Incremental value of restingthree-dimensional speckle-tracking echocardiography in detecting coronaryartery disease





The aim of thepresent study was to investi­gate the incremental value of resting three-dimensionalspeckle-trackingechocardiography (3D-STE) in the detection of early‑stage left ventriculardysfunction in patients with coro­nary artery disease (CAD). A total of 110patients suspected of having CAD were recruited. All patients underwent 3D‑STEand coronary artery angiography (CAG). They were divided to a CAD group and anormal group according to the results of CAG. Using 3D‑STE software, the peakvalues of longitudinal strain (LS), circumferential strain (CS), radial strain(RS) and area strain (AS) and the time to peak value of these strains (T‑LS, T‑CS,T‑RS and T‑AS) were measured. A receiver operator characteristic curve (ROC)was used to analyze the sensitivity of these strains for the diagnosis of CAD.ROC analysis indicated that T‑LS and composite indices combining the peakstrain value and time to peak of LS, CS and AS have diagnostic value for theearly detection of CAD; the area under the curve (AUC) values were 0.667,0.692, 0.621 and 0.672 respectively (P<0.005). The composite index oflongitudinal strain demonstrated the highest diagnostic value for CAD with 62%sensitivity and 76% specificity. These results indicate that 3D-STE hasincremental value for the diagnosis of CAD in patients at rest.





Coronary arterydisease (CAD) is the most common cardio­vascular disease, and its incidenceincreases in the elderly (1). Early detection and intervention for leftventricular (LV) dysfunction are of great importance in patients with CAD (2).Echocardiography is noninvasive, inexpensive and convenient. The detection ofpatients with CAD by this method is likely to improve the benefit‑risk ratio inpatients.


During recentyears, strain and strain rate imaging echo­cardiography have emerged asvaluable tools that provide a comprehensive and reliable assessment ofmyocardial function. However, 2D‑STE has intrinsic limitations, such as longexamination times, a geometric assumption of LV morphology, and mistracking ifspeckles move out of the scanning plane.


three‑dimensionalspeckle‑tracking echocar­diography (3D‑STE) is free of geometric assumptionsand speckles moving out of the scanning plane. In the present study, the aimwas to investigate the incremental value of resting 3D‑STE in the detection ofLV dysfunction in patients with CAD and to identify sensitive indicators forthe detection of CAD.



目前,应变(StrainS)和应变率(Strain RateSR)成为较全面、可靠的评价心肌运动的重要指标。已经有研究证实了二维斑点追踪技术可以评价左室收缩功能。但二维‑STE具有局限性,如检查时间长,LV立体形态假设,及失追踪等。三维斑点追踪技术可以避免部分弊端。本研究的目的是探讨三维斑点追踪技术检测早期冠心病患者静息状态下左室功能不全的价值及确定诊断冠心病的敏感性指标

Patients and methods

Study population

Thestudy evaluated 110 patients (68 males; age, 63±9 years) who were suspected ofhaving CAD and who were referred for coronary artery angiography (CAG) atBeijing Hospital (Beijing, China) between June 2010 and August 2012.

Echocardiographicimage acquisition and analyses

Echocardiographicdata were acquired with a ultrasound system, which was equipped with a 2Dtransducer (frequency, 1‑5 MHz), 3D volume transducer, 3D speckle‑trackinganalysis software and a background processing workstation. Firstly, general 2Dultrasonic detectors were used to collect images of the para­sternal long axis,short axis and apical four‑chamber view. Secondly, following conversion to the3D volume transducer with an apical four‑chamber view, the pre‑4D mode wasselected to obtain a clear image of the LV endocardium. Then, as the full‑4Dmode was entered, the patient was asked to hold their breath at the end ofexpiration.


Thestored dynamic cardiac images were extracted in the off‑line state. Thisinvolved entering ‘3DT (three‑dimensional speckle tracking) mode’, adjustingthe baseline and angle, clearly displaying the image of the endocardium, mitralannulus and apex, and speckle assignment on plane A and plane B separately. Theinstrument then speckle‑tracked automatically when the ‘start’ button waspressed. The curve of the endocardium was drawn, and the software analyzed thedata and calculated the peak value strains (LS), (CS), (RS) (AS) of 16 sectionsof the left ventricle and the time to peak value of the strains.

Observational indices

Strainvalues and the time to peak value of the strain for 16 local myocardialsegments from the 3D‑STI were observed. The strain values were RS, CS, LS and AS.The time to peak values for the strains were (T‑RS), (T‑CS), (T‑LS) and (T‑AS).

Statistical analysis

The observation indices were numerical variable data. Afterestablishing the database, with the use of Excel (Microsoft Corporation,Redmond, WA, USA) and SPSS version 16.0, binary logistic regression analysisand receiver operator character­istic curves (ROCs) were used to analyze thesensitivity of the strains for the diagnosis of the CAD.


