Department of Biomedical Engineering, Hanyang University, Seoul, South Korea.
Bonbridge hospital, 562, Songpa-daero, Songpa-gu, Seoul, South Korea.
Comput Methods Programs Biomed. 2019 Dec;182:105063. doi: 10.1016/j.cmpb.2019.105063. Epub 2019 Sep 3.
Rotator cuff muscle tear is one of the most frequent reason of operations in orthopedic surgery. There are several clinical indicators such as Goutallier grade and occupation ratio in the diagnosis and surgery of these diseases, but subjective intervention of the diagnosis is an obstacle in accurately detecting the correct region.
Therefore, in this paper, we propose a fully convolutional deep learning algorithm to quantitatively detect the fossa and muscle region by measuring the occupation ratio of supraspinatus in the supraspinous fossa. In the development and performance evaluation of the algorithm, 240 patients MRI dataset with various disease severities were included.
As a result, the pixel-wise accuracy of the developed algorithm is 0.9984 ± 0.073 in the fossa region and 0.9988 ± 0.065 in the muscle region. The dice coefficient is 0.9718 ± 0.012 in the fossa region and 0.9463 ± 0.047 in the muscle region.
We expect that the proposed convolutional neural network can improve the efficiency and objectiveness of diagnosis by quantifying the index used in the orthopedic rotator cuff tear.
肩袖肌肉撕裂是矫形外科最常见的手术原因之一。在这些疾病的诊断和手术中，有几个临床指标，如 Goutallier 分级和占有率，但诊断中的主观干预是准确检测正确区域的障碍。
因此，在本文中，我们提出了一种完全卷积深度学习算法，通过测量冈上肌在冈上窝中的占有率来定量检测窝和肌肉区域。在算法的开发和性能评估中，包括了 240 名具有不同疾病严重程度的患者的 MRI 数据集。
开发的算法在窝区域的像素精度为 0.9984 ± 0.073，在肌肉区域的像素精度为 0.9988 ± 0.065。窝区域的骰子系数为 0.9718 ± 0.012，肌肉区域的骰子系数为 0.9463 ± 0.047。
我们期望所提出的卷积神经网络能够通过量化矫形肩袖撕裂中使用的指标来提高诊断的效率和客观性。
