Feasibility and Clinical Utility of Low-Field Magnetic Resonance Imaging in Critically Ill Children: An Experience from Pakistan

Background: Low-field portable magnetic resonance imaging (pMRI) systems have been approved for clinical use, but their feasibility, efficacy, and most appropriate clinical application in children are unknown. The objective of this study was to evaluate the diagnostic accuracy of pMRI in detecting acute brain injury (ABI) in critically ill children compared to conventional MRI (cMRI) and or computed tomography (CT). Methods: This prospective diagnostic accuracy study included children (1 month to < 18 years) admitted between May 2021 and June 2022 who underwent pMRI scans within a 24-h window of standard neuroimaging (CT or cMRI). pMRI images were assessed for quality of images and interpreted for findings by two independent pediatric radiologists blinded to the findings of the standard imaging modalities. Diagnostic accuracy was assessed using sensitivity, specificity, and agreement statistics. Results: pMRI scans were successfully completed in 83% of the cases (73 of 88 patients) with a median scan time of 48 min (interquartile range 43–54 min), with no patient- or machine-related adverse event. Neuroimaging indications were seizures in 38 (52.1%), unexplained encephalopathy in 29 (39.7%), and focal neurologic deficits in 19 (26%) patients. Radiological findings of pMRI included edema in 28 (38.4%), hydrocephalus in 12 (16.4%), infarction in 13 (17.8%), midline shift in 12 (16.4%), and intraparenchymal hemorrhage in 7 (9.6%). The pMRI demonstrated good agreement with cMRI and CT scans for detecting edema (87% agreement, κ = 0.7), hydrocephalus (94% agreement, κ = 0.8) and intraparenchymal hemorrhage (87% agreement, κ = 0.6). Compared to cMRI, pMRI showed 72% agreement (κ = 0.41, P = 0.0002), with 98% specificity and 73% sensitivity for hydrocephalus. Overall agreement between the two observers for pMRI was 90% (κ = 0.81, P < 0.001). Image quality was adequate for T1-weighted (n = 58, 79.5%), T2-weighted (n = 61, 83.6%), diffusion-weighted imaging (DWI) (n = 49, 67.1%), and apparent diffusion coefficient (n = 48, 65.8%) sequences. The highest number of uninterpretable images were for the DWI sequence (n = 9, 12.3%). Conclusions: pMRI is a safe and feasible bedside imaging modality that shows promising results in diagnosing ABI in children.