基于深度学习的川崎病合并冠状动脉瘤超声分类诊断

doi:10.12372/jcp.2024.23e0717

摘要/Abstract

摘要：

目的探讨基于深度学习的川崎病合并冠状动脉瘤超声分类诊断的可行性。方法收集复旦大学附属儿科医院诊断为川崎病患者的心超图像，选取胸骨旁大动脉短轴切面，非彩色多普勒成像的二维图像，剔除冠状动脉图像缺失、图像质量差，图像不完善的病例，共纳入研究图片1 328张，其中包括冠状动脉瘤664张图片，冠状动脉正常664张图片。利用所采集图片基于典型的深度神经网络AlexNet、LeNet、VggNet19、ResNet18进行分类诊断，其中1 000张作为训练集，164张作为验证集，164张作为测试集，约各占总图像数量的80%、10%、10%。结果基于AlexNet的深度神经网路对川崎病合并冠状动脉瘤的超声图像分类结果最好，可达98%准确率，且该深度分类模型在参数数量及训练效率上均有明显优势。结论基于深度卷积神经网络的川崎病合并冠状动脉瘤的超声分类诊断具有可行性。

关键词: 川崎病, 超声心动图, 深度学习, 分类诊断

Abstract:

Objective To investigate the feasibility of ultrasonic classification diagnosis based on deep learning for Kawasaki disease complicated with coronary aneurysms. Methods Echocardiography images of Kawasaki patients diagnosed in the Pediatric Hospital of Fudan University were collected. The parasternal short axis sections were selected and two-dimensional images without color Doppler imaging were included. In addition, the cases with missing images, poor image quality and imperfect images were eliminated. The collected dataset contains 664 images of coronary aneurysms and 664 images of normal coronary arteries. The collected images are used for classification diagnosis based on four classical deep neural networks including AlexNet, LeNet, VggNet19 and ResNet18. The dataset was divided into a training set of 1000 images, a validation set of 164 images, and a test set of 164 images, accounting for approximately 80%, 10%, and 10% of the total image count, respectively. Results The AlexNet based deep neural network has the best results for ultrasonic image classification of Kawasaki disease complicated with coronary aneurysms and the diagnostic accuracy can reach 98%. In addition, this depth model has obvious advantages in the parameter number and training efficiency. Conclusion The ultrasonic classification diagnosis of Kawasaki disease complicated with coronary aneurysms based on deep convolutional neural network is feasible.

Key words: Kawasaki disease, echocardiography, deep learning, classification diagnosis

党艳, 张璟, 高燕, 黄国英, 马晓静. 基于深度学习的川崎病合并冠状动脉瘤超声分类诊断[J]. 临床儿科杂志, 2024, 42(10): 849-856.

DANG Yan, ZHANG Jing, GAO Yan, HUANG Guoying, MA Xiaojing. Ultrasonic classification diagnosis of Kawasaki disease complicated with coronary aneurysms based on deep learning[J]. Journal of Clinical Pediatrics, 2024, 42(10): 849-856.

