切换至 "中华医学电子期刊资源库"

中华医学超声杂志(电子版) ›› 2022, Vol. 19 ›› Issue (12) : 1329 -1335. doi: 10.3877/cma.j.issn.1672-6448.2022.12.003

心血管超声影像学

基于人工智能的三维超声心动图定量评估右心室功能的价值
包雨微1, 朱英1, 郑康超1, 周玮1, 刘娅妮1,(), 邓又斌1   
  1. 1. 430030 武汉,华中科技大学同济医学院附属同济医院超声医学科
  • 收稿日期:2021-05-21 出版日期:2022-12-01
  • 通信作者: 刘娅妮

Value of artificial intelligence based three-dimensional echocardiography in quantitative evaluation of right ventricular function

Yuwei Bao1, Ying Zhu1, Kangchao Zheng1, Wei Zhou1, Yani Liu1,(), Youbin Deng1   

  1. 1. Department of Medical Ultrasound, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, China
  • Received:2021-05-21 Published:2022-12-01
  • Corresponding author: Yani Liu
引用本文:

包雨微, 朱英, 郑康超, 周玮, 刘娅妮, 邓又斌. 基于人工智能的三维超声心动图定量评估右心室功能的价值[J/OL]. 中华医学超声杂志(电子版), 2022, 19(12): 1329-1335.

Yuwei Bao, Ying Zhu, Kangchao Zheng, Wei Zhou, Yani Liu, Youbin Deng. Value of artificial intelligence based three-dimensional echocardiography in quantitative evaluation of right ventricular function[J/OL]. Chinese Journal of Medical Ultrasound (Electronic Edition), 2022, 19(12): 1329-1335.

目的

探讨基于人工智能(AI)的三维超声心动图(3DE)算法在定量评估右心室(RV)功能中的价值。

方法

连续性收集自2021年1月至2021年2月在华中科技大学同济医学院附属同济医院就诊且行心脏磁共振(CMR)的患者51例,CMR检查后24 h内进行3DE检查,使用基于AI的3DE软件自动分析右心室(RV)功能参数,记录每位患者所有参数值及分析时间。对RV 3DE图像质量进行分级评估(分为优、中等、差)。以CMR测量的RVEF(CMR-RVEF)为金标准,比较AI-3DE计算的RVEF与CMR结果的相关性及一致性;以CMR-RVEF<45%为RV收缩功能减低,比较AI-3DE所获取的定量参数在RV功能正常与功能减低组间的差异,并绘制ROC曲线分析各参数对RV收缩功能异常的诊断价值。

结果

51例患者均应用AI-3DE完成了RV功能的多参数定量分析,19例(19/51,37.3%)未经手动校正的患者,全自动分析平均时间为(15±1)s;32例需要手动校正的患者中,AI-3DE RV分析时间在图像质量优、中等、差的患者中分别为(100±12)、(105±6)、(162±3)s,三者比较差异有统计学意义(F=1954.9,P<0.05)。CMR与AI-3DE测量的RVEF分别为51%(44%,54%)和47%(43%,50%),2种方法的测值差异无统计学意义(P>0.05)。相关性分析散点图显示,2种方法测量结果呈一定程度的正相关关系(r=0.735,P<0.05)。Bland-Altman图表明,2种方法获得的差异分布在平均值附近,二者具有良好的一致性。基于AI-3DE的RVEF、三尖瓣环收缩期位移(TAPSE)、右心室面积变化分数(FAC)、室间隔长轴应变(LS)和RV游离壁LS在RV功能正常组与功能减低组之间差异均有统计学意义(P均<0.05)。ROC曲线显示,基于AI-3DE的RVEF诊断RV功能减低的ROC曲线下面积为0.854,最佳截断值为43%,敏感度为94.4%,特异度为66.6%。

结论

基于AI的3DE算法能够快速、准确地获得反映RV功能的多参数指标,这些定量参数尤其是RVEF对RV功能减低具有良好的诊断价值。

Objective

To evaluate the value of a new three-dimensional echocardiography (3DE) algorithm based on artificial intelligence (AI) for quantitative assessment of right ventricular (RV) function.

