Detection of Global Brain Injury Using Point-of-Care Neonatal MRI Scanner

Abstract

Abstract Background Conventional magnetic resonance imaging (MRI) neuroimaging of infants is complicated by the need to transport infants outside the neonatal intensive care unit (NICU), often to distant areas of the hospital. Primary Objective The main aim of this study was to evaluate and compare scoring of images from a novel 1T MRI, which enables neuroimaging within the NICU, with those from a conventional MRI. Secondary Objective The second aim of this study was to document improved expediency, and thereby greater patient safety, as reflected by decreased transport time. Materials and Methods Thirty premature infants (mean gestational age: 28.8 ± 2.1 weeks) were scanned consecutively on the novel 1T and 1.5T conventional scanners at term-equivalent age. Orthogonal T1- and T2-weighted images were acquired and reviewed. A global brain abnormality score (Kidokoro) was assigned independently to all images by two radiologists. Interrater agreement was evaluated using the kappa statistic and interscanner agreement was evaluated by Bland–Altman analysis. Transport time to and from both scanners was monitored and compared. Results Weighted kappas were 0.77 (standard error of measurement [SEM] 0.08; confidence interval [CI]: 0.62–0.92) and 0.86 (SEM: 0.07; CI: 0.73–1), for the 1T and 1.5T scanners, respectively, reflecting substantial interrater agreement. Bland–Altman analysis showed excellent agreement between the two scanners.Transport time was 8 ± 6 minutes for the 1T MRI versus 46 ± 21 minutes for the conventional MRI (p < 0.00001). No adverse events were recorded during transport. Standard transport times will vary from institution to institution. Conclusions Kidokoro scores are similar when comparing images obtained from a 1T MRI with those of a conventional 1.5T MRI, reflecting comparable image quality. Transport time was significantly decreased using the 1T neonatal MRI.