基于人工智能的医学影像学检查在动脉粥样硬化性狭窄定量评估的研究进展
Research Progress of AI-Based Medical Imaging Examination in Quantitative Evaluation of Atherosclerotic Stenosis
DOI: 10.12677/acm.2024.1441068, PDF, HTML, XML, 下载: 31  浏览: 65  科研立项经费支持
作者: 宦欣玥, 郭大静*:重庆医科大学附属第二医院放射科,重庆
关键词: 动脉粥样硬化性狭窄人工智能医学影像学定量评估Atherosclerotic Stenosis Artificial Intelligence Medical Imaging Quantitative Assessment
摘要: 动脉狭窄是动脉粥样硬化疾病的主要结局之一,其进展和稳定性对动脉粥样硬化疾病的预后和治疗产生重大影响。现有的影像学方法对于动脉粥样硬化性狭窄(Atherosclerotic stenosis, AS)的准确评估存在局限性,而人工智能(Artificial intelligence, AI)在医学影像分析中发挥重要作用,可以实现对病变严重程度和进展速度的定量评估及风险预测。目前,基于AI的医学影像学在动脉狭窄定量评估方面取得了显著进展,尤其是基于深度学习(Deep learning, DL)的算法在血管狭窄预测、斑块分类和识别中表现出良好的性能。本文对基于AI的医学影像学检查在AS定量评估中的研究进展进行综述,并对未来基于AI技术在动脉狭窄定量评估中可能存在的挑战和机遇进行展望。
Abstract: Arterial stenosis is one of the main outcomes of atherosclerotic diseases, and its progression and stability have a significant impact on the prognosis and treatment of atherosclerotic diseases. Existing imaging methods have limitations for accurate assessment of atherosclerotic stenosis (AS), and artificial intelligence (AI) plays an important role in medical image analysis. The quantitative evaluation and risk prediction of the severity and progression of the disease can be realized. At present, AI-based medical imaging has made remarkable progress in the quantitative assessment of arterial stenosis, especially the algorithm based on deep learning (DL) has shown good performance in the prediction of arterial stenosis, plaque classification and recognition. This article reviews the research progress of AI-based medical imaging in quantitative assessment of AS, and looks forward to the possible challenges and opportunities of AI-based quantitative assessment of arterial stenosis in the future.
文章引用:宦欣玥, 郭大静. 基于人工智能的医学影像学检查在动脉粥样硬化性狭窄定量评估的研究进展[J]. 临床医学进展, 2024, 14(4): 615-622. https://doi.org/10.12677/acm.2024.1441068

1. 引言

动脉粥样硬化性狭窄(Atherosclerotic stenosis, AS)影响冠状动脉、脑动脉及外周动脉从而造成的缺血性心脏病和缺血性脑卒中仍是全球死亡和残疾的主要原因 [1] [2] 。AS的严重程度能够指导临床医生评估病因,估计风险,并确定适当的干预措施,因此,作为动脉狭窄程度的诊断基础,动脉狭窄的准确量化对于建立最佳的临床管理至关重要 [3] [4] [5] 。

传统医学影像学检查在AS定量评估中发挥了重要作用,已发展到表征病变的各个方面,包括程度狭窄(%)、最小管腔面积(mm2)、斑块体积(mm3)、斑块负荷(%)、重塑比(%)、偏心指数和与组织炎症相关的增强率(%) [6] [7] [8] ,临床上目前主要依靠超声成像、CTA、MRI的无创方法进行诊断 [8] [9] [10] [11] ,然而传统的影像学判读由于大量的图像和广泛的成像范围会导致诊治被延误从而造成不可逆转的损伤 [12] 。另外目前常用的AS定量评估的目测法及手动测量法则存在主观性及耗时、耗力的缺点 [13] [14] 。相比之下,基于人工智能(Artificial intelligence, AI)的医学影像学检查擅长识别图像数据中复杂的表现形式,能够对图像进行自动分割与自动定量诊断,从而缩短图像后处理及诊断时间,缩小放射科诊断医师之间的差异,避免医生在超负荷工作后由于注意力不集中或视觉感知疲劳而造成的诊断错误,在AS的诊断、治疗及预后评估等方面发挥重要作用 [15] [16] [17] [18] 。本文就基于AI的医学影像学检查在AS定量评估中的相关研究进行综述。

