A 78 Seconds Complete Brain MRI Examination in Ischemic Stroke: A Prospective Cohort Study

• Published in: Journal of magnetic resonance imaging Vol. 56; no. 3; pp. 884 - 892
• Main Authors: af Burén, Siri, Kits, Annika, Lönn, Lucas, De Luca, Francesca, Sprenger, Tim, Skare, Stefan, Falk Delgado, Anna
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.09.2022; Wiley Subscription Services, Inc
• Subjects: brain infarction; Cohort analysis; diagnostic performance; Diagnostic systems; EPIMix; fast imaging; Field strength; Ischemia; ischemic stroke; Lesions; Magnetic resonance imaging; Maximum likelihood estimation; Medicin och hälsovetenskap; multicontrast MRI; Patients; Performance evaluation; Population studies; Sensitivity analysis; Statistical analysis; Statistical tests; Stroke; diagnostic performance; brain infarction; multicontrast MRI; EPIMix; ischemic stroke; fast imaging
• ISSN: 1053-1807; 1522-2586; 1522-2586
• DOI: 10.1002/jmri.28107

Background Fast 78‐second multicontrast echo‐planar MRI (EPIMix) has shown good diagnostic performance for detecting infarctions at a comprehensive stroke center, but its diagnostic performance has not been evaluated in a prospective study at a primary stroke center. Purpose To prospectively determine whether EPIMix was noninferior in detecting ischemic lesions compared to routine clinical MRI. Study type Prospective cohort study. Population A total of 118 patients with acute MRI and symptoms of ischemic stroke. Field Strength and Sequence A 3 T. EPIMix (echo‐planar based: T1‐FLAIR, T2‐weighted, T2‐FLAIR, T2*, DWI) and routine clinical MRI sequences (T1‐weighted fast spin echo, T2‐weighted PROPELLER, T2‐weighted‐FLAIR fast spin echo, T2* gradient echo echo‐planar, and DWI spin echo echo‐planar). Assessment Three radiologists, blinded for clinical information, assessed signs of ischemic lesions (DWI↑, ADC↓, and T2/T2‐FLAIR↑) on EPIMix and routine clinical MRI, with disagreements solved in consensus with a fourth reader to establish the reference standard. Statistical tests Diagnostic performance including sensitivity and specificity against the reference standard was evaluated. EPIMix sensitivity was tested for noninferiority compared to the reference standard using Nam's restricted maximum likelihood estimation (RMLE) Score. A P‐value < 0.05 was considered statistically significant. Results Of 118 patients (mean age 62 ± 16 years, 58% males), 25% (n = 30) had MRI signs of acute infarcts. EPIMix was noninferior with 97% (95% CI 83–100) sensitivity for reader 1, 100% (95% CI 88–100) sensitivity for reader 2, and 90% (95% CI 88–98) sensitivity for reader 3 vs. 93% (95% CI 78–99) sensitivity for readers 1 and 2 and 90% (95% CI 74–98) for reader 3 on routine clinical MRI. Specificity was 99% (95% CI 94–100) for reader 1, 100% (95% CI 96–100) for reader 2, and 98% (95% CI 92–100) for reader 3 on EPIMix vs. 100% (95% CI 96–100) for all readers on routine clinical MRI. Conclusion EPIMix was noninferior to routine clinical MRI for the diagnosis of acute ischemic stroke. Evidence Level 2 Technical Efficacy Stage 2