机器学习在脊柱疾病临床诊断中的应用研究进展

Survey of Machine Learning Applications in the Clinical Diagnosis of Spinal Diseases

脊柱相关疾病是现代社会中的多发病、常见病.鉴于脊柱生理结构的复杂性,脊柱疾病的临床诊断需要经验丰富的医师.机器学习技术可有助于脊柱疾病的快速、精准诊断,辅助医生进行术前规划以及术后结果预测,有助于提高诊断效率,减少医护人员的负担并降低误诊率.回顾了常用于脊柱疾病临床诊断领域的机器学习技术的研究现状,从脊柱的分割、椎骨定位和标记以及脊柱疾病临床辅助诊断等三个方面对机器学习在脊柱疾病临床诊断中的应用研究进行综述,分析目前机器学习技术在脊柱疾病临床诊断中的不足和挑战.

Spine-related diseases are frequently-occurring and common diseases in modern society.In view of the complexity of the physiological structure of the spine,the clinical diagnosis of spinal diseases needs experienced doctors.However,the machine learning technology can contribute to the rapid and accurate diagnosis of spinal diseases,and assist doctors in preoperative planning and postoperative outcome prediction to improve the efficiency of diagnosis,and reduce the burden of medical staff and the misdiagnosis rate.The current work summarizes the research status of machine learning technology commonly used in clinical diagnosis of spinal diseases,further,and analyzes the following three aspects:spinal segmentation,vertebral location and labeling and the main application progress of clinical auxiliary diagnosis.Meanwhile,the shortcomings and challenges of machine learning technology in clinical diagnosis of spinal diseases are proposed and the prospects for the future are put forward.