[1]于立恒,林锡祥,陈 煦,等.人工智能技术在心脏超声常规参数测量及左室舒张性慢性心力衰竭诊断中的应用[J].陕西医学杂志,2023,52(7):826-830.[doi:DOI:10.3969/j.issn.1000-7377.2023.07.011]
　YU Liheng,LIN Xixiang,CHEN Xu,et al.Application of artificial intelligence technology in measurement of conventional echocardiographic parameters and diagnosis of left ventricular diastolic chronic heart failure[J].,2023,52(7):826-830.[doi:DOI:10.3969/j.issn.1000-7377.2023.07.011]
点击复制

人工智能技术在心脏超声常规参数测量及左室舒张性慢性心力衰竭诊断中的应用
分享到：

《陕西医学杂志》[ISSN:1000-7377/CN:61-1281/TN]

卷:
52
期数:
2023年7期
页码:
826-830
栏目:
临床研究
出版日期:
2023-07-05

文章信息/Info

Title:
Application of artificial intelligence technology in measurement of conventional echocardiographic parameters and diagnosis of left ventricular diastolic chronic heart failure
作者:
于立恒1; 2; 林锡祥1; 2; 陈 煦1; 2; 何昆仑1
(1.解放军总医院医学大数据研究中心,北京 100853; 2.解放军总医院研究生院,北京 100853)
Author(s):
YU Liheng; LIN Xixiang; CHEN Xu; HE Kunlun
(Chinese PLA General Hospital,Beijing 100853,China)
关键词:
心脏超声;  人工智能;  深度学习;  图像分割卷积神经网络;  慢性心力衰竭;  左室舒张功能
Keywords:
Cardiac ultrasound;  Artificial intelligence;  Deep learning;  Image segmentation convolutional neural network;  Chronic heart failure;  Left ventricular diastolic function
分类号:
R 540.4
DOI:
DOI:10.3969/j.issn.1000-7377.2023.07.011
文献标志码:
A
摘要:
目的:探究人工智能算法模型在心脏超声常规参数测量及左室舒张性慢性心力衰竭(CHF)诊断中的应用效果。方法:收集410例左室舒张性CHF疑似病例者心脏超声图像为研究对象。由一组高年资超声医师完成纳入病例者的心脏超声参数测量; 同时,采用人工智能深度学习模型,即图像分割卷积神经网络,对纳入者心脏超声图像进行自动智能分割,并设计专门算法计算基于心脏超声影像的心脏结构和功能参数。通过比较两种方法获得心脏超声参数的一致性、偏差情况及受试者工作特征曲线(ROC)来探究人工智能深度学习技术在左室舒张性CHF诊断中的价值。结果:人工智能与医师所测心脏超声参数左室舒张末期容积(LVEDV)、左室收缩末期容积(LVESV)、左室射血分数(LVEF)间一致性一般(0.60≤ICC<0.80),E/A、E/e'、主动脉瓣峰值流速(AVPW)一致性极好(ICC≥0.80)。Bland-Altman分析表明,人工智能与医师所测心脏超声参数LVEDV、LVESV、LVEF间偏差较大,平均偏差分别为-16.9%、-7.0%和1.0%,E/A、E/e'、AVPW间偏差较低,平均偏差分别为0.0%、-0.4%和-0.0%。ROC曲线显示,人工智能所测E/A、E/e'、AVPW均对左心室舒张性CHF具有一定诊断价值(AUC=0.853、0.777、0.770,P<0.05)。结论:人工智能可快速识别并分割处理心脏超声图像,自动计算常规心脏参数,且关键参数与高年资超声医师结果一致性较好,并能用于左室舒张性CHF的临床辅助诊断。
Abstract:
Objective:To explore the application effect of artificial intelligence algorithm model in the measurement of conventional echocardiographic parameters and the diagnosis of left ventricular diastolic chronic heart failure(CHF).Methods:The ultrasound images of 410 suspected cases of left ventricular diastolic CHF were collected.Cardiac ultrasound parameters were measured by the same group of senior sonographers.At the same time,artificial intelligence model(image segmentation convolutional neural network)was used to segment and process the input cardiac ultrasound images,and the parameters of cardiac ultrasound were calculated and output.The consistency and deviation of cardiac ultrasound parameters obtained by two methods were compared.The value of artificial intelligence parameters in the diagnosis of left ventricular diastolic CHF was explored by receiver operating characteristic curve(ROC).Results:The artificial intelligence was consistent with the measured cardiac ultrasound parameters left ventricular end-diastolic volume(LVEDV),left ventricular end-systolic volume(LVESV)and left ventricular ejection fraction(LVEF)(0.60≤ICC<0.80),and the consistency of E/A,E/e' and AVPW was very good(ICC≥0.80).Blan-altman analysis showed that the deviation between artificial intelligence and the measured cardiac ultrasound parameters LVEDV,LVESV and LVEF was large,the average deviation was -16.9%,-7.0% and 1.0% respectively,and the deviation between E/A,E/e' and AVPW was low,the mean deviations were 0.0%,-0.4% and -0.0% respectively.ROC curve showed that E/A,E/e' and AVPW measured by artificial intelligence had certain diagnostic value for left ventricular diastolic CHF(AUC=0.853,0.777,0.770,P<0.05).Conclusion:Artificial intelligence can quickly identify and segment cardiac ultrasound images,automatically calculate conventional ultrasound parameters,and some conventional parameters can be highly consistent with the results of senior sonographers,and can be used for the clinical diagnosis of left ventricular diastolic CHF.

