Rapid MRI evaluation of acute intracranial hemorrhage in pediatric head trauma

Rapid MRI with ultrafast T2 sequences can be performed without sedation and is often used in place of computed tomography (CT) to evaluate pediatric patients for indications such as hydrocephalus. This study investigated the sensitivity of rapid magnetic resonance imaging (MRI) for detection and follow-up of acute intracranial hemorrhage in comparison to CT, which is commonly the first-line imaging.

Patients presenting to a pediatric hospital with acute intracranial hemorrhage on CT and follow-up rapid MRI within 48 h were included. Rapid MRI studies consisted of three plane ultrafast T2 sequences either with or without axial gradient echo (GRE) sequences. Identification of hemorrhage on rapid MRI was assessed by readers both blinded and unblinded to prior CT results.

One hundred two acute hemorrhages in 61 patients were identified by CT. Rapid MRI detection of subdural and epidural hemorrhages was modest in the absence of prior CT for comparison (sensitivity 61-74 %), but increased with review of the prior CT (sensitivity 80-86 %). Hemorrhage size was a significant predictor of detection (p < 0.0001). Three plane fast T2 images alone without GRE sequences were poor at detecting subarachnoid hemorrhage (sensitivity 10-25 %); rapid MRI with GRE sequences identified the majority of subarachnoid hemorrhage (sensitivity 71-93 %). GRE modestly increased detection of other extra-axial hemorrhages.

Rapid MRI with GRE sequences is sensitive for most acute intracranial hemorrhages only when a prior CT is available for review. Rapid MRI is not adequate to replace CT in initial evaluation of intracranial hemorrhages but may be helpful in follow-up of known hemorrhages.