2. 人工智能的概述

AI于1956年美国达特茅斯会议上首次被提出,因其能够分析复杂数据、自动关联和预测的能力,优于传统的统计方法,在现实世界中显示出巨大的应用前景 [19] 。AI的核心是机器学习(Machine learning, ML),随着计算机设备的不断更新,深度学习(Deep learning, DL)作为ML的一个子领域问世 [20] ,DL可在没有明确标识的情况下接收输入的数据并学习其特征,形成端–端机器学习模式,广泛应用于众多学科领域中,尤其在临床医疗解决实际疑难问题时有较强的表现,例如基于图像的卷积神经网络(Convolutional neural networks, CNN)融合了学习卷积滤波器进行特征学习的思想,利用隐含层之间简单的局部非线性连接,创建每个深层的抽象表示,降低了模型的计算代价,还可以跨层共享权重,提取独立于其位置的图像特征,并使用尽可能多的隐藏层进行学习,是医疗领域的热点 [19] [21] [22] [23] 。目前在血管成像领域,AI已被开发应用于心血管疾病的定量诊断和预测等方面 [24] [25] [26] 。

3. 基于人工智能的医学影像学检查在动脉粥样硬化性狭窄定量评估中的应用

3.1. 人工智能在CTA中的应用

CT血管造影(CT angiography, CTA)是一种快速、准确、无创的评估AS的工具,特别是在急性情况下其应用更加有必要 [27] [28] ,疑似心血管疾病的患者通常会行冠状动脉CT血管造影(Coronary CT angiography, CCTA)或头颈部CTA检查,是当下基于AI定量分析中最流行的成像方式 [13] [14] [22] 。

Fan等 [29] 建立了一个基于三维CNN模型的用于头颈部CTA图像的自动成像重建系统,并对该模型的算法性能、图像重建质量是否满足诊断需求、模型增强后的临床应用效率三个方面进行了全面的定量评估,结果发现该算法对各种增强CTA扫描(包括主动脉、颈动脉和颅内动脉)中骨和血管的分割性能好(训练集和验证集的骰子相似系数分别为0.979和0.960、0.975和0.944,分割的总体准确率为93.1%),且图像质量的合格率为92%;该研究团队在后来的研究中 [14] ,又建立了一个结合二维和三维ResU-Net的CNN模型进行准确的狭窄检测和斑块分类,该模型对狭窄检测(κ = 0.84)、斑块分类(κ = 0.78)及狭窄定量诊断(R = 0.87, P < 0.001) 的一致性好,并能显著减少放射科医生诊断和报告书写的时间。即使在早期,一些半自动化人工智能软件对动脉狭窄程度的定量诊断也有较好的一致性(组内相关系数(Interclass correlation coefficient, ICC > 80%)),但由于AI算法的不成熟,对于斑块的分割与识别错误是造成诊断错误的主要原因之一 [9] [30] [31] 。

基于AI的CCTA分析也取得了先进成果。Griffin等 [13] [32] 评估基于ML的CCTA狭窄程度量化的性能,分别得到良好和极好的一致性(ICC分别为0.73、0.88)。对斑块钙化积分、斑块的体积 [26] [32] 的定量分析中均表现出极好的一致性(ICC分别为0.96、0.96)。除了对管径和管壁的量化,AI还能应用于血流的定量诊断。侵入性血流储备分数(Fractional flow reserve, FFR)是评估冠状动脉生理功能的“金标准”,但该方法为有创检查且费用昂贵,FFRCT是基于CCTA三维数据模型运用计算流体力学模拟真实冠状动脉血流而获取的冠状动脉血流动力学信息,现有基于ML的FFRCT能够实现在数分钟内快速分析冠状动脉解剖结构与血流动力学变化 [33] ,Schuessler等 [34] 研究结果表明,基于ML的FFRCT与FFRCT具有良好的相关性,基于量化评估的分层分析显示其诊断性能优于单纯CCTA (曲线下面积(Area under the curve, AUC)分别为0.84、0.69)。