20106月至20128月于卫生部北京医院心内科就诊疑似冠心病并拟行CAG术的患者1 10例,其中男性68例,女性42例,年龄(632±89)岁,身高(1664±78)cm,体重(727±102)kg。排除标准:心律失常;陈旧心肌梗死病史、既往有PCI治疗史;心功能衰竭者;心肌病、严重瓣膜病、先心病史等其它心脏病变,三维图像质量差,无法进一步分析者


研究仪器采用日本东芝Toshiba SSH880H心血管专用彩色超声诊断仪,配有二维探头(频率为1-5MHz)和三维容积探头(频率-5MHz)3DT三维斑点追踪分析软件和后台处理工作站。首先对所有受检者同步连接心电图,嘱受检者左侧卧位,使用常规二维探头(频率为l5MHz)采集胸骨旁左室长轴、短轴、心尖四腔心、两腔心、三腔心等切面图像。(2)在心尖四腔心切面换用三维容积探头(探头频率卜5MHz),选择pre4D模式调整图像至左心室内膜显像清晰后,进入全容积(full4D)模式,嘱受检者呼气末屏气,成像角度为(7080)×(70-80),采集并存储心脏三维全容积动态图。






上述各观察指标均为计量资料,采用EXCELSPSS 160统计分析软件建立相应数据库并行统计学分析。计量资料用均数±标准差表示,多组间比较采用单因素方差分析,运用Logistic二元回归分析和ROC曲线对各指标诊断冠心病的敏感性进行分析。PO05表示具有统计学差异。


Resultsof CAG

According to the CAG results, of the 110 patients enrolled, 68 exhibitedCAD (levels 1-4), which were assigned to the CAD group. The other patients(level 0) were designated the non‑CAD group .


3D‑STE evaluation of LV local function in different gradedstenoses of coronary arteries. Generally, it was observed that the peak valueof strain of local segments of the left ventricle decreased as the stenosisdegree of the supplied vessel increased, the higher the degree of stenosis thelonger the time to the peak value. In addition, these relationships were mostevident at the highest graded degree of stenosis. For example, the peak valueof AS of certain segments supplied by the left circumflex (LCX) arteryexhibited a reduction in differently graded coronary stenosis groups. Thehigher the grade of stenosis the lower the peak value of AS . In addition, theT‑LS of certain segments supplied by the left anterior descending arteryascended in differently graded coronary stenosis groups .

ROCcurve and sensitive indicators

In this study, three kinds of ROC were generated. One was for thepeak strains RS, CS, LS and AS (Fig. 3), one was for the time to the peakstrain, which comprised T‑RS, T‑CS, T‑LS and T‑AS (Fig. 4), and the third wasfor the composite index. The composite indices were designated RS1, CS1, LS1and AS1


Through the ROC analysis, an area under the curve (AUC) wasdetermined for each index.


Analysis of the ROC curves and AUCs, revealed that each individualpeak value of strain, namely RS, CS, LS and AS, had no diagnostic value for theearly detection of CAD (P>0.05). Only T‑LS had statistical significance inthe diagnosis of CAD.


However, the composite index of peak strain and time to peakstrain had significantly higher sensitivity for the detection of CADand wasfound to have 62% sensitivity and 76% specificity for the detection of CAD.

Excellentdiagnostic value and individual prediction

Through the analysis of the ROC of LS1, it was determined that themaximum Youden index was 0.38 and the excellent diagnostic probability value (Pexce) of the LS1 was Pexce=0.615.Pexceis an ROC‑derived statistical index thatindicates the likelihood of a positive diagnosis.


Pindiis aprobability value that can be used to predict whether the individual patient islikely to have the disease or not, and can be compared with Pexce. For calculating PindiofLS1, this study established a binary logistic regression equation throughlogistic regression analysis.












将所有可能的切点作为阈值进行灵敏度和特异度计算后,以敏感性为纵坐标代表真阳性率,(1—特异性)为横坐标代表假阳性率,作图绘成ROC曲线。一般以Youden指数(约登指数)最大,即(灵敏度+特异度一1)达到最大所对应的值为最佳诊断界值。关于预测概率值的计算,需要首先建立Logistic回归方程。用SPSSBinary Logistic方程对变量LS峰值和LS达峰时间进行Logistic回归分析,可得到两变量的回归系数,进而求得Logistic回归方程。然后将各变量值(LS峰值和LS达峰时间)代入上述方程,即可得出个体预测概率值P,与最佳诊断界值P=0615进行比较,即可早期判断患者是否为冠心病。


Speckle‑tracking strain imaging has been demonstrated to be anappropriate method for the examination of global and regional LV functionalproperties (15,16), which was confirmed in the present study. In this study, itwas found that changes in the peak value and time to peak value of certain strainsaccorded with the physiological characteristics of the left ventricle.Secondly,through the ROC analysis of each parameter measured by 3D‑STE

By convention, myocardial deformation is divided into threedirections (17). LS and CS are considered to reflect the contraction oflongitudinally arranged endocardial fibers and circumferentially arrangedmidlayer fibers, respectively. In general, the longitudinally arrangedsubendocardial fibers are more vulnerable due to their direct exposure to theintraventricular blood pressure and the anatomy of the coro­nary circulation(18,19). As a result, longitudinal function is impaired first in many diseases,including CAD.

In the present study, it was demonstrated that resting 3D‑STE hadincremental value for the prompt detection of coronary lesions and a logisticregression equation was estab­lished

In addition, an evaluation model may be established through theaddition of risk factors such as hyper­tension, diabetes, smoking andhyperlipidemia.