中文摘要（第一部分）

全身成像三维量化法定量评估主动脉瓣反流：与心脏磁共振的对比研究

目的：探讨全身成像三维量化（general imaging three-dimensional quantification， GI3DQ）法定量评估主动脉瓣反流（aortic regurgitation, AR）严重程度的可行性及准确性，以及GI3DQ法测量的主动脉瓣反流容积（aortic regurgitant volume, ARVol）在识别重度AR中的诊断价值。

方法：纳入2020年11月至2021年9月于我院超声科行二维经胸超声心动图检查诊断为AR的患者共55例，根据主动脉瓣反流方向将患者分为中心性反流组（n=27）和偏心性反流组（n=28）。所有患者进行常规二维、三维经胸超声心动图及心脏磁共振（cardiac magnetic resonance, CMR）检查。使用GI3DQ法直接测量避开二尖瓣前向血流的AR最大反流束体积代表ARVol，以CMR直接法测量的ARVol为参照标准，评估GI3DQ法定量评估中心性及偏心性AR严重程度的准确性；以CMR测量的ARVol分级结果为参照标准，探讨GI3DQ法测量的ARVol在识别重度AR中的诊断价值。

结果：（1）整体人群，GI3DQ法与CMR测量的ARVol之间相关系数为0.88,（P＜0.0001），一致性检验显示GI3DQ法较CMR平均低估ARVol1.0ml，差异无统计学意义（P＞0.05），GI3DQ法和CMR分级结果之间一致性良好，kappa值为0.85；（2）中心性反流组，GI3DQ法与CMR测量的ARVol之间相关系数为0.84（P＜0.0001）；一致性检验显示GI3DQ法较CMR平均低估ARVol0.7ml，差异无统计学意义（P＞0.05），GI3DQ法和CMR分级结果之间一致性良好，kappa值为0.81；（3）偏心性反流组，GI3DQ法和CMR测量的ARVol之间相关系数为0.84（P＜0.0001），一致性检验显示GI3DQ法较CMR平均低估ARVol1.4ml，差异无统计学意义（P＞0.05），GI3DQ法和CMR分级结果之间一致性也良好，kappa值为0.79；（4）ROC曲线分析显示：GI3DQ法测量的ARVol＞59.9ml时，ROC曲线下面积为0.89（P＜0.0001），诊断重度AR的敏感性和特异性均为93.3%；（5）GI3DQ法测量的ARVol观察者自身及观察者间可重复性均良好，组内相关系数分别为0.94、0.93。

结论：

1.对于本研究纳入人群，使用GI3DQ法测量避开二尖瓣前向血流的AR最大反流束体积代替ARVol是可行且可重复的；

2.以CMR为参照方法，GI3DQ法能够对AR患者进行相对准确的定量评估和程度分级，且与反流方向无关；

3.GI3DQ法测量的ARVol在诊断重度AR时有良好的敏感性和特异性。

关键词：全身成像三维量化；实时三维超声心动图；主动脉瓣反流；定量

中文摘要（第二部分）

实时三维超声心动图平面法定量评估主动脉瓣反流:与心脏磁共振的对比研究

目的：探讨实时三维超声心动图（real-time three-dimensional echocardiography，RT3DE）平面法定量评估主动脉瓣反流（aortic regurgitation, AR）严重程度的可行性及准确性，以及RT3DE平面法测量的主动脉瓣反流容积（aortic regurgitant volume, ARVol）在识别重度AR中的诊断价值。

方法：纳入2020年11月至2021年9月于我院超声科行二维经胸超声心动图检查诊断为AR的患者共51例，根据主动脉瓣反流方向将患者分为中心性反流组（n=24）和偏心性反流组（n=27），根据心脏磁共振（cardiac magnetic resonance, CMR）测量的ARVol分级结果将患者分为轻度反流组（n=11）和轻度以上反流组（n=40）。所有患者进行常规二维、三维经胸超声心动图及CMR检查。使用二维近端等速表面（proximal isovelocity surface area, PISA）法及RT3DE平面法分别测量AR有效反流口面积，然后分别与AR速度时间积分相乘获得ARVol。以CMR直接法测量的ARVol为参照标准，评估PISA法和RT3DE平面法测量结果的准确性。以CMR测量的ARVol分级结果为参照，探讨RT3DE平面法测量的ARVol在识别重度AR中的诊断价值。

