Manasse Sharmiladevi, Koskas Patricia, Savatovsky Julien, Deschamps Romain, Vignal-Clermont Catherine, Boudot de la Motte Marine, Papeix Caroline, Trunet Stéphanie, Lecler Augustin
Department of Neuroradiology, Fondation Adolphe de Rothschild Hospital, 75019 Paris, France.
Department of Neuroradiology, Fondation Adolphe de Rothschild Hospital, 75019 Paris, France.
Diagn Interv Imaging. 2024 Dec;105(12):481-488. doi: 10.1016/j.diii.2024.08.001. Epub 2024 Sep 5.
The purpose of this study was to compare the capabilities of contrast-enhanced fat-suppressed (CE FS) three-dimensional fluid-attenuated inversion recovery (3D FLAIR) brain magnetic resonance imaging (MRI) with those of coronal T2-weighted orbital MRI obtained at 3 Tesla for the diagnosis of optic neuritis (ON).
Patients who presented to our center with acute visual loss and underwent MRI examination of the orbits and the brain between November 2014 and February 2020 were retrospectively included. Three radiologists independently and blindly analyzed CE FS 3D FLAIR and coronal T2-weighted images. Disagreements in image interpretation were resolved by consensus with an independent neuroradiologist who was not involved in the initial reading sessions. The primary adjudication criterion for the diagnosis of ON was the presence of an optic nerve hypersignal. Sensitivity, specificity, and accuracy of CE 3D FLAIR brain images were compared with those of coronal T2-weighted orbital images using McNemar test. Artifacts were classified into three categories and compared between the two image sets.
A total of 1023 patients were included. There were 638 women and 385 men with a mean age of 42 ± 18.3 (standard deviation) years (age range: 6-92 years). Optic nerve hyperintensities were identified in 375/400 (94%) patients with ON using both 3D FLAIR and coronal T2-weighted images. Sensitivity, specificity, and accuracy of both sequences were 94% (95% CI: 91.3-96.1), 79% (95% CI: 75.5-82.2), and 89% (95% CI: 86.8-90.7), respectively. Optic disc hypersignal was detected in 120/400 patients (30%) using 3D FLAIR compared to 3/400 (0.75%) using coronal T2-weighted images (P < 0.001). Optic radiation hypersignal was observed in 2/400 (0.5%) patients using 3D FLAIR images. Significantly more artifacts (moderate or severe) were observed on coronal T2-weighted images (801/1023; 78%) by comparison with 3D FLAIR images (448/1023; 44%) (P < 0.001).
The performance of 3D FLAIR brain MRI for the diagnosis of ON is not different from that of coronal T2-weighted orbital MRI and its use for optic nerve analysis may be beneficial.
本研究旨在比较对比增强脂肪抑制(CE FS)三维液体衰减反转恢复(3D FLAIR)脑磁共振成像(MRI)与在3特斯拉场强下获得的冠状位T2加权眼眶MRI对视神经炎(ON)的诊断能力。
回顾性纳入2014年11月至2020年2月期间因急性视力丧失到本中心就诊并接受眼眶和脑部MRI检查的患者。三名放射科医生独立且盲法分析CE FS 3D FLAIR和冠状位T2加权图像。图像解读中的分歧通过与未参与初始读片的独立神经放射科医生达成共识来解决。ON诊断的主要判定标准是视神经高信号的存在。使用McNemar检验比较CE 3D FLAIR脑图像与冠状位T2加权眼眶图像的敏感性、特异性和准确性。将伪影分为三类并在两组图像之间进行比较。
共纳入1023例患者。其中女性638例,男性385例,平均年龄42±18.3(标准差)岁(年龄范围:6 - 92岁)。使用3D FLAIR和冠状位T2加权图像在375/400(94%)例ON患者中均发现视神经高信号。两种序列的敏感性、特异性和准确性分别为94%(95%可信区间:91.3 - 96.1)、79%(95%可信区间:75.5 - 82.2)和89%(95%可信区间:86.8 - 90.7)。使用3D FLAIR在120/400例患者(30%)中检测到视盘高信号,而使用冠状位T2加权图像在3/400例患者(0.75%)中检测到视盘高信号(P < 0.001)。使用3D FLAIR图像在2/400例患者(0.5%)中观察到视辐射高信号。与3D FLAIR图像(448/1023;44%)相比,冠状位T2加权图像上观察到的伪影(中度或重度)明显更多(801/1023;78%)(P < 0.001)。
3D FLAIR脑MRI对ON的诊断性能与冠状位T2加权眼眶MRI无异,其用于视神经分析可能有益。