腰椎X线摄影人工智能测量技术研究进展-《中国医学物理学杂志》

文章信息/Info

Title:: Advances in artificial intelligence technology for parameter measurement in lumbar X-ray photograph

作者:: 姚红艳1; 邓兴1; 3; 陈晓飞2; 王闻奇2; 周晟2; 1.甘肃中医药大学第一临床医学院，甘肃兰州 730000； 2.甘肃省中医院放射影像科，甘肃兰州 730050； 3.南充市中心医院介入放射科，四川南充 637000

Author(s):: YAO Hongyan1; DENG Xing1; 3; CHEN Xiaofei2; WANG Wenqi2; ZHOU Sheng2; 1. The First Clinical Medical School of Gansu University of Chinese Medicine, Lanzhou 730000, China 2. Department of Medical Imaging, Gansu Provincial Hospital of TCM, Lanzhou 730050, China 3. Department of Interventional Radiology, Nanchong Central Hospital, Nanchong 637000, China

Keywords:: Keywords: low back pain lumbar X-ray deep learning artificial intelligence review

摘要:: 腰痛是全球范围内普遍存在的公共健康卫生问题，但临床医师对腰椎X线图像的视觉分析和主观判断已不能满足临床精准化、定量化的诊疗需求。近年来，人工智能（AI）及大数据技术的蓬勃发展，为医疗图像的智能检测和定量分析提供了条件和基础。众多学者通过人工神经网络、支持向量机和卷积神经网络等模型的建立和算法的改进，使腰椎X线定量分析及诊断成为可能。AI技术与医疗图像的交叉融合在减轻临床医师工作负荷、有效降低或消除手工测量误差和辅助临床医师从定量角度更客观地评估脊柱畸形等疾病具有很好的应用前景。目前，AI技术辅助腰椎疾病诊断的发展仍处于早期阶段，AI算法的改进、高质量数据库建立及制定新参数的量化标准等需要进一步探索。

Abstract:: Abstract: Low back pain (LBP) is a prevalent public health problem worldwide, but the visual analysis and subjective judgment of lumbar spine X-ray photographs by clinicians no longer meet the demands for clinical precision and quantification of diagnosis and treatment. In recent years, the flourishing development of artificial intelligence (AI) and big data technology has provided conditions and foundations for intelligent detection and quantitative analysis of medical images. The establishment of models such as artificial neural networks, support vector machines and convolutional neural networks and the improvement of algorithms made the quantitative analysis and diagnosis of lumbar spine X-ray photographs possible. The cross-integration of AI technology with medical images has a promising application in reducing the workload of clinicians, effectively decreasing or eliminating manual measurement errors, and assisting clinicians in assessing spinal deformities and other related diseases more objectively from a quantitative perspective. At present, the development of AI technology-assisted diagnosis of lumbar spine diseases is still at an early stage, and the improvement of AI algorithms, the establishment of high-quality databases and the development of new criteria for parameter quantification need to be further explored.