Role of a modified ultrafast MRI brain protocol in clinical paediatric neuroimaging

• Published in: Clinical radiology 2020-12, Vol.75 (12), p.914-920
• Main Authors: Ahamed, S.H., Lee, K.J., Tang, P.H.
• Format: Article
• Language: English
• Subjects: Adolescent; Brain Diseases - diagnostic imaging; Child; Child, Preschool; Female; Humans; Imaging, Three-Dimensional; Infant; Magnetic Resonance Imaging - methods; Male; Neuroimaging - methods; Prospective Studies
• ISSN: 0009-9260; 1365-229X
• DOI: 10.1016/j.crad.2020.07.009

To establish a role for modified ultrafast magnetic resonance imaging (MRI) of the brain in clinical paediatric patients based on clinically acceptable image quality and diagnostic accuracy. A prospective study was conducted with institutional review board approval on an ultrafast MRI brain protocol consisting of sagittal T1-weighted, axial T2-weighted, axial fluid-attenuated inversion recovery (FLAIR), axial diffusion-weighted imaging (DWI), and axial T2∗-weighted sequences. Preliminary investigations revealed that the default ultrafast T2-weighted sequence was prone to pulsation artefacts. A modified ultrafast T2-weighted sequence was therefore developed to replace the default ultrafast T2-weighted sequence. Thirty-five patients with clinical indication for neuroimaging underwent ultrafast MRI, modified ultrafast T2-weighted sequence and standard MRI at 3 T. Image quality of ultrafast MRI sequences were graded as clinically “diagnostic” or “non-diagnostic” and compared against the corresponding standard MRI sequences as the reference standard. The modified ultrafast T2-weighted sequence surpassed the default ultrafast T2-weighted sequence in image quality. The ultrafast MRI protocol was therefore replaced with the modified ultrafast T2-weighted sequence creating a modified ultrafast MRI protocol. The clinical reports of modified ultrafast MRI were compared against standard MRI for diagnostic concordance, categorised further as “normal”, “clinically significant”, or “clinically minor” abnormalities. Ultrafast T1-weighted, FLAIR, and DWI sequences had comparable image quality to standard MRI sequences. The ultrafast T2∗-weighted sequence had significantly higher non-diagnostic images (42.9%) compared to the standard MRI sequence (2.9%). The default ultrafast T2-weighted sequence had significantly higher non-diagnostic images compared to the modified ultrafast T2-weighted sequence and standard T2-weighted sequence (82.9%, 5.7%, 8.6%, respectively). There was 100% concordance for normal and clinically significant abnormalities and 23% discordance for clinically minor abnormalities. Modified ultrafast MRI takes 5 minutes 41 seconds compared to standard MRI time of 14 minutes 57 seconds. The modified ultrafast MRI protocol for brain imaging demonstrates clinically acceptable image quality in four out of five sequences and has high accuracy in diagnosing normal and clinically significant abnormalities when compared against the standard MRI protocol for brain im