3.2. 人工智能在MRI中的应用

高分辨率MRI血管壁成像作为一种新兴的非侵入性成像方式,因其优越的软组织对比度,已被证明是直接观察颅内和颈动脉血管壁和分类各种斑块特征的首选诊断方法 [35] ,近年来,3D涡轮自旋回波被引入颅内及颈动脉血管壁的成像更显示出优越的分辨能力,在对管壁及管腔参数的评估中,ICC可达0.988 [36] [37] [38] ,。研究者将CNN应用于颅内或颅外颈动脉血管壁分析,包括血管壁分割和分类。Shi等 [39] [40] 利用重建的2D图像切片进行了血管分析,在对管壁面积、管腔面积、最小壁厚、最大壁厚和斑块范围等参数进行定量评估发现一致性极好(ICC范围0.77~0.93)。Wan等 [41] 开发了一种自动分析颅内外动脉血管壁3D图像的方法,对血管中心线跟踪、血管矫直、血管壁分割和形态学量化评估同样发现极好的一致性(骰子系数为0.88~0.90),与前期Gao等 [42] 基于3D模型的方法结果一致,该研究发现从双序列MRI中分割颈动脉壁在所有颈动脉切片人工和自动识别诊断轮廓的一致性也极好(骰子系数均大于0.87)。在最新的研究中,Wu等 [43] 利用多任务学习技术对黑血MRI上进行建模并在神经网络中完成颈动脉管腔分割、外壁分割和颈动脉粥样硬化诊断,该方法在“CAREII测试”数据集中获得了目前最好的分割性能(管腔分割骰子系数为0.97,外壁分割骰子系数为0.97),对颈动脉粥样硬化诊断的准确性也较好(AUC为0.95,准确度为0.90)。

3.3. 人工智能在超声中的应用

超声优异的软组织分辨率和亚毫米空间分辨率非常适合血管细节成像,尤其适用于大的、浅表的、接近颈总动脉分叉处的血管探测 [2] [44] [45] 。与CCTA检查相比,心脏研究相对较少,因为体外超声不适用于冠状动脉检查,识别无症状性颈动脉疾病并提供临床决策支持的ML工具的开发是当前超声研究的主要焦点 [46] 。由于颈动脉内膜厚度(Intimamedia thickness, IMT)在CTA检查中的空间分辨率低使得超声检查成为IMT的独特诊断方法。IMT的自动分割基于深度学习单层前馈神经网络使用了几种超声成像参数,如Hough转换、频域分析、像素强度分析,与专家手动分割相比,基于定量分析的分层分析得出灵敏度和特异性均大于97% [47] 。Huang等 [48] 对单个斑块进行灰度分布分析,将斑块分为回声丰富、中等和回声清晰并建立回声类型和斑块易感性之间的关联,斑块类型的诊断准确率为77.5%。Roy等 [49] [50] 利用多种参数研究回声强度与颈动脉斑块易损性的相关性,AUC可高达0.97。

3.4. 基于人工智能的预后预测与风险评估

正确识别具有AS高危因素的患者,及时给予干预政策是心脑血管医学面临的一个重大挑战。尽管当前有各式各样的预测模型应用于预测心血管不良事件中,但传统的预测模型是基于对人群的分析而建立的,只能分析部分变量,无法有效地进行个体风险评估,而AI模型能够识别并综合分析大量的风险预测,因子及影像学参数,对心脑血管不良事件进行预测,帮助临床医生提前识别出不同风险层的患者,制定个性化治疗计划和随访方案,从而降低心脑血管疾病的发病率和死亡率 [51] 。