结果：（1）整体人群，PISA法和RT3DE平面法与CMR测量的ARVol之间相关系数分别为0.85、0.89（P均＜0.0001）；一致性检验显示PISA法较CMR平均高估ARVol0.3ml，差异无统计学意义（P＞0.05），RT3DE平面法较CMR平均高估ARVol5.2ml，差异有统计学意义（P＜0.05），但RT3DE平面法和CMR之间的95%一致性界限较PISA法和CMR之间的95%一致性界限窄；PISA法和RT3DE平面法与CMR分级结果之间kappa值分别为0.76、0.82；（2）中心性反流组，PISA法和RT3DE平面法与CMR测量的ARVol之间相关系数分别为0.65、0.67（P均＜0.001）；一致性检验显示PISA法较CMR平均高估ARVol3.5ml，差异无统计学意义（P＞0.05）,RT3DE平面法较CMR平均高估ARVol6.9ml，差异有统计学意义（P＜0.05），但RT3DE平面法和CMR之间的95%一致性界限也较PISA法和CMR之间的95%一致性界限窄；PISA法和RT3DE平面法与CMR分级结果之间kappa值分别0.70、0.74；（3）偏心性反流组，PISA法和RT3DE平面法与CMR测量的ARVol之间相关系数均为0.84（P均＜0.0001）；一致性检验显示PISA法较CMR平均低估ARVol2.7ml，RT3DE平面法较CMR平均高估ARVol3.8ml，差异均无统计学意义（P均＞0.05）；PISA法和RT3DE平面法与CMR分级结果之间kappa值分别为0.68、0.77；（4）轻度反流组，一致性检验显示RT3DE平面法较CMR平均高估ARVol15.7ml，差异有统计学意义（P＜0.05），两种方法分级结果一致率为45.5%；轻度以上反流组，一致性检验显示RT3DE平面法较CMR平均高估ARVol2.3ml，差异无统计学意义（P＞0.05），两种方法分级结果之间kappa值为0.86；（5）ROC曲线分析显示：RT3DE平面法测量的ARVol＞58.7ml时，ROC曲线下面积为0.93（P＜0.0001），诊断重度AR的敏感性为100.0%，特异性为86.7%；（6）RT3DE平面法测量ARVol的观察者自身及观察者间可重复性均良好，组内相关系数分别为0.89、0.87。

结论：

1.对于本研究纳入人群，使用RT3DE平面法测量ARVol是可行且可重复的；

2.以CMR为参照方法，RT3DE平面法对轻度AR患者出现了明显高估，但对轻度以上AR患者能够进行相对准确的定量评估和程度分级，且RT3DE平面法测量的ARVol在诊断重度AR时有良好的敏感性和特异性；

3.RT3DE平面法评估AR严重程度分级的准确性优于传统二维PISA法。

关键词：实时三维超声心动图；平面；主动脉瓣反流；定量

中文摘要（第三部分）

动态心脏模型成像联合三维自动右室定量评估主动脉瓣反流：与心脏磁共振的对比研究

目的：探讨动态心脏模型成像（Dynamic HeartModel, DHM）联合三维自动右室定量（3D Auto RV）法在定量评估主动脉瓣反流（aortic regurgitation, AR）患者心脏大小、功能及主动脉瓣反流容积（aortic regurgitant volume, ARVol）中的可行性及准确性，以及DHM联合3D Auto RV法测量的ARVol在识别重度AR中的诊断价值。