Methods

Fifty-one patients who underwent cardiac magnetic resonance (CMR) examination at Tongji Hospital affiliated to Tongji Medical College of Huazhong University of Science and Technology from January 2021 to February 2021 were enrolled. 3DE images were obtained within 24 hours, and AI based RV 3DE software (AI-3DE) was used to automatically analyze and measure the RV functional parameters. RV 3DE image quality was graded, and the process and time of analyzing RV function by AI software under different image quality were recorded. The RV ejection fraction (RVEF) measured by CMR (CMR-RVEF) was regarded as the gold standard. And RVEF measured by AI-3DE was compared to the CMR results. CMR-RVEF<45% was defined as reduced RV systolic function. The differences between the quantitative parameters obtained by AI-3DE in the normal RV function and reduced RV function groups were compared, and the diagnostic performance of each parameter in identifying RV dysfunction was assessed by receiver operating characteristic (ROC) curve analysis.

Results

Multi-parameter quantitative analysis of RV function was completed in all patients by AI-3DE, of which 19 patients (37.3%) achieved fully RV automatic quantitative analysis within 15±1 s. The RV endometrium tracing needed to be adjusted manually in 32 patients (62.7%). And the analysis time was 100±12 s, 105±6 s, and 162±3 s in the patients with good, moderate, and poor image quality, respectively (F=1964.9, P<0.05). Both in the overall population and in patients with moderate and poor image quality, the RVEF measured by AI-3DE was statistically correlated with the results of CMR analysis (r=0.735, P<0.05). Bland-Altman curve showed that the RVEF measured by AI-3DE was slightly underestimated compared with CMR, but they had a good consistency. RV function parameters measured by AI-3DE, including RVEF, tricuspid ring systolic displacement (TAPSE), area change fraction (FAC), and longitudinal strain of septum and right ventricular free wall (LS), showed excellent diagnostic performance in identifying the reduced RV systolic function (P<0.05). The cutoff value of RVEF measured by AI-3DE for the diagnosis of RV dysfunction was 43%, with a sensitivity of 94.4% and specificity of 66.6%.

Conclusion

The right ventricular 3DE algorithm based on AI can quickly and accurately evaluate the RV function, providing multiple quantitative parameters for the identification of RV dysfunction.

