Abstract WP439: Diagnostic Yield of MRI/MRA in 400 Patients With Spontaneous Intracerebral Hemorrhage

Abstract

Background: It is common practice to perform MRI and MRA to identify secondary causes of intracerebral hemorrhage (ICH). The utility of MRI/A in this setting has not been adequately studied, and the guidelines of the AHA/ASA on ICH do not indicate which patients require further imaging with MRI/A. Objective: To determine the yield of MRI/MRA in spontaneous ICH and stratify the results in different subgroups of patients. Methods: Patients with spontaneous ICH as diagnosed on non-contrast head CT who underwent further evaluation with an MRI/A between 2015 and 2017 at our institution were included in the study. Patients with known vascular/neoplastic lesions prior to admission were excluded. MRIs and MRAs were reviewed to identify underlying neoplastic or vascular lesions namely arteriovenous malformations, aneurysms, cavernous malformations, arteriovenous fistulas, sinus thrombosis, or moyamoya disease. Results: A total of 400 patients were identified. Mean age was 65.3±15.8 years. ICH locations were as follows: lobar (48%), basal ganglia (35.3%), cerebellar (12.3%), and pontine (4/5%). The overall diagnostic yield of MRI/A was 12.5%. In multivariate analysis, factors predicting positive findings on MRI/A were younger patients (p<0.001), nonbasal ganglia location (p=0.006), and absence of vascular risk factors (p=0.007). Importantly, the diagnostic yield of MRI/A was 0% in patients > 63 years with a basal ganglia hemorrhage, and those > 85 years with any ICH location. Conversely, the yield of MRI/A was 45% in those ≤50 years. Conclusion: Based on the results from a large cohort of patients with spontaneous ICH, we do not recommend obtaining MRI/A in patients > 85 years and those > 63 years with ICH in the basal ganglia. In all other situations, MRI/A remains indicated to rule out underlying lesions.