方法：纳入2020年11月至2021年9月于我院超声科行二维经胸超声心动图检查初步诊断为轻度以上单纯AR的患者42例。所有患者进行常规二维、三维经胸超声心动图及心脏磁共振（cardiac magnetic resonance, CMR）检查。使用双平面Simpson法及DHM分别测量左心室（left ventricular, LV）舒张末期容积（end-diastolic volume, EDV）、收缩末期容积（end-systolic volume, ESV）、每搏量（stroke volume, SV）及射血分数（ejection fraction, EF），使用3D Auto RV测量右心室（right ventricular, RV）EDV、ESV、SV及EF，然后通过计算DHM测量的LVSV和3D Auto RV测量的RVSV的差值得到ARVol。以CMR为参照方法，评估DHM联合3D Auto RV法各测量结果的准确性；并以CMR直接法测量的ARVol分级结果为参照，探讨DHM联合3D Auto RV法测量的ARVol在识别重度AR中的诊断价值。

结果：（1）左心室容积和功能：双平面Simpson法测量的LVEDV、LVESV、LVSV和LVEF与CMR测量结果之间相关系数分别为0.83、0.82、0.80、0.75（P均＜0.0001），一致性检验显示双平面Simpson法较CMR明显低估了LVEDV、LVESV及LVSV，高估了LVEF，差异均有统计学意义（P均＜0.0001）；DHM测量的LVEDV、LVESV、LVSV和LVEF与CMR测量结果之间相关系数分别为0.87、0.81、0.83、0.73（P均＜0.0001），一致性检验显示DHM也较CMR低估了LVEDV、LVESV及LVSV，高估了LVEF，差异均有统计学意义（P均＜0.0001），但DHM和CMR之间的平均偏差小于双平面Simpson法与CMR之间的平均偏差；（2）右心室容积和功能：3D Auto RV测量的RVEDV、RVESV、RVSV和RVEF与CMR测量结果之间相关系数分别为0.82、0.79、0.81、0.74（P均＜0.0001），一致性检验显示3D Auto RV较CMR低估了RVEDV、RVESV及RVSV，高估了RVEF，差异均具有统计学意义（P均＜0.05）；（3）DHM联合3D Auto RV法测量的ARVol与CMR测量结果之间相关系数为0.89（P＜0.0001），一致性检验显示DHM联合3D Auto RV法较CMR平均高估ARVol0.8ml，差异无统计学意义（P＞0.05），DHM联合3D Auto RV法和CMR分级结果之间的kappa值为0.81；（4）ROC曲线分析显示：DHM联合3D Auto RV法测量的ARVol＞62.1ml时，ROC曲线下面积为0.95（P＜0.0001），诊断重度AR的敏感性为100%，特异性为93.3%；（5）DHM联合3D Auto RV测量LVEDV、LVESV、LVSV、LVEF、RVEDV、RVESV、RVSV、RVEF及ARVol的观察者自身及观察者间可重复性均良好，组间相关系数均≥0.89。

结论：

1.对于本研究纳入人群，DHM联合3D Auto RV对于左、右心室容积和EF及ARVol的测量是可行且可重复的；

2.以CMR为参照方法，DHM和3D Auto RV能够对左、右心室容积和功能进行相对准确的测量，但存在对左、右心室容积一定程度的低估，和对左、右心室EF一定程度的高估，DHM准确性仍优于常规双平面Simpson法；

3.以CMR为参照方法，DHM联合3D Auto RV法能够对轻度以上单纯AR进行相对准确的定量评估和程度分级，且DHM联合3D Auto RV法测量的ARVol在诊断重度AR时有良好的敏感性和特异性。

关键词：动态心脏模型成像；三维自动右室定量；实时三维超声心动图；主动脉瓣反流；定量

Abstract (Part 1)

Quantitative evaluation of aortic regurgitation by general imaging three-dimensional quantification: a comparison with cardiac magnetic resonance

Objectives: To investigate the feasibility and accuracy of general imaging three-dimensional quantification (GI3DQ) in the quantitative evaluation of aortic regurgitation (AR) severity and discuss the diagnostic value of the GI3DQ-derived aortic regurgitant volume (ARVol) in identifying severe AR.