图/表 11

图1

图2

图3

图4

图5

图6

图7

图8

图9

表1

图10

参考文献 24

[1]	McCrindle BW, Rowley AH, Newburger JW, et al. Diagnosis, treatment, and long-term management of Kawasaki disease: a scientific statement for health professionals from the American Heart Association[J]. Circulation, 2017, 135(17): 927-999. doi: 10.1161/CIR.0000000000000484 pmid: 28356445
[2]	王雷, 夏焙. 超声心动图在川崎病诊断、治疗及长期随访中的应用进展[J]. 中华医学超声杂志, 2019, 16(3): 161-165.
[3]	Litjens G, Ciompi F, Wolterink J M, et al. State-of-the-art deep learning in cardiovascular image analysis[J]. JACC Cardiovasc Imaging, 2019, 12(8): 1549-1565.
[4]	Kusunose K, Haga A, Abe T, et al. Utilization of artificial intelligence in echocardiography[J]. Circ J, 2019, 83: 1623-1629. doi: 10.1253/circj.CJ-19-0420 pmid: 31257314
[5]	陶攀, 付忠良, 朱锴, 等. 基于深度学习的超声心动图切面识别方法[J]. 计算机应用, 2017, 37(5): 1434-1438. doi: 10.11772/j.issn.1001-9081.2017.05.1434
[6]	Tan W, Cao Y, Ma X, et al. Bayesian inference and dynamic neural feedback promote the clinical application of intelligent congenital heart disease diagnosis[J]. Engineering, 2023, 23: 90-102.
[7]	Yang C, Ojha BD, Aranoff ND, et al. Classification of aortic stenosis using conventional machine learning and deep learning methods based on multi-dimensional cardio mechanical signals[J]. Sci Rep, 2020, 10(1): 17521.
[8]	宋美琪, 徐皓煊, 汝童, 等. 人工智能在超声影像中的应用现状[J]. 中国中西医结合影像学杂志, 2020, 18(5): 528-533.
[9]	Zhou J, Du M, Chang S, et al. Artificial intelligence in echocardiography: detection, functional evaluation, and disease diagnosis[J]. Cardiovasc Ultrasound, 2021, 19(1): 29.
[10]	Song D, Kim E, Huang X, et al. Multimodal entity coreference for cervical dysplasia diagnosis[J]. IEEE Trans Med Imaging, 2014, 34(1): 229-245.
[11]	Lawrence S, Giles CL, Tsoi A, et al. Face recognition: a convolutional neural-network approach[J]. IEEE Transactions on Neural Netw, 1997, 8(1): 98-113.
[12]	He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]. IEEE conference on computer vision and pattern recognition. Las Vegas, NV, USA, 2016.
[13]	陈思文, 刘玉江, 刘冬, 等. 基于AlexNet模型和自适应对比度增强的乳腺结节超声图像分类[J]. 计算机科学, 2019, 46(6): 146-153.
[14]	Sarvamangala DR, Kulkarni RV. Convolutional neural networks in medical image understanding[J]. Evol Intell, 2021, 15(1): 1-22.
[15]	Xie LP, Yan WL, Huang M, et al. Epidemiologic features of Kawasaki disease in Shanghai from 2013 through 2017[J]. J Epidemiol, 2020, 30(10): 429-435.
[16]	Huang GY, Ma XJ, Huang M, et al. Epidemiologic pictures of Kawasaki disease in Shanghai from 1998 through 2002[J]. J Epidemiol. 2006, 16(1): 9-14.
[17]	Ma XJ, Yu CY, Huang M, et al. Epidemiologic features of Kawasaki disease in Shanghai from 2003 through 2007[J]. Chin Med J (Engl), 2010, 123(19): 2629-2634.
[18]	Chen JJ, Ma XJ, Liu F, et al. Epidemiologic features of Kawasaki disease in Shanghai from 2008 through 2012[J]. Pediatr Infect Dis J, 2016, 35(1): 7-12.
[19]	Zhang J, Gajjala S, Tison GH, et al. Fully automated echocardiogram interpretation in clinical practice[J]. Circulation, 2018, 138: 1623-1635. doi: 10.1161/CIRCULATIONAHA.118.034338 pmid: 30354459
[20]	Madani A, Arnaout R, Mofrad M, et al. Fast and accurate view classification of echocardiograms using deep learning[J]. NPJ Digit Med, 2018, 1:6.
[21]	Kusunose K, Abe T, Haga A, et al. A Deep learning approach for assessment of regional wall motion abnormality from echocardiographic images[J]. ACC Cardiovasc Imaging, 2020, 13(1): 374-381.
[22]	Ouyang D, He B, Ghorbani A, et al. Video-based AI for beat-to-beat assessment of cardiac function[J]. Nature, 2022, 580(7802): 252-256.
[23]	Leha A, Hellenkamp K, Unsold B, et al. A machine learning approach for the prediction of pulmonary hypertension[J]. PLoS ONE, 2021, 14(10): e0224453.
[24]	Xu E, Nemati S, Tremoulet AH. A deep convolutional neural network for Kawasaki disease diagnosis[J]. Sci Rep, 2022, 12(1): 11438. doi: 10.1038/s41598-022-15495-x pmid: 35794205

模型	评估指标
模型	准确率/%	测试损失率
AlexNet LeNet VggNet19 ResNet18	98.78 94.51 64.63 91.46	6.0×10^-6 0.000 63 2.4×10^-7 0.005