Center for Imaging Science and Institute for Computational Medicine, Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, 21218, USA.
Vanderbilt University, Nashville, TN, 37235, USA.
Neuroradiology. 2020 Sep;62(9):1157-1167. doi: 10.1007/s00234-020-02436-6. Epub 2020 May 19.
It has long been thought that the acoustic radiation (AR) white matter fibre tract from the medial geniculate body of the thalamus to the Heschl's gyrus cannot be reconstructed via single-fibre analysis of clinical diffusion tensor imaging (DTI) scans. A recently developed single-fibre probabilistic method suggests otherwise. The method uses dynamic programming (DP) to compute the most probable paths between two regions of interest. This study aims to observe the ability of single-fibre probabilistic analysis via DP to visualise the AR in clinical DTI scans from legacy pilot cohorts of subjects with normal hearing (NH) and profound hearing loss (HL).
Single-fibre probabilistic analysis via DP was applied to reconstruct 3D models of the AR in the two cohorts. DTI and T1 data at 1.5 T for subjects with NH (n = 11) and HL (n = 5), as well as 3 T for NH (n = 1) and HL (n = 1), were used.
The topographical features of AR previously observed in post-mortem and multi-fibre analyses can be visualised in DTI scans of 16 subjects and 2 atlases with a success rate of 100%. Relative to MNI coordinates, there was no significant difference in the varifold distances between the topography of the tracts in the 1.5 T cohort.
The AR can be visualised in clinical 1.5 T and 3 T DTI scans using single-fibre probabilistic analysis via DP, hence, the potential for DP to visualise the AR in medical and pre-surgical applications in pathologies such as vestibular schwannoma, multiple sclerosis, thalamic tumours and stroke as well as hearing loss.
长期以来，人们一直认为，从丘脑内侧膝状体到 Heschl 回的声学辐射（AR）白质纤维束无法通过临床扩散张量成像（DTI）扫描的单纤维分析来重建。一种新开发的单纤维概率方法则持不同意见。该方法使用动态规划（DP）来计算两个感兴趣区域之间最可能的路径。本研究旨在观察通过 DP 进行单纤维概率分析来可视化正常听力（NH）和重度听力损失（HL）受试者的临床 DTI 扫描中 AR 的能力。
通过 DP 对来自具有 NH（n=11）和 HL（n=5）的两个队列的 NH 和 HL 的受试者的 NH 和 HL 的 1.5 T 数据以及 3 T 的 NH（n=1）和 HL（n=1）的单纤维概率分析进行应用，以重建 AR 的 3D 模型。
可以在 16 名受试者和 2 个图谱的 DTI 扫描中可视化以前在尸体和多纤维分析中观察到的 AR 的地形特征，成功率为 100％。与 MNI 坐标相比，1.5 T 队列中束的地形之间的变体距离没有显着差异。
可以使用通过 DP 进行的单纤维概率分析来可视化临床 1.5 T 和 3 T DTI 扫描中的 AR，因此 DP 有可能在诸如前庭神经鞘瘤，多发性硬化症，丘脑肿瘤和中风以及听力损失等病理中用于医学和术前应用中的 AR 可视化。