Hilbert等 [52] 建立DL模型以利用CTA图像直接预测血管内治疗后良好的再灌注和良好功能结果,AUC分别为0.65、0.71。此外,MRA可以显示与年龄相关的脑动脉变化,但测量或评估这种变化的工具或方法有限,Nam [53] 等开发了一种三维CNN模型并在三维MRA上量化评估年龄相关的脑动脉变化,可预测年龄差以此定量分析脑血管衰老。Mu等 [26] 评估基于DL的CCTA和非对比CT冠状动脉钙化评分的风险分类一致极好(加权κ = 0.94),分类准确率为93%。Johnson等 [54] 利用ML算法开发了一个血管特征模型,用来识别未来有无发生心血管不良事件的患者,发现该模型在预测冠状动脉疾病死亡率和全因死亡率时的预测效能均优于传统冠状动脉血管评分。先前Motwani等 [55] 研究的模型预测性能也高于传统CCTA指标评分,但其性能低于Johnson等开发的模型。这些研究观察到的结果都展现了在预测疑似心脑血管疾病患者的预后风险方面,AI模型已具备传统模型所没有的优势。

4. 人工智能的限制

近年来,随AI技术在医学影像学领域的不断发展,相关研究表明 [13] [14] [49] [52] ,AI技术能够通过对图像信息、病变特征和生物标记物进行分析,可从管腔直径、管腔面积、斑块成分等多个参数评估血管并预测AS患者的发病风险、疾病转归和临床结局,并为AS患者的危险分层提供参考依据。但其研究仍存在一些问题:1) 目前尚无统一的数据库,且AI技术应用于AS领域是一个跨学科、多领域的交叉学科,如何将多个领域的相关数据进行整合仍是一个需要解决的问题;2) AI技术对AS患者发病风险预测及预后评估仍停留在初步阶段,如何进一步优化模型、提高模型泛化能力、完善预测体系是目前需要解决的问题;3) AI技术在临床应用中仍存在许多伦理问题,如何平衡好患者的隐私保护和AI技术发展之间的关系是未来需要解决的问题。

5. 小结与展望

总之,虽然AI技术的发展为AS评估带来了新的机遇,但目前仍存在许多不足之处,需要进一步改进和完善。首先,目前的AI技术在分类准确性方面仍有待提高,未来研究需进一步完善算法,开发新的分类特征或改进已有特征。其次,对于斑块风险的评估需要更多高质量的前瞻性临床研究来证明其价值。最后,虽然AI技术在AS风险评估中表现出了一定的优势,但其对斑块的准确识别还存在很多问题,如斑块异质性、图像质量等。随着研究的深入,AI技术对动脉粥样硬化斑块的分类将更加准确、定量评估将更加精确,为放射科医师及临床医师提供更准确的诊断依据。