图1 基于人工智能的三维超声心动图(AI-3DE)多参数定量分析右心室功能。图a为右心室非容积功能参数测量;图b为右心室容积参数及右心室射血分数(RVEF)测量
图2 基于人工智能的三维超声心动图与心脏磁共振测量的右心室射血分数相关性分析散点图注:CMR为心脏磁共振;AI-3DE为基于人工智能的三维超声心动图;RVEF为右心室射血分数
图3 基于人工智能的三维超声心动图与心脏磁共振测量的右心室射血分数的Bland-Altman分析图。图中红色虚线表示平均偏差值,上下两端的黑色虚线表示95%的一致性界限,中间的黑色虚线表示差值为0的水平。如图所示,多数心脏磁共振和三维超声心动图的差异值都在95%一致性界限内注:3DE为三维超声心动图;CMR为心脏磁共振;RVEF为右心室射血分数;RVEF平均值为(3DE测值+CMR测值)/2;RVEF差值为3DE测值-CMR测值
表1 RV功能正常组与功能减低组一般资料及AI-3DE测量参数的比较
表2 基于AI-3DE的RV功能参数诊断RV功能减低的ROC曲线分析
表3 AI-3DE测量RV功能参数的观察者间和观察者内一致性
1
Badano LP, Addetia K, Pontone G, et al. Advanced imaging of right ventricular anatomy and function [J]. Heart, 2020, 106(19): 1469-1476.
2
Mrad AK, Burri M, Cleuziou J, et al. Preoperative predictability of right ventricular failure following surgery for Ebstein's anomaly [J]. Eur J Cardiothorac Surg, 2019, 55(6): 1187-1193.
3
Bolat I. Preoperative right ventricular echocardiographic parameters predict perioperative cardiovascular complications in patients undergoing non-cardiac surgery [J]. Heart Lung Circ, 2020, 29(8): 1146-1151.
4
Chou J, Ma M, Gylys M, et al. Preexisting right ventricular dysfunction is associated with higher postoperative cardiac complications and longer hospital stay in high-risk patients undergoing nonemergent major vascular surgery [J]. J Cardiothorac Vasc Anesth, 2019, 33(5): 1279-1286.
5
Park JB, Lee SP, Lee JH, et al. Quantification of right ventricular volume and function using single-beat three-dimensional echocardiography: a validation study with cardiac magnetic resonance [J]. J Am Soc Echocardiogr, 2016, 29(5): 392-401.
6
Zhang QB, Sun JP, Gao RF, et al. Feasibility of single-beat full-volume capture real-time three-dimensional echocardiography for quantification of right ventricular volume: validation by cardiac magnetic resonance imaging [J]. Int J Cardiol, 2013, 168(4): 3991-3995.
7
Nesser HJ, Tkalec W, Patel AR, et al. Quantitation of right ventricular volumes and ejection fraction by three-dimensional echocardiography in patients: comparison with magnetic resonance imaging and radionuclide ventriculography [J]. Echocardiography, 2006, 23(8): 666-680.
8
Genovese D, Rashedi N, Weinert L, et al. Machine learning-based three-dimensional echocardiographic quantification of right ventricular size and function: validation against cardiac magnetic resonance [J]. J Am Soc Echocardiogr, 2019, 32(8): 969-977.
9
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]. Eur Heart J Cardiovasc Imaging, 2015, 16(3): 233-270.
10
Nowak-Machen M, Lang T, Schilling A, et al. Regional right ventricular volume and function analysis using intraoperative 3-dimensional echocardiography-derived mesh models [J]. J Cardiothorac Vasc Anesth, 2019, 33(6): 1527-1532.
11
Muraru D, Badano LP, Nagata Y, et al. Development and prognostic validation of partition values to grade right 20 ventricular dysfunction severity using 3D echocardiography [J]. Eur Heart J Cardiovasc Imaging, 2020, 21(1): 10-21.
12
Chang SA, Lee SC, Kim EY, et al. Feasibility of single-beat full-volume capture real-time three-dimensional echocardiography and auto-contouring algorithm for quantification of left ventricular volume: validation with cardiac magnetic resonance imaging [J]. J Am Soc Echocardiogr, 2011, 24(8): 853-859.
13
