慢性阻塞性肺疾病患者呼吸音频谱分析与肺功能的相关性研究

Study on the correlation between breath sound audio frequency spectrum analysis and lung function in patients with chronic obstructive pulmonary disease

目的研究不同肺功能分级慢性阻塞性肺疾病(chronic obstructive pulmonary disease,COPD)患者呼吸音声谱特征,初步确定不同肺功能分级COPD呼吸音声谱图特征性的参数和指标,探讨不同肺功能分级COPD患者呼吸听诊差异的可视化和可测量化。方法入选2018年10月至2020年12月在上海交通大学医学院附属新华医院崇明分院诊治的经肺功能检查确诊为COPD52例患者[中轻-中度阻塞性通气功能障碍22例(轻-中度组)、重度阻塞性通气功能障碍30例(重度组)]和通气功能正常对照者36例(对照组)。分别用电子听诊器对呼吸音采样,传输至Audition音频软件,计算不同肺功能分级样本呼吸音声压级(sound pressure level,SPL)的相关数据及与每分钟通气量(minute ventilation,VE)的比值。采用回顾性病例对照研究,正态分布的计量资料采用单因素方差分析,两两比较采用SNK-q检验。结果轻-中度组、重度组COPD患者及对照组的呼吸音频在吸气相与呼气相50~200 Hz频段内的声压差分别为(-0.5±2.2)、(-1.6±6.1)、(0.7±4.0)dB,3组间比较差异有统计学意义(F=2.64,P=0.038);轻-中度组、重度组COPD患者与对照组的呼吸音吸气相单位分钟通气量声压级(sound pressure level per minute ventilation,SPL50~200/VE)分别为(5.7±1.8)、(6.1±2.3)、(5.4±0.9)dB/L;呼气相SPL50~200/VE分别为(5.8±1.7)、(6.3±2.2)、(5.3±1.2)dB/L,3组间比较差异均有统计学意义(F吸=3.26,P=0.048;F呼=2.44,P=0.045)。ROC曲线结果显示,吸气相SPL50~200/VE、呼气相SPL50~200/VE、声压差评估不同分级阻塞性通气功能障碍具有诊断价值(AUC值分别为0.697、0.725、0.686,P值分别为0.015、0.005、0.022)。部分COPD患者呼吸音时频图中可出现异常高能量水平线。结论不同程度阻塞性通气功能障碍患者与通气功能正常者呼吸音频谱分析可出现明显差异,可为辅助判断阻塞性通气功能障碍提供依据。

Objective To study the characteristics of breath sound spectroscopy in chronic obstructive pulmonary disease(COPD)patients with different lung function grades,to preliminarily determine the characteristic parameters and indicators of breath sound spectrograms with different lung function grades of COPD,and to explore the visualization and measurability of the differences in respiratory auscultation in patients with different lung function grades.Methods Patients diagnosed and treated in Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine Chongming Branch from October 2018 to December 2020 were selected.Fifty-two patients with COPD(22 patients with moderate to moderate obstructive ventilation dysfunction(mild to moderate group),30 patients with severe obstructive ventilation dysfunction(severe group))and 36 normal controls(control group)were selected.The respiratory sounds were sampled with an electronic stethoscope and transmitted to the audit audio software to calculate the relevant data of respiratory sound pressure level(SPL)and the ratio to minute ventilation(VE)of samples with different lung function grades.A retrospective case-control study was used.The data with normal distribution were analyzed by one-way ANOVA,and the pairwise comparison was performed by SNK-q test.Results The breathing audio frequency of mild-moderate,severe COPD patients and the control group has a sound pressure difference of(-0.5±2.2)dB and(-1.6±6.1)dB,(0.7±4.0)dB,and there was significant difference between each group(F=2.64,P=0.038).The sound pressure level per minute ventilation(SPL50-200/VE)of respiratory sound inspiratory phase in COPD patients with mild to moderate and severe obstructive ventilation dysfunction and the control group were(5.7±1.8)dB/L,(6.1±2.3)dB/L and(5.4±0.9)dB/L,respectively.The expiratory SPL50-200/VE were(5.8±1.7),(6.3±2.2)and(5.3±1.2)respectively.There was significant difference among the three groups(Finhalation=3.26,P=0.048;Frespiration=2.44,P=0.045).ROC curve results showed that SPL50-200/VE in the inspiratory phase,SPL50-200/VE in the expiratory phase,and sound pressure difference to assess different grades of obstructive ventilatory dysfunction had diagnostic value(AUC values were 0.697,0.725 and 0.686,respectively;and P values were 0.015,0.005 and 0.022,respectively).In some patients with COPD,abnormally high energy levels may appear in the time-frequency diagram of breath sounds.Conclusion There are significant differences in respiratory audio spectrum analysis between patients with different degrees of obstructive ventilation dysfunction and those with normal ventilation function,which can provide a basis for auxiliary judgment of obstructive ventilation dysfunction.