基金项目

重庆市科卫联合医学科研项目(2023MSXM014)。

NOTES

*通讯作者。

参考文献

[1] Pasterkamp, G., Den Ruijter, H.M. and Giannarelli, C. (2022) False Utopia of One Unifying Description of the Vulnerable Atherosclerotic Plaque: A Call for Recalibration That Appreciates the Diversity of Mechanisms Leading to Atherosclerotic Disease. Arteriosclerosis, Thrombosis, and Vascular Biology, 42, e86-e95.
https://doi.org/10.1161/ATVBAHA.121.316693
[2] Roth, G.A., Mensah, G.A., Johnson, C.O., et al. (2020) Global Burden of Cardiovascular Diseases and Risk Factors, 1990-2019. Journal of the American College of Cardiology, 76, 2982-3021.
https://doi.org/10.1016/j.jacc.2020.11.010
[3] Libby, P. (2021) The Changing Landscape of Atherosclerosis. Nature, 592, 524-533.
https://doi.org/10.1038/s41586-021-03392-8
[4] Nardi, V., Benson, J., Bois, M., et al. (2022) Carotid Plaques from Symptomatic Patients with Mild Stenosis Is Associated with Intraplaque Hemorrhage. Hypertension, 79, 271-282.
https://doi.org/10.1161/HYPERTENSIONAHA.121.18128
[5] Fox, A. (1993) How to Measure Carotid Stenosis. Radiology, 186, 316-318.
https://doi.org/10.1148/radiology.186.2.8421726
[6] Merino, J.G. and Warach, S. (2010) Imaging of Acute Stroke. Nature Reviews Neurology, 6, 560-571.
https://doi.org/10.1038/nrneurol.2010.129
[7] Flachskampf, F.A., Benson, J., Bois, M., et al. (2011) Cardiac Imaging after Myocardial Infarction. European Heart Journal, 32, 272-283.
https://doi.org/10.1093/eurheartj/ehq446
[8] Su, M.Y. (2021) Editorial for “the Occurrence and Outcome of Mild Intracranial Atherosclerotic Stenosis: A Prospective High-Resolution MRI Study”. Journal of Magnetic Resonance Imaging, 54, 89-90.
https://doi.org/10.1002/jmri.27571
[9] Siepmann, T., Barlinn, K. Floegel, T., et al. (2021) CT Angiography Manual Multiplanar Vessel Diameter Measurement vs. Semiautomated Perpendicular Area Minimal Caliber Computation of Internal Carotid Artery Stenosis. Frontiers in Cardiovascular Medicine, 8, Article 740237.
https://doi.org/10.3389/fcvm.2021.740237
[10] Bae, Y., Kang, S.J., Kim, G., et al. (2019) Prediction of Coronary Thin-Cap Fibroatheroma by Intravascular Ultrasound-Based Machine Learning. Atherosclerosis, 288, 168-174.
https://doi.org/10.1016/j.atherosclerosis.2019.04.228
[11] Forssen, H., Patel, R., Fitzpatrick, N., et al. (2017) Evaluation of Machine Learning Methods to Predict Coronary Artery Disease Using Metabolomic Data. Studies in Health Technology and Informatics, 235, 111-115.
[12] Yang, Y., Patel, R., Fitzpatrick, N., et al. (2023) Performance of Deep Learning-Based Autodetection of Arterial Stenosis on Head and Neck CT Angiography: An Independent External Validation Study. La Radiologia Medica, 128, 1103-1115.
https://doi.org/10.1007/s11547-023-01683-w
[13] Griffin, W.F., Patel, R., Fitzpatrick, N., et al. (2023) AI Evaluation of Stenosis on Coronary CTA, Comparison with Quantitative Coronary Angiography and Fractional Flow Reserve: A CREDENCE Trial Substudy. JACC: Cardiovascular Imaging, 16, 193-205.
https://doi.org/10.1016/j.jcmg.2021.10.020
[14] Fu, F., Shan, Y., Yang, G., et al. (2023) Deep Learning for Head and Neck CT Angiography: Stenosis and Plaque Classification. Radiology, 307, e220996.
https://doi.org/10.1148/radiol.220996
[15] Wardlaw, J.M., Mair, G., Von Kummer, R., et al. (2022) Accuracy of Automated Computer-Aided Diagnosis for Stroke Imaging: A Critical Evaluation of Current Evidence. Stroke, 53, 2393-2403.