Methods: A total of 55 patients who were diagnosed with AR by two-dimensional transthoracic echocardiography in our Ultrasonography Department from November 2020 to September 2021 were included and divided into central (n=27) and eccentric (n=28) AR groups according to the direction of the AR jet. All patients underwent routine two-dimensional and three-dimensional transthoracic echocardiography and cardiac magnetic resonance (CMR) examinations. The volume of maximum AR jet which is not mixed with mitral froward flow was measured by GI3DQ as ARVol. Taking the ARVol measured by CMR direct method as the reference standard, the accuracy of the ARVol measured by GI3DQ method was evaluated. And using the grading results according to the ARVol measured by CMR as the reference standard, the diagnostic value of the ARVol measured by GI3DQ in identifying severe AR was evaluated.

Results: (1) For the entire population, the GI3DQ method demonstrated a good correlation and agreement with CMR in ARVol measurement (r=0.88, P＜0.0001; mean difference=-1.0ml, P＞0.05). The kappa agreement between GI3DQ and CMR in grading AR severity was good (k=0.85). (2) In central and eccentric AR groups, GI3DQ method also had good correlations (central, r=0.84; eccentric, r=0.84; both P＜0.0001) and agreements (central, mean difference=-0.7ml; eccentric, mean difference= -1.4ml; both P＞0.05) with CMR in ARVol measurement. The kappa agreements between GI3DQ and CMR in grading AR severity in both groups were good (central, k=0.81; eccentric, k=0.79). (3) As assessed by receiver operating characteristic curve analysis, the ARVol derived from GI3DQ method at a cutoff value of 59.9 ml yielded 93.3% sensitivity and 93.3% specificity in identifying severe AR (area under the curve=0.89, P＜0.0001). (4) The intraobserver and interobserver reproducibility of the ARVol derived from GI3DQ method were good (ICC=0.94 and 0.93, respectively).

Conclusions:

1.Measuring the maximum AR jet volume which is not mixed with mitral froward flow by GI3DQ method to replace the ARVol is feasible and reproducible in this study population.

2.Taking CMR as the reference method, GI3DQ method can make relatively accurate measurement of the ARVol and correct grade of AR severity no matter what is the direction of the AR jet.

3.The ARVol derived from the GI3DQ method can identify severe AR with good sensitivity and specificity.

Key words: general imaging three-dimensional quantification; real-time three-dimensional echocardiography; aortic regurgitation; quantification

Abstract (Part 2)

Quantitative evaluation of aortic regurgitation by real-time three-dimensional echocardiography planimetry method: a comparison with cardiac magnetic resonance

Objectives: To investigate the feasibility and accuracy of real-time three-dimensional echocardiography (RT3DE) planimetry method in the quantitative evaluation of aortic regurgitation (AR) severity and discuss the diagnostic value of the aortic regurgitant volume (ARVol) derived from RT3DE planimetry method in identifying severe AR.

Methods: A total of 51 patients who were diagnosed with AR by two-dimensional transthoracic echocardiography in our Ultrasonography Department from November 2020 to September 2021 were included and divided into central (n=24) and eccentric (n=27) AR groups according to the direction of the AR jet as well as mild AR group (n=11) and more than mild AR group (n=40) according the grading results based on the ARVol measured by cardiac magnetic resonance (CMR). All patients underwent routine two-dimensional and three-dimensional transthoracic echocardiography and CMR examinations. The effective regurgitant orifice area was measured by the proximal isovelocity surface area (PISA) method and RT3DE planimetry method respectively, and then multiplied by the AR velocity time integral to obtain ARVol. Taking the ARVol measured by CMR direct method as the reference standard, the accuracy of the ARVol derived from the PISA and RT3DE planimetry method was evaluated. And using the grading results based on the ARVol measured by CMR as the reference standard, the diagnostic value of the ARVol derived from RT3DE planimetry method in identifying severe AR was evaluated.

