Associations between cerebral magnetic resonance imaging infarct volume and acute ischemic stroke etiology

Abstract

Background Identifying ischemic stroke etiology is necessary for proper treatment and secondary prevention. We sought to define associations between infarct volume and stroke subtypes. Materials and methods Inclusion criteria necessitated a Johns Hopkins Hospital inpatient admission (2017–2019) for ischemic stroke with confirmatory brain magnetic resonance imaging. Infarct volume was calculated using MRIcron© by a masked reviewer. Ischemic strokes were adjudicated using TOAST classification. Multivariable/multinomial logistic regression determined associations between infarct volume and stroke subtypes with interaction terms for infarct number and location. Stepwise adjustment accounted for potential confounders. Results Patients (N = 150) were on average 61 years old, male (58%), and black (57%). Each 5mL increase in infarct volume was associated with cardioembolic (OR 1.07, 95%CI 1.01–1.14) and large-artery occlusions (OR 1.10, 95%CI 1.02–1.18), but lower odds of lacunar stroke (OR 0.18, 95%CI 0.06–0.55). There was no difference in risk of cardioembolic (base) and large-artery atherosclerotic strokes with increasing infarct volume (RRR 1.01, 95%CI 0.94–1.09), but risk of lacunar stroke was decreased (RRR 0.17, 95%CI 0.06–0.53). Infarct number (single vs multiple) modified the association between volume and subtype for large-artery occlusions (p-interaction 0.09). Conclusions In this study, larger volume infarcts were significantly associated with both cardioembolic and large-artery atherosclerotic strokes (no difference in the degree of association) and decreased odds of lacunar stroke. A single, large-volume stroke was associated with large-artery atherosclerosis, while multiple infarcts were associated with cardioembolism. Given the differential associations between volume, number of lesions, and stroke etiology, defining stroke subtypes in light of infarct volume might aid in clinical practice.