https://doi.org/10.1161/STROKEAHA.121.036204
[16] Wiklund, P., Medson, K. and Elf, J. (2023) Incidental Pulmonary Embolism in Patients with Cancer: Prevalence, Underdiagnosis and Evaluation of an AI Algorithm for Automatic Detection of Pulmonary Embolism. European Radiology, 33, 1185-1193.
https://doi.org/10.1007/s00330-022-09071-0
[17] Salerno, A., Strambo, D., Nannoni, S., et al. (2022) Patterns of Ischemic Posterior Circulation Strokes: A Clinical, Anatomical, and Radiological Review. International Journal of Stroke, 17, 714-722.
https://doi.org/10.1177/17474930211046758
[18] Langlotz, C.P. (2019) Will Artificial Intelligence Replace Radiologists? Radiology: Artificial Intelligence, 1, e190058.
https://doi.org/10.1148/ryai.2019190058
[19] Pesapane, F., Codari, M. and Sardanelli, F. (2018) Artificial Intelligence in Medical Imaging: Threat or Opportunity? Radiologists again at the Forefront of Innovation in Medicine. European Radiology Experimental, 2, Article No. 35.
https://doi.org/10.1186/s41747-018-0061-6
[20] Bizzo, B.C., Dasegowda, G. Bridge, C., et al. (2023) Addressing the Challenges of Implementing Artificial Intelligence Tools in Clinical Practice: Principles from Experience. Journal of the American College of Radiology, 20, 352-360.
https://doi.org/10.1016/j.jacr.2023.01.002
[21] 俞益洲, 石德君, 马杰超, 等. 人工智能在医学影像分析中的应用进展[J]. 中国医学影像技术, 2019, 35(12): 1808-1812.
[22] Hosny, A., Parmar, C., Quackenbush, J., et al. (2018) Artificial Intelligence in Radiology. Nature Reviews. Cancer, 18, 500-510.
https://doi.org/10.1038/s41568-018-0016-5
[23] Kriegeskorte, N. and Golan, T. (2019) Neural Network Models and Deep Learning. Current Biology, 29, R231-R236.
https://doi.org/10.1016/j.cub.2019.02.034
[24] Sheahan, M., Ma, X., Paik, D., et al. (2018) Atherosclerotic Plaque Tissue: Noninvasive Quantitative Assessment of Characteristics with Software-Aided Measurements from Conventional CT Angiography. Radiology, 286, 622-631.
https://doi.org/10.1148/radiol.2017170127
[25] Sieren, M.M., Widmann, C., Weiss, N., et al. (2022) Automated Segmentation and Quantification of the Healthy and Diseased Aorta in CT Angiographies Using a Dedicated Deep Learning Approach. European Radiology, 32, 690-701.
https://doi.org/10.1007/s00330-021-08130-2
[26] Mu, D., Widmann, C., Weiss, N., et al. (2022) Calcium Scoring at Coronary CT Angiography Using Deep Learning. Radiology, 302, 309-316.
https://doi.org/10.1148/radiol.2021211483
[27] Luijten, S.P.R., Wolff, L., Duvekot, M.H.C., et al. (2022) Diagnostic Performance of an Algorithm for Automated Large Vessel Occlusion Detection on CT Angiography. Journal of NeuroInterventional Surgery, 14, 794-798.
https://doi.org/10.1136/neurintsurg-2021-017842
[28] Torres, C., Lum, C., Puac-Polanco, P., et al. (2021) Differentiating Carotid Free-Floating Thrombus from Atheromatous Plaque Using Intraluminal Filling Defect Length on CTA: A Validation Study. Neurology, 97, e785-e793.
https://doi.org/10.1212/WNL.0000000000012368
[29] Fu, F., Wei, J., Zhang, M., et al. (2020) Rapid Vessel Segmentation and Reconstruction of Head and Neck Angiograms Using 3D Convolutional Neural Network. Nature Communications, 11, Article No. 4829.
https://doi.org/10.1038/s41467-020-18606-2
