Prakkamakul Supada, Witzel Thomas, Huang Susie, Boulter Daniel, Borja Maria J, Schaefer Pamela, Rosen Bruce, Heberlein Keith, Ratai Eva, Gonzalez Gilberto, Rapalino Otto
Neuroradiology Division, Department of Radiology, Massachusetts General Hospital, Boston, MA.
Department of Radiology, King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bangkok, Thailand.
J Neuroimaging. 2016 Sep;26(5):503-10. doi: 10.1111/jon.12365. Epub 2016 Jun 7.
To compare an ultrafast brain magnetic resonance imaging (MRI) protocol to the conventional protocol in motion-prone inpatient clinical settings.
This retrospective study was HIPAA compliant and approved by the Institutional Review Board with waived inform consent. Fifty-nine inpatients (30 males, 29 females; mean age 55.1, range 23-93 years)who underwent 3-Tesla brain MRI using ultrafast and conventional protocols, both including five sequences, were included in the study. The total scan time for five ultrafast sequences was 4 minutes 59 seconds. The ideal conventional acquisition time was 10 minutes 32 seconds but the actual acquisition took 15-20 minutes. The average scan times for ultrafast localizers, T1-weighted, T2-weighted, fluid-attenuated inversion recovery (FLAIR), diffusion-weighted, T2*-weighted sequences were 14, 41, 62, 96, 80, 6 seconds, respectively. Two blinded neuroradiologists independently assessed three aspects: (1) image quality, (2) gray-white matter (GM-WM) differentiation, and (3) diagnostic concordance for the detection of six clinically relevant imaging findings. Wilcoxon signed-rank test was used to compare image quality and GM-WM scores. Interobserver reproducibility was calculated.
The ultrafast T1-weighted sequence demonstrated significantly better image quality (P = .005) and GM-WM differentiation (P < .001) compared to the conventional sequence. There was high agreement (>85%) between both protocols for the detection of mass-like lesion, hemorrhage, diffusion restriction, WM FLAIR hyperintensities, subarachnoid FLAIR hyperintensities, and hydrocephalus.
The ultrafast protocol achieved at least comparable image quality and high diagnostic concordance compared to the conventional protocol. This fast protocol can be a viable option to replace the conventional protocol in motion-prone inpatient clinical settings.
在易出现运动的住院临床环境中，将一种超快脑磁共振成像（MRI）方案与传统方案进行比较。
本回顾性研究符合健康保险流通与责任法案（HIPAA）要求，并经机构审查委员会批准，豁免了知情同意。纳入59例住院患者（30例男性，29例女性；平均年龄55.1岁，范围23 - 93岁），这些患者使用超快和传统两种方案进行了3特斯拉脑MRI检查，两种方案均包括五个序列。五个超快序列的总扫描时间为4分59秒。传统采集的理想时间为10分32秒，但实际采集时间为15 - 20分钟。超快定位像、T1加权、T2加权、液体衰减反转恢复（FLAIR）、扩散加权、T2*加权序列的平均扫描时间分别为14、41、62、96、80、6秒。两名盲法神经放射科医生独立评估三个方面：（1）图像质量，（2）灰白质（GM - WM）区分，（3）对六个临床相关影像表现检测的诊断一致性。采用Wilcoxon符号秩检验比较图像质量和GM - WM评分。计算观察者间的可重复性。
与传统序列相比，超快T1加权序列的图像质量（P = .005）和GM - WM区分度（P < .001）明显更好。两种方案在检测类肿块病变、出血、扩散受限、白质FLAIR高信号、蛛网膜下腔FLAIR高信号和脑积水方面的一致性较高（>85%）。
与传统方案相比，超快方案实现了至少相当的图像质量和高诊断一致性。在易出现运动的住院临床环境中，这种快速方案可以作为替代传统方案的可行选择。