Results: (1) For the entire population, the PISA and RT3DE planimetry method correlated well with CMR in ARVol measurement (r=0.85 and 0.89, respectively; both P＜0.0001). Compared with CMR, PISA method made a mild overestimation (mean difference=0.3ml, P＞0.05), while RT3DE planimetry method made a significant overestimation (mean difference=5.2ml, P＜0.05) in ARVol measurement based on consistency analysis. But the 95% limits of agreement between the RT3DE planimetry method and CMR was smaller than that between the PISA method and CMR. The kappa agreements between PISA method and CMR，and between RT3DE planimetry method and CMR in AR severity grading were good (k=0.76 and 0.82, respectively). (2) In central AR group, the PISA and RT3DE planimetry method correlated well with CMR in ARVol measurement (r=0.65 and 0.67, respectively; both P＜0.001). Compared with CMR, PISA method made a mild overestimation (mean difference=3.5ml, P＞0.05), while RT3DE planimetry method made a significant overestimation (mean difference=6.9ml, P＜0.05) in ARVol measurement based on consistency analysis. The 95% limits of agreement between the RT3DE planimetry method and CMR was also smaller than that between the PISA method and CMR. The kappa agreements between PISA method and CMR，and between RT3DE planimetry and CMR in AR severity grading were good (k=0.70 and 0.74, respectively). (3) In eccentric AR group, the PISA and RT3DE planimetry method also correlated well with CMR in ARVol measurement (both r=0.84; both P＜0.0001). Compared with CMR, PISA method made a mild underestimation (mean difference=-2.7ml, P＞0.05), while RT3DE planimetry method made a mild overestimation (mean difference=3.8ml, P＞0.05) in ARVol measurement based on consistency analysis. The kappa agreement between RT3DE planimetry method and CMR in AR severity grading was good (k=0.77)，while that between PISA method and CMR was inferior (k=0.68). (4) In mild AR group, RT3DE planimetry method made a significant overestimation (mean difference=15.7ml, P＜0.05) compared with CMR in ARVol measurement based on consistency analysis. The grading consistency between these two methods was 45.5%. In more than mild AR group, RT3DE planimetry method only made a mild overestimation (mean difference=2.3ml, P＞0.05) compared with CMR in ARVol measurement based on consistency analysis. The kappa agreement between these two methods was good (k=0.86). (5) As assessed by receiver operating characteristic curve analysis, the ARVol derived from RT3DE planimetry method at a cutoff value of 58.7ml yielded 100.0% sensitivity and 86.7% specificity in identifying severe AR (area under the curve=0.93, P＜0.0001). (6) The intraobserver and interobserver reproducibility of the ARVol derived from RT3DE planimetry method were good (ICC=0.89 and 0.87, respectively).

Conclusions:

1.ARVol measurement using RT3DE planimetry method is feasible and reproducible in this study population.

2.Taking CMR as the reference method, in mild AR group，RT3DE planimetry method make a significant overestimation of ARVol, while in more than mild AR group, RT3DE planimetry method can make relatively accurate measurement of the ARVol and correct grade of AR severity. The ARVol derived from the RT3DE planimetry method can identify severe AR with good sensitivity and specificity.

3.The accuracy of RT3DE planimetry method in the grading of AR severity is better than that of conventional two-dimensional PISA method.

Key words: real-time three-dimensional echocardiography; planimetry； aortic regurgitation; quantification

Abstract (Part 3)

Evaluation of aortic regurgitation by Dynamic HeartModel plus 3D Auto RV: a comparison with cardiac magnetic resonance

Objectives: To investigate the feasibility and accuracy of Dynamic HeartModel (DHM) plus 3D Auto RV method in the measurement of left and right ventricular volumes, function and aortic regurgitant volume (ARVol) for patients with aortic regurgitation (AR)，and discuss the diagnostic value of the ARVol derived from DHM plus 3D Auto RV method in identifying severe AR.

