目的：使用代价敏感的机器学习方法，依据超声参数对右心衰竭进行诊断，以增进超声在右心衰竭诊断中的价值。方法：选取右心衰竭患者82例为病例组，非右心衰竭者106例为对照组，使用的主要超声参数包括：三尖瓣环收缩期位移（TAPSE）、右室面积变化率（RVFAC）、三尖瓣环收缩期运动速度（Sm）及舒张早期运动速度（Em）、Tei指数、三尖瓣环等容收缩期加速度（IVA）及右室内径、右房内径、左室内径等。通过代价敏感的朴素贝叶斯分析，以上述参数为特征建立诊断模型，对该模型进行交叉验证，比较代价敏感的判别模型与代价均等的模型的诊断效果。结果：较单一参数诊断右心衰竭，机器学习模型能达到更好的正确率（91%），代价敏感的朴素贝叶斯方法可达到更小的总体代价。结论：通过机器学习建立的诊断模型较单一超声参数的效果更佳，较常规方案可降低诊断代价，该方法在右心衰竭的影像筛查中具有潜在实用价值。
Abstract
Objective: To improve the efficacy of echocardiographic method in evaluating right ventricular function, and to explore the value of cost-sensitive machine learning method model based on ultrasound parameters in the diagnosis of right ventricular failure. Methods: Eighty-two subjects with right heart failure(RHF group) and 106 subjects without right heart failure(non-RHF group) were enrolled in this study. Multi-echocardiographic indices were selected as features, including the right ventricular fractional area change(RVFAC), tricuspid annular plane systolic excursion(TAPSE), tricuspid flow peak velocity at early diastole(E), tricuspid annular peak systolic velocity(Sm) and peak early diastolic velocity(Em), Tei index, isovolumic acceleration(IVA) at tricuspid annular, and the diameter of the left ventricle, right ventricle, right atrium were measured and calculated. Cost-sensitive naive bayes analysis was used to build diagnosis model by using those parameters. The model was tested by 10-fold cross-validation. The diagnostic results of the cost-sensitive model and the classical naive bayes model were compared. Results: Compared with any single parameter, machine learning model can achieve better accuracy in diagnosis of right heart failure(91%). The cost-sensitive naive bayes method achieved a smaller overall cost. Conclusion: The diagnostic model established by machine learning is better than any single ultrasound parameter, which can reduce the diagnostic cost compared with conventional method. This method has potential practical value in screening of right ventricular failure by imaging.
关键词
心力衰竭 /
充血性 /
超声心动描记术
Key words
Heart failure, congestive /
Echocardiography
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]Selly J, Iriart X, Roubertie F, et al. Multivariable Assessment of the Right Ventricle By Echocardiography in Patients with Repaired Tetralogy of Fallot Undergoing Pulmonary Valve Replacement: a Comparative Study with Magnetic Resonance Imaging[J]. Archf Cardiovasc Dis, 2015, 108(1): 5-15.
[2]中华医学会心血管病学分会. 右心衰竭诊断和治疗中国专家共识[J]. 中华心血管病杂志，2012，40（6）：449-461.
[3]Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the Echocardiographic Assessment of the Right Heart in Adults: a Report From the American Society of Echocardiography: Endorsed By the European Association of Echocardiography, a Registered Branch of the European Society of Cardiology, and the Canadian Society of Echocardiography[J]. J Am Soc Echocardiogr, 2010, 23(7): 685-713.
[4]Lang RM, Badano LP, Mor-avi V, et al. Recommendations for Cardiac Chamber Quantification By Echocardiography in Adults: an Update From the American Society of Echocardiography and the European Association of Cardiovascular Imaging[J]. J Am Soc Echocardiogr, 2015, 28(1): 1-39.
[5]Simon MA. Assessment and Treatment of Right Ventricular Failure[J]. Nat Rev Cardiol, 2013, 10(4): 204-218.
[6]Dutta T, Aronow WS. Echocardiographic Evaluation of the Right Ventricle: Clinical Implications[J]. Clin Cardiol, 2017, 40(8): 542-548.
[7]Venkatachalam S, Wu G, Ahmad M. Echocardiographic assessment of the right ventricle in the current era: Application in clinical practice[J]. Echocardiography, 2017, 34(12): 1930-1947.
[8]Focardi M, Cameli M, Carbone S, et al. Traditional and Innovative Echocardiographic Parameters for the Analysis of Right Ventricular Performance in Comparison with Cardiac Magnetic Resonance[J]. Eur Heart J Cardiovasc Imaging, 2015, 16(1): 47-52.
[9]Yang F, Wang H, Mi H, et al. Using Random Forest for Reliable Classification and Cost-sensitive Learning for Medical Diagnosis[J]. Bmc Bioinformatics, 2009, 10(Suppl 1): S22.
[10]Dinunzio GM. A New Decision to Take for Cost-sensitive Naive Bayes Classifiers[J]. Inf Process Manag, 2014, 50(5): 653-674.
[11]Shapiro SD. Refining Lung Cancer Screening Criteria in the Era of Value-based Medicine[J]. PLoS Med, 2017, 14(2): e1002226.
[12]Gatos I, Tsantis S, Spiliopoulos S, et al. A Machine-learning Algorithm Toward Color Analysis for Chronic Liver Disease Classification, Employing Ultrasound Shear Wave Elastography[J]. Ultrasound Med Biol, 2017, 43(9): 1797-1810.
[13]Medvedofsky D, Addetia K, Hamilton J, et al. Semi-automated Echocardiographic Quantification of Right Ventricular Size and Function[J]. Int J Cardiov Imaging, 2015, 31(6): 1149-1157.
基金
河北省医学科学研究课题计划（编号20150529）。
