as the reference standard on 24-hour follow-up CT, adjusted for hemorrhagic transformation and the use of thrombectomy. Association between edema markers and day 90 functional outcomes (modified Rankin Scale) was assessed using ordinal logistic regression. Results: Overall (n=144), there was no correlation between NWU and rHV (r s =0.055, P =0.51). In sub-group analyses, a weak correlation between NWU with rHV was observed after excluding patients with any degree of hemorrhagic transformation (r s =0.211, P =0.015), which further improved after excluding thrombectomy patients (r s =0.453, P =0.001). Midline-shift correlated strongly with rHV in all sub-group analyses (r s >0.753, P =0.001). Functional outcome at 90 days was negatively associated with rHV (adjusted common odds ratio, 0.46 [95% CI, 0.32–0.65]; P <0.001) and midline-shift (adjusted common odds ratio, 0.85 [95% CI, 0.78–0.92]; P <0.001) but not NWU (adjusted common odds ratio, 1.00 [95% CI, 0.97–1.03]; P =0.84), adjusted for age, baseline National Institutes of Health Stroke Scale, and thrombectomy. Prognostic performance of NWU improved after excluding patients with hemorrhagic transformation and thrombectomy (adjusted odds ratio, 0.90 [95% CI, 0.80–1.02]; P =0.10). Conclusions: NWU correlated poorly with conventional markers of cerebral edema and was not associated with clinical outcome in the presence of hemorrhagic transformation and thrombectomy. Measuring NWU postthrombectomy requires validation before implementation into clinical research. At present, the use of NWU should be limited to baseline CT, or follow-up CT only in patients without hemorrhagic transformation or treatment with thrombectomy.