de Blank Peter, Fisher Michael J, Gittleman Haley, Barnholtz-Sloan Jill S, Badve Chaitra, Berman Jeffrey I
Division of Oncology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States; University of Cincinnati, Department of Pediatrics, Cincinnati, OH, United States.
Division of Oncology, The Children's Hospital of Philadelphia, Philadelphia, PA, United States.
Exp Neurol. 2018 Jan;299(Pt B):308-316. doi: 10.1016/j.expneurol.2017.06.004. Epub 2017 Jun 3.
Fractional anisotropy (FA) of the optic radiations has been associated with vision deficit in multiple intrinsic brain pathologies including NF1 associated optic pathway glioma, but hand-drawn regions of interest used in previous tractography methods limit consistency of this potential biomarker. We created an automated method to identify white matter tracts in the optic radiations and compared this method to previously reported hand-drawn tractography.
Automated tractography of the optic radiation using probabilistic streamline fiber tracking between the lateral geniculate nucleus of the thalamus and the occipital cortex was compared to the hand-drawn method between regions of interest posterior to Meyer's loop and anterior to tract branching near the calcarine cortex. Reliability was assessed by two independent raters in a sample of 20 healthy child controls. Among 50 children with NF1-associated optic pathway glioma, the association of FA and visual acuity deficit was compared for both tractography methods.
Hand-drawn tractography methods required 2.6±0.9min/participant; automated methods were performed in <1min of operator time for all participants. Cronbach's alpha was 0.83 between two independent raters for FA in hand-drawn tractography, but repeated automated tractography resulted in identical FA values (Cronbach's alpha=1). On univariate and multivariate analyses, FA was similarly associated with visual acuity loss using both methods. Receiver operator characteristic curves of both multivariate models demonstrated that both automated and hand-drawn tractography methods were equally able to distinguish normal from abnormal visual acuity.
Automated tractography of the optic radiations offers a fast, reliable and consistent method of tract identification that is not reliant on operator time or expertise. This method of tract identification may be useful as DTI is developed as a potential biomarker for visual acuity.
在包括与神经纤维瘤病1型（NF1）相关的视路胶质瘤在内的多种脑内疾病中，视辐射的各向异性分数（FA）与视力缺陷有关，但以往的纤维束成像方法中使用的手工绘制的感兴趣区域限制了这种潜在生物标志物的一致性。我们创建了一种自动方法来识别视辐射中的白质纤维束，并将该方法与先前报道的手工绘制纤维束成像进行比较。
使用概率性流线型纤维追踪技术对丘脑外侧膝状体和枕叶皮质之间的视辐射进行自动纤维束成像，并与在迈耶袢后方和距状沟皮质附近纤维束分支前方的感兴趣区域之间的手工绘制方法进行比较。由两名独立评估者在20名健康儿童对照样本中评估可靠性。在50名患有NF1相关视路胶质瘤的儿童中，比较了两种纤维束成像方法中FA与视力缺陷的相关性。
手工绘制纤维束成像方法每个参与者需要2.6±０.9分钟；自动方法在所有参与者中操作时间不到1分钟。在手工绘制纤维束成像中，两名独立评估者之间FA的Cronbach's α为0.83，但重复自动纤维束成像产生相同的FA值（Cronbach's α = 1）。在单变量和多变量分析中，两种方法中FA与视力丧失的相关性相似。两个多变量模型的受试者操作特征曲线表明，自动和手工绘制纤维束成像方法在区分正常和异常视力方面同样有效。
视辐射的自动纤维束成像提供了一种快速、可靠且一致的纤维束识别方法，不依赖于操作人员的时间或专业知识。随着扩散张量成像（DTI）被开发为视力的潜在生物标志物，这种纤维束识别方法可能会有用。