Methods: A total of 42 patients who were initially diagnosed with more than mild isolated AR by two-dimensional transthoracic echocardiography in our Ultrasonography Department from November 2020 to September 2021 were included. All patients underwent routine two-dimensional, three-dimensional transthoracic echocardiography and cardiac magnetic resonance (CMR) examinations. Left ventricular (LV) end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV) and ejection fraction (EF) were measured by biplane Simpson’s method and DHM respectively. Right ventricular (RV) EDV, ESV, SV and EF were measured by 3D Auto RV. The difference between the LVSV and RVSV measured by DHM and 3D Auto RV was calculated as ARVol. Taking CMR as the reference method, the accuracy of DHM and 3D Auto RV measurement results was analyzed. And using the grading results according to the ARVol measured by CMR direct method as the reference standard, the diagnostic value of the ARVol measured by DHM plus 3D Auto RV method in identifying severe AR was evaluated.

Results: (1) LV volumes and function: There were good correlations between the LVEDV, LVESV, LVSV and LVEF measured by biplane Simpson’s method and CMR respectively (r=0.83, 0.82, 0.80 and 0.75, respectively; all P＜0.0001). Consistency analysis showed that the Biplane Simpson’s method made significant underestimations of LVEDV, LVESV and LVSV (all P＜0.0001) and a significant overestimation of LVEF (P＜0.0001) compared with CMR. There were also good correlations between the LVEDV, LVESV, LVSV and LVEF measured by DHM and CMR respectively (r=0.87, 0.81, 0.83 and 0.73, respectively; all P＜0.0001). Consistency analysis showed that DHM also made underestimations of LVEDV, LVESV and LVSV (all P＜0.0001) and a overestimation of LVEF (P＜0.0001) compared with CMR. But the mean difference between DHM and CMR was smaller than that between biplane Simpson’s method and CMR. (2) RV volumes and function: There were good correlations between the RVEDV, RVESV, RVSV and RVEF measured by 3D Auto RV and CMR respectively (r=0.82, 0.79, 0.81 and 0.74, respectively; all P＜0.0001). Consistency analysis showed that 3D Auto RV made underestimations of RVEDV, RVESV and RVSV (all P＜0.05) and a overestimation of RVEF (P＜0.05) compared with CMR. (3) DHM plus 3D Auto RV method correlated well with CMR in ARVol measurement (r=0.89, P＜0.0001), with no significant difference based on consistency analysis (mean difference=0.8ml, P＞0.05). The kappa agreement between DHM plus 3D Auto RV method and CMR in grading CAR severity was good (k=0.81). (4) As assessed by receiver operating characteristic analysis, ARVol measured by DHM plus 3D Auto RV method at a cutoff value of 62.1ml yielded 100.0% of sensitivity and 93.3% of specificity in identifying severe AR (area under the curve=0.95, P＜0.0001). (5) The intraobserver and interobserver reproducibility of the LVEDV, LVESV, LVSV, LVEF, RVEDV, RVESV, RVSV, RVEF and ARVol measured by DHM plus 3D Auto RV method were all good (all ICC≥0.89).

Conclusions:

1.Left and right ventricular volumes, EF and ARVol measurement using DHM plus 3D Auto RV method is feasible and reproducible in this study population.

2.Taking CMR as the reference method, DHM and 3D Auto RV can make relatively accurate measurement of left and right ventricular volumes and EF, but still with underestimations of left and right ventricular volumes and overestimations of left and right ventricular EF to some extent. However, the accuracy of the measurement results by DHM is better than that by conventional biplane Simpson’s method.

3.Taking CMR as the reference method, DHM plus 3D Auto RV method can make relatively accurate measurement of the ARVol and correct grade of AR severity. The ARVol measured by DHM plus 3D Auto RV method can identify severe AR with good sensitivity and specificity.

Key words: Dynamic HeartModel; 3D Auto RV; real-time three-dimensional echocardiography; aortic regurgitation; quantification