[30] Borst, J., Marquering, H.A., Kappelhof, M., et al. (2015) Diagnostic Accuracy of 4 Commercially Available Semiautomatic Packages for Carotid Artery Stenosis Measurement on CTA. American Journal of Neuroradiology, 36, 1978-1987.
https://doi.org/10.3174/ajnr.A4400
[31] Budoff, M.J., Dowe, D., Jollis, J.G., et al. (2008) Diagnostic Performance of 64-Multidetector Row Coronary Computed Tomographic Angiography for Evaluation of Coronary Artery Stenosis in Individuals without Known Coronary Artery Disease: Results from the Prospective Multicenter ACCURACY (Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography) Trial. Journal of the American College of Cardiology, 52, 1724-1732.
https://doi.org/10.1016/j.jacc.2008.07.031
[32] Lin, A., Manral, N., McElhinney, P., et al. (2022) Deep Learning-Enabled Coronary CT Angiography for Plaque and Stenosis Quantification and Cardiac Risk Prediction: An International Multicentre Study. The Lancet Digital Health, 4, E256-E265.
https://doi.org/10.1016/S2589-7500(22)00022-X
[33] Liu, X., Mo, X., Zhang, H., et al. (2021) A 2-Year Investigation of the Impact of the Computed Tomography-Derived Fractional Flow Reserve Calculated Using a Deep Learning Algorithm on Routine Decision-Making for Coronary Artery Disease Management. European Radiology, 31, 7039-7046.
https://doi.org/10.1007/s00330-021-07771-7
[34] Schuessler, M., Saner, F., Al-Rashid, F., et al. (2022) Diagnostic Accuracy of Coronary Computed Tomography Angiography-Derived Fractional Flow Reserve (CT-FFR) in Patients before Liver Transplantation Using CT-FFR Machine Learning Algorithm. European Radiology, 32, 8761-8768.
https://doi.org/10.1007/s00330-022-08921-1
[35] Saba, L., Yuan, C., Hatsukami, T.S., et al. (2018) Carotid Artery Wall Imaging: Perspective and Guidelines from the ASNR Vessel Wall Imaging Study Group and Expert Consensus Recommendations of the American Society of Neuroradiology. American Journal of Neuroradiology, 39, E9-E31.
https://doi.org/10.3174/ajnr.A5488
[36] Van Der Kolk, A.G., Hendrikse, J., Brundel, M., et al. (2013) Multi-Sequence Whole-Brain Intracranial Vessel Wall Imaging at 7.0 Tesla. European Radiology, 23, 2996-3004.
https://doi.org/10.1007/s00330-013-2905-z
[37] Yang, H., Zhang, X., Qin, Q., et al. (2016) Improved Cerebrospinal Fluid Suppression for Intracranial Vessel Wall MRI. Journal of Magnetic Resonance Imaging, 44, 665-672.
https://doi.org/10.1002/jmri.25211
[38] Fan, Z., Zhang, Z., Chung, Y.C., et al. (2010) Carotid Arterial Wall MRI at 3T Using 3D Variable-Flip-Angle Turbo Spin-Echo (TSE) with Flow-Sensitive Dephasing (FSD). Journal of Magnetic Resonance Imaging, 31, 645-654.
https://doi.org/10.1002/jmri.22058
[39] Shi, F., Yang, Q., Guo, X., et al. (2019) Intracranial Vessel Wall Segmentation Using Convolutional Neural Networks. IEEE Transactions on Biomedical Engineering, 66, 2840-2847.
https://doi.org/10.1109/TBME.2019.2896972
[40] Niu, P.P., Yu, Y., Zhou, H.W., et al. (2016) Vessel Wall Differences Between Middle Cerebral Artery and Basilar Artery Plaques on Magnetic Resonance Imaging. Scientific Reports, 6, Article No. 38534.
https://doi.org/10.1038/srep38534
[41] Wan, L., Li, H., Zhang, L., et al. (2022) Automated Morphologic Analysis of Intracranial and Extracranial Arteries Using Convolutional Neural Networks. British Journal of Radiology, 95, Article ID: 20210031.
https://doi.org/10.1259/bjr.20210031
