Analysis of the utility of early MRI/MRA in 400 patients with spontaneous intracerebral hemorrhage

Objective: MRI and MRA studies are routinely obtained to identify the etiology of intracerebral hemorrhage (ICH). The diagnostic yield of MRI/MRA in the setting of an acute ICH, however, remains unclear. The authors' goal was to determine the utility of early MRI/MRA in detecting underlying structural lesions in ICH and to identify patients in whom additional imaging during hospitalization could safely be foregone. Methods: The authors reviewed data obtained in 400 patients with spontaneous ICH diagnosed on noncontrast head CT scans who underwent MRI/MRA between 2015 and 2017 at their institution. MRI/MRA studies were reviewed to identify underlying lesions, such as arteriovenous malformations, aneurysms, cavernous malformations, arteriovenous fistulas, tumors, sinus thrombosis, moyamoya disease, and abscesses. Results: The median patient age was 65 ± 15.8 years. Hypertension was the most common (72%) comorbidity. Structural abnormalities were detected on MRI/MRA in 12.5% of patients. Structural lesions were seen in 5.7% of patients with basal ganglia/thalamic ICH, 14.1% of those with lobar ICH, 20.4% of those with cerebellar ICH, and 27.8% of those with brainstem ICH. Notably, the diagnostic yield of MRI/MRA was 0% in patients > 65 years with a basal ganglia/thalamic hemorrhage and 0% in those > 85 years with any ICH location, whereas it was 37% in patients < 50 years and 23% in those < 65 years. Multivariate analysis showed that decreasing age, absence of hypertension, and non-basal ganglia/thalamic location were predictors of finding an underlying lesion. Conclusions: The yield of MRI/MRA in ICH is highly variable, depending on patient age and hemorrhage location. The findings of this study do not support obtaining early MRI/MRA studies in patients ≥ 65 years with basal ganglia/ thalamic ICH or in any ICH patients ≥ 85 years. In all other situations, early MRI/MRA remains valuable in ruling out underlying lesions.