自制心肌超声造影定量分析软件的设计与评价

Design and evaluation of a quantitative analysis software for myocardial contrast echocardiography

目的评价自制心肌超声造影(MCE)图像定量分析软件对心肌灌注研究的可行性。方法根据阻断和再灌注冠状动脉时间不同,家兔被分为两组:阻断15min再灌注30min(Ⅰ组)和阻断120min再灌注60min(Ⅱ组)。分别于基础状态、冠状动脉阻断时和再灌注后行MCE,应用基于淘汰粒子群优化(EPSO)聚类算法的自制计算机图像定量分析软件对造影图像进行自动处理,获得多个心肌灌注定量参数。结果 (1)阻断时,Ⅰ组与Ⅱ组危险心肌的标化CI值均明显减低,与基础状态比较差异有统计学意义(t=5.104和t=4.327,P<0.01)。再灌注后,Ⅰ组危险心肌标化CI值比阻断时增高(t=2.933,P<0.01),与基础状态时相比仍减低(t=2.653,P<0.01),Ⅱ组危险心肌标化CI值与阻断时相比差异无统计学意义(P>0.05);(2)Ⅱ组危险心肌红色编码区和TTC中梗死心肌面积分别为(21.4±12.3)%和(18.0±9.5)%,且与TTC结果呈正相关(r=0.89,P<0.01);(3)阻断时,危险心肌节段直方图呈偏态分布,再灌注后,Ⅰ组危险心肌基本恢复至正态分布,而Ⅱ组危险心肌仍呈偏态分布。结论基于淘汰粒子群优化聚类算法的心肌超声图像分析软件在定量评价心肌微灌注和识别灌注异常方面具有较好的可行性和较高的应用价值。

Objective To evaluate the feasibility of quantitative analysis software for myocardial contrast echocardiography （MCE） in assessment of myocardial perfusion. Methods According to coronary occlusion and reperfusion at different times,rabbits were divided into two groups：15 min occlusion/30 min reperfusion （group Ⅰ） and 120 min occlusion/60 min reperfusion （group Ⅱ）. MCE was performed on all rabbits at baseline,occlusion and after reperfusion,and its images were analyzed by a new quantitative analysis software based on eliminating particle swarm optimization （EPSO） clustering algorithm,by which obtain myocardial perfusion parameters. Results （1） The values of calibrated contrast intensity （CI） in risk segments of Groups Ⅰ and Ⅱ were significantly lower than those at baseline during occlusion （t =5.104 and t =4.327,P0.01）. After reperfusion,calibrated CI in risk segments significantly improved in Group Ⅰ （t =2.933,P0.01） while those remained unchanged in Group Ⅱ （P0.05）. （2） The areas of red-coded region in color-coded map and myocardial infarction in triphenyl-tetrazolium chloride （TTC） were （21.4±12.3）% and （18.0±9.5）%,respectively. The correlation between color-coded image and TTC was 0.89 （P0.01）. （3） The histogram in all risk segments was skew distribution during occlusion. After reperfusion,the histogram in Group Ⅰ was normal distribution while that was still skewed distribution in Group Ⅱ. Conclusion The MCE image analysis software based on EPSO clustering algorithm in the quantitative assessment of myocardial microperfusion and identification of myocardial perfusion abnormalities was feasible and of high value.