机器学习在心血管疾病诊断中的研究进展  

Review on machine learning approaches for cardiovascular disease diagnosis

作　　者：赵梦蝶 孙九爱[2] ZHAO Mengdie;SUN Jiuai(School of Medical Instrument and Food Engineering,University of Shanghai for Science and Technology,Shanghai 200093;Shanghai University of Medicine&Health Sciences,Shanghai 201318)

机构地区：[1]上海理工大学医疗器械与食品学院,上海200093 [2]上海健康医学院,上海201318

出　　处：《北京生物医学工程》2020年第2期208-214,共7页Beijing Biomedical Engineering

基　　金：上海健康医学院协调创新专项(SPCI-17-18-001)资助。

摘　　要：当前医生对心血管疾病的诊断主要依赖对患者心血管影像的分析,同时,医生还需要考虑患者的各项生理健康指标、既往病史、生活环境等信息,该方法存在效率低和成本高等问题。因此,人们试图利用机器学习方法辅助心血管疾病的诊断。本文首先总结了机器学习在冠状动脉计算机断层扫描、超声心动图、心电图等多种心血管影像处理中的应用;其次,对现有的机器学习模型进行了评估和分析;最后,本文认为虽然现有的基于机器学习的心血管疾病诊断方法已经可以媲美专业医生的水平,但是,该方法仍面临医学数据难以大量采集、医学成像信噪比低等困难。未来的研究方向应在小样本诊断模型的性能、多模态医学数据的融合、医学数据的共享等方面继续改进。At present,the diagnosis of cardiovascular diseases mainly depends on the analysis of cardiovascular images.Meanwhile,doctors also need to take into account the health indicators,medical history,living environment and other information of patients.This method has the disadvantages of low efficiency and high cost,and therefore,machine learning method has been adapted to solve these problems.This paper,summarizes the applications of machine learning in coronary computed tomography,echocardiography,electrocardiogram and so on.By evaluating and analyzing the existing models,the existing machine learning based methods may achieve similar level as clinical doctors.However the machine learning approaches need to solve the problems such as less training data and low signal-to-noise ratio of medical imaging data.The future research direction should continue to improve the performance of small sample diagnostic model,the fusion of multimodal medical data,the sharing of medical diagnostic data and so on.

关 键 词：机器学习 心血管疾病 医学影像 多模态数据 辅助诊断 

分 类 号：R318[医药卫生—生物医学工程]