Joint Department of Biomedical Engineering, UNC Chapel Hill & NC State University, Chapel Hill 27599, NC, USA.
Joint Department of Biomedical Engineering, UNC Chapel Hill & NC State University, Chapel Hill 27599, NC, USA.
Comput Methods Programs Biomed. 2021 Jul;206:106120. doi: 10.1016/j.cmpb.2021.106120. Epub 2021 Apr 27.
Direct measurement of muscle-tendon junction (MTJ) position is important for understanding dynamic tendon behavior and muscle-tendon interaction in healthy and pathological populations. Traditionally, obtaining MTJ position during functional activities is accomplished by manually tracking the position of the MTJ in cine B-mode ultrasound images - a laborious and time-consuming process. Recent advances in deep learning have facilitated the availability of user-friendly open-source software packages for automated tracking. However, these software packages were originally intended for animal pose estimation and have not been widely tested on ultrasound images. Therefore, the purpose of this paper was to evaluate the efficacy of deep neural networks to accurately track medial gastrocnemius MTJ positions in cine B-mode ultrasound images across tasks spanning controlled loading during isolated contractions to physiological loading during treadmill walking.
Cine B-mode ultrasound images of the medial gastrocnemius MTJ were collected from 15 subjects (6M/9F, 23 yr, 71.9 kg, 1.8 m) during treadmill walking at 1.25 m/s and during maximal voluntary isometric plantarflexor contractions (MVICs). Five deep neural networks were trained using 480 manually-labeled images collected during walking, defined as the ground truth, and were then used to predict MTJ position in images from novel subjects: 1) during walking (novel-subject) and 2) during MVICs (novel-condition).
We found an average mean absolute error of 1.26±1.30 mm and 2.61±3.31 mm between the ground truth and predicted MTJ positions in the novel-subject and novel-condition evaluations, respectively.
Our results provide support for the use of open-source software for creating deep neural networks to reliably track MTJ positions in B-mode ultrasound images. We believe this approach to MTJ position tracking is an accessible and time-saving solution, with broad applications for many fields, such as rehabilitation or clinical diagnostics.
直接测量肌肉肌腱结合部（MTJ）的位置对于理解健康和病理人群中动态肌腱的行为和肌肉肌腱的相互作用非常重要。传统上，在功能活动中获得 MTJ 位置是通过手动跟踪 Cine B 模式超声图像中 MTJ 的位置来实现的，这是一个繁琐且耗时的过程。深度学习的最新进展使得用户友好的开源软件包可用于自动跟踪。然而，这些软件包最初是为动物姿势估计而设计的，尚未在超声图像上进行广泛测试。因此，本文的目的是评估深度神经网络在从单独收缩的受控加载到跑步机行走时的生理加载等多个任务中准确跟踪 Cine B 模式超声图像中内侧腓肠肌 MTJ 位置的功效。
从 15 名受试者（6 名男性/9 名女性，23 岁，71.9 千克，1.8 米）在跑步机上以 1.25 米/秒的速度行走和进行最大自愿等长跖屈肌收缩（MVIC）期间收集内侧腓肠肌 MTJ 的 Cine B 模式超声图像。使用在行走期间（定义为地面实况）收集的 480 张手动标记图像对 5 个深度神经网络进行训练，然后用于预测来自新受试者的图像中的 MTJ 位置：1）行走期间（新受试者）和 2）MVIC 期间（新条件）。
我们发现，在新受试者和新条件评估中，地面实况和预测 MTJ 位置之间的平均绝对误差分别为 1.26±1.30 毫米和 2.61±3.31 毫米。
我们的结果为使用开源软件创建深度神经网络以可靠地跟踪 B 模式超声图像中的 MTJ 位置提供了支持。我们相信，这种 MTJ 位置跟踪方法是一种易于访问且节省时间的解决方案，在康复或临床诊断等许多领域都有广泛的应用。