Liu BY, Wu WC, Zeng QX, et al. The value of three-dimensional echocardiography in risk stratification in pulmonary arterial hypertension: a cross-sectional study [J]. Int J Cardiovasc Imaging, 2020, 36(4): 577-584.
14
Kusunose K, Haga A, Abe T, et al. Utilization of artificial intelligence in echocardiography [J]. Circ J, 2019, 83(8): 1623-1629.
15
Narang A, Mor-Avi V, Prado A, et al. Machine learning based automated dynamic quantification of left heart chamber volumes [J]. Eur Heart J Cardiovasc Imaging, 2019, 20(5): 541-549.
16
Namisaki H, Nabeshima Y, Kitano T, et al. Prognostic value of the right ventricular ejection fraction, assessed by fully automated three-dimensional echocardiography: a direct comparison of analyses using right ventricular-focused views versus apical four-chamber views [J]. J Am Soc Echocardiogr, 2021, 34(2): 117-126.
17
Smolarek D, Gruchala M, Sobiczewski W. Echocardiographic evaluation of right ventricular systolic function: The traditional and innovative approach [J]. Cardiol J, 2017, 24(5): 563-572.
18
Kuhn A, Meierhofer C, Rutz T, et al. Non-volumetric echocardiographic indices and qualitative assessment of right ventricular systolic function in Ebstein's anomaly: comparison with CMR-derived ejection fraction in 49 patients [J]. Eur Heart J Cardiovasc Imaging, 2016, 17(8): 930-935.
19
Liu S, Bose R, Ahmed A, et al. Artificial intelligence-based assessment of indices of right ventricular function [J]. J Cardiothorac Vasc Anesth, 2020, 34(10): 2698-2702.
20
Ahmad A, Ibrahim Z, Sakr G, et al. A comparison of artificial intelligence-based algorithms for the identification of patients with depressed right ventricular function from 2-dimentional echocardiography parameters and clinical features [J]. Cardiovasc Diagn Ther, 2020, 10(4): 859-868.
[1] 李洋, 蔡金玉, 党晓智, 常婉英, 巨艳, 高毅, 宋宏萍. 基于深度学习的乳腺超声应变弹性图像生成模型的应用研究[J/OL]. 中华医学超声杂志(电子版), 2024, 21(06): 563-570.
[2] 杨敬武, 周美君, 陈雨凡, 李素淑, 何燕妮, 崔楠, 刘红梅. 人工智能超声结合品管圈活动对低年资超声医师甲状腺结节风险评估能力的作用[J/OL]. 中华医学超声杂志(电子版), 2024, 21(05): 522-526.
[3] 明昊, 肖迎聪, 巨艳, 宋宏萍. 乳腺癌风险预测模型的研究现状[J/OL]. 中华乳腺病杂志(电子版), 2024, 18(05): 287-291.
[4] 孔德铭, 刘铮, 李睿, 钱文伟, 王飞, 蔡道章, 柴伟. 人工智能辅助全髋关节置换三维术前规划准确性评价[J/OL]. 中华关节外科杂志(电子版), 2024, 18(04): 431-438.
[5] 何淳诺, 田志敏, 李焕玺, 吴昊越, 庄凯鹏, 周胜虎, 张浩强. 小儿发育性髋关节发育不良诊治的研究进展[J/OL]. 中华关节外科杂志(电子版), 2024, 18(04): 497-504.
[6] 叶莉, 杜宇. 深度学习在牙髓根尖周病临床诊疗中的应用[J/OL]. 中华口腔医学研究杂志(电子版), 2024, 18(06): 351-356.
[7] 熊鹰, 林敬莱, 白奇, 郭剑明, 王烁. 肾癌自动化病理诊断:AI离临床还有多远?[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(06): 535-540.
[8] 李伟, 宋子健, 赖衍成, 周睿, 吴涵, 邓龙昕, 陈锐. 人工智能应用于前列腺癌患者预后预测的研究现状及展望[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(06): 541-546.
[9] 黄俊龙, 李文双, 李晓阳, 刘柏隆, 陈逸龙, 丘惠平, 周祥福. 基于盆底彩超的人工智能模型在女性压力性尿失禁分度诊断中的应用[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(06): 597-605.
[10] 莫淇舟, 苏劲, 黄健, 李健维, 李思宁, 柳建军. 智能控压输尿管软镜碎石吸引取石术在直径10~25 mm上尿路结石中的应用[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(05): 497-502.
[11] 李义亮, 苏拉依曼·牙库甫, 麦麦提艾力·麦麦提明, 克力木·阿不都热依木. 机器人与腹腔镜食管裂孔疝修补术联合Nissen 胃底折叠术短期疗效分析[J/OL]. 中华疝和腹壁外科杂志(电子版), 2024, 18(05): 512-517.
[12] 王洪, 王骏华, 范建楠. 人工智能技术在肩袖损伤中的研究进展[J/OL]. 中华肩肘外科电子杂志, 2024, 12(04): 356-361.
[13] 孙铭远, 褚恒, 徐海滨, 张哲. 人工智能应用于多发性肺结节诊断的研究进展[J/OL]. 中华临床医师杂志(电子版), 2024, 18(08): 785-790.
[14] 杨松林, 黄仕豪, 王丽珠, 李禧萌, 邹飞翔, 李坤炜, 梁明柱, 陈炳辉. 良性肺结节生长变化的影像学评价[J/OL]. 中华介入放射学电子杂志, 2024, 12(04): 344-350.
[15] 张玮玮, 霍晓川. 人工智能时代医学生批判性思维培养的重要性[J/OL]. 中华脑血管病杂志(电子版), 2024, 18(04): 357-359.
阅读次数
全文


摘要