基于MRI深度学习在膝关节前交叉韧带损伤的研究进展

Research progress of anterior cruciate ligament injury of knee using MRI:Based on deep learning

磁共振成像是膝关节前交叉韧带损伤的首选无创评估方法,目前对前交叉韧带损伤的诊断主要靠放射科医生的临床经验且耗时耗力。深度学习作为机器学习的一个重要新兴分支,它以神经网络为架构、对数据进行表征学习。近年来,深度学习在膝关节前交叉韧带的应用主要集中在前交叉韧带分割和损伤分类(包括二分类和多分类),对于损伤韧带的分割及基于韧带损伤对相关疾病预测的研究仍处在起步阶段。尽管如此,深度学习既可快速实现前交叉韧带的自动分割,又可对前交叉韧带损伤进行自动分类评估,这将显著提高放射科医生的工作效率。本文就基于MRI深度学习在膝关节前交叉韧带损伤方面的研究予以综述。

Magnetic resonance imaging is the preferred non-invasive assessment method for anterior cruciate ligament injuries of the knee.At present,the diagnosis of anterior cruciate ligament injury mainly relies on the clinical experience of radiologists and is time-consuming and labor-intensive.Deep learning is an arising meaningful branch of machine learning,which uses neural networks as the architecture and characterizes data for learning.In recent years,the main application of deep learning in anterior cruciate ligament of the knee joint focused on anterior cruciate ligament segmentation and injury classification(including binary classification and multi-classification),but researches on segmentation of injured ligaments and prediction of related diseases are still in an initial stage.Nevertheless,deep learning can quickly achieve the automatic segmentation of anterior cruciate ligament and the classification assessment of anterior cruciate ligament injury simultaneously,which can significantly improve the productiveness of radiologists.In this paper,we review the research on MRI-based on deep learning in anterior cruciate ligament injuries of the knee.