A Novel Automated CT Biomarker to Predict Outcomes in Acute Ischemic Stroke: Net Water Uptake

Monica Mallavarapu¹Hyun Woo Kim²Ananya Iyyangar¹Sergio Salazar-Marioni¹Albert Yoo²Luca Giancardo¹Sunil Sheth¹Jerome Jeevarajan^1*

• ¹University of Texas Health Science Center at Houston, Houston, United States
• ²Texas Stroke Institute, Plano, United States

Background: Recent trials of large core thrombectomy have shown that our traditional understanding of infarct characteristics and reperfusion benefit may be incomplete for patients with acute ischemic stroke (AIS). The Alberta Stroke Program Early CT Score (ASPECTS) has wide inter-rater variability, and modern studies have also shown that reperfusion therapies can benefit some patients regardless of the ASPECTS. Reproducible imaging metrics that account for the degree of hypo-attenuation on non-contrast computed tomography (NCCT) may be better suited to guide treatments. Here, we evaluate Net Water Uptake (NWU), a novel NCCT metric that can be calculated in a rapid and automated fashion, to determine its predictive performance for identifying clinical outcomes in patients with AIS compared to ASPECTS. Methods: From our prospectively collected registry encompassing eleven certified stroke centers, we identified patients with AIS. CT images were pre-processed and segmented, then NWU was calculated by automated comparison of density on ipsilateral and contralateral brain regions.Primary outcome was the area under the receiver operating characteristic curve (AUROC) for competing multivariable regression models with Average NWU versus ASPECTS to predict 90day outcome measured by modified Rankin Scale (mRS). Regression models were adjusted for age, National Institutes of Health Stroke Scale (NIHSS), tPA administration, and endovascular therapy. Secondary analyses included subgroup comparisons of patients with large infarct core and late time window. Results: Among 402 subjects with anterior circulation AIS, median age was 69 [IQR 57-80], 49.3% were female, median NIHSS was 11 [IQR 5-19], median ASPECTS was 9 [IQR 7-10], and median 90-day mRS was 3 [IQR 1-5]. The ASPECTS-based model performance was not significantly different from the NWU-based model to classify 90-day mRS outcome, with AUROC 0.732 and 0.749 respectively (p=0.513 with Delong test). Among the subgroups, performance was again similar, including patients with large infarct core (AUROC 0.795 vs 0.863, p=0.312) and late time window (AUROC 0.638 vs 0.677, p=0.267). Conclusion: NWU is a quantitative metric that can be rapidly and automatically obtained from non-contrast CT with comparable performance to ASPECTS when predicting 90-day functional outcome across a wide range of AIS presentations.