Arterial spin labelling MRI for assessment of cerebral perfusion in children with moyamoya disease: comparison with dynamic susceptibility contrast MRI

Goetti, Robert ; O'Gorman, Ruth ; Khan, Nadia ; Kellenberger, Christian ; Scheer, Ianina

In: Neuroradiology, 2013, vol. 55, no. 5, p. 639-647

Add to personal list
    Summary
    Introduction: This study seeks to evaluate the diagnostic accuracy of cerebral perfusion imaging with arterial spin labelling (ASL) MR imaging in children with moyamoya disease compared to dynamic susceptibility contrast (DSC) imaging. Methods: Ten children (7 females; age, 9.2 ± 5.4years) with moyamoya disease underwent cerebral perfusion imaging with ASL and DSC on a 3-T MRI scanner in the same session. Cerebral perfusion images were acquired with ASL (pulsed continuous 3D ASL sequence, 32 axial slices, TR = 5.5s, TE = 25ms, FOV = 24cm, matrix = 128 × 128) and DSC (gradient echo EPI sequence, 35 volumes of 28 axial slices, TR = 2,000ms, TE = 36ms, FOV = 24cm, matrix = 96 × 96, 0.2ml/kg Gd-DOTA). Cerebral blood flow maps were generated. ASL and DSC images were qualitatively assessed regarding perfusion of left and right ACA, MCA, and PCA territories by two independent readers using a 3-point-Likert scale and quantitative relative cerebral blood flow (rCBF) was calculated. Correlation between ASL and DSC for qualitative and quantitative assessment and the accuracy of ASL for the detection of reduced perfusion per territory with DSC serving as the standard of reference were calculated. Results: With a good interreader agreement (κ = 0.62) qualitative perfusion assessment with ASL and DSC showed a strong and significant correlation (ρ = 0.77; p < 0.001), as did quantitative rCBF (r = 0.79; p < 0.001). ASL showed a sensitivity, specificity and accuracy of 94%, 93%, and 93% for the detection of reduced perfusion per territory. Conclusion: In children with moyamoya disease, unenhanced ASL enables the detection of reduced perfusion per vascular territory with a good accuracy compared to contrast-enhanced DSC