参考文献/References:

[1] 中华医学会心血管病学分会心力衰竭学组,中国医师协会心力衰竭专业委员会中华心血管病杂志编辑委员会.中国心力衰竭诊断和治疗指南2018[J].中华心血管病杂志,2018,46(10):760-789.
[2] Bueno H,Moura B,Lancellotti P,et al.The year in cardiovascular medicine 2020:Heart failure and cardiomyopathies[J].Eur Heart J,2021,42(6):657-670.
[3] Hao G,Wang X,Chen Z,et al.Prevalence of heart failure and left ventricular dysfunction in China:The China hypertension survey,2012-2015[J].Eur J Heart Fail,2019,21(11):1329-1337.
[4] Jones N,Smith M,Lay-Flurrie S,et al.Survival among people with heart failure and atrial fibrillation; A population cohort study[J].Eur Heart J,2022,43(S2):899.
[5] Volpe M,Francia P,Tocci G,et al.Prediction of long-term survival in chronic heart failure by multiple biomarker assessment:A 15-year prospective follow-up study[J].Clin Cardiol,2010,33(11):700-707.
[6] 许明芳,张 方,罗季平,等.心力衰竭患者左室舒张功能异常的心脏超声评价及其可靠性研究[J].影像科学与光化学,2020,38(5):897-900.
[7] 徐 杰,潘如梅,潘如凤.超声心动图对左心室舒张性心力衰竭的评价[J].中华解剖与临床杂志,2022,19(2):147-150.
[8] 周小雪,张莹莹,张 烨,等.人工智能技术在胎儿超声心动图四腔心切面筛查中的应用[J].中华超声影像学杂志,2020,29(8):668-672.
[9] Knackstedt C,Bekkers S,Schummers G,et al.Fully automated versus standard tracking of left ventricular ejection fraction and longitudinal strain the FAST-EFs multicenter study[J].J Am Coll Cardiol,2015,66(13):1456-1466.
[10] Anwar SM,Majid M,Qayyum A,et al.Medical Image Analysis using Convolutional Neural Networks:A review[J].J Med Syst,2018,42(11):226-238.
[11] Shi X,Chen Z,Wang H,et al.Convolutional LSTM network:A machine learning approach for precipitation nowcasting[J].Adv Neural Inf Process Syst,2015,1(1):802-810.
[12] Yu M,Duan Y,Li Z,et al.Prediction of peptide detectability based on capsnet and convolutional block attention module[J].Int J Mol Sci,202,22(21):12080-12088.
[13] Shelhamer E,Long J,Darrell T.Fully convolutional networks for semantic segmentation[J].IEEE Trans Pattern Anal Mach Intell,2017,39(4):640-651.
[14] Zhang B,Goel B,Sarhan MH,et al.Surgical workflow recognition with temporal convolution and transformer for action segmentation[J].Int J Comput Assist Radiol Surg,2023,18(4):785-794.
[15] Simpson T.An attempt to account for the regularity of the tides[J].Philos Trans R Soc Lond B Bio Sci,1743,38(1):45-88.
[16] Gomaa S,Soliman AA,Nasr K,et al.Development of artificial neural network models to calculate the areal sweep efficiency for direct line,staggered line drive,five-spot,and nine-spot injection patterns[J].Fuel,2022,317:123564.
[17] Pranjal A,Nitish- Katal1 CA,Nishtha H.Segmentation and classification of brain tumor using 3D-UNet deep neural networks[J].Int J Cog Com Eng,2022,3:199-210.
[18] 兰 津,尹璐瑶,鲁洪涛,等.斑点追踪分层应变联合心肌做功技术评价射血分数保留的心力衰竭患者左心室功能[J].中华超声影像学杂志,2021,30(10):836-842.
[19] 范 苗,任卫东,宋 光,等.三维径向应变评价射血分数保留的心力衰竭患者左心室收缩功能变化[J].中国医科大学学报,2018,47(7):581-584.

相似文献/References:

[1]彭 赛,卢丛兰,孙中洋.人工智能在骨质疏松症诊疗中的应用进展[J].陕西医学杂志,2024,(3):425.[doi:DOI:10.3969/j.issn.1000-7377.2024.03.030]
　PENG Sai,LU Conglan,SUN Zhongyang.Advances in application of artificial intelligence in diagnosis and treatment of osteoporosis[J].,2024,(7):425.[doi:DOI:10.3969/j.issn.1000-7377.2024.03.030]

备注/Memo

备注/Memo:
基金项目:中国2020医疗健康智能筛查诊断项目(2020-0103-3-1)
更新日期/Last Update: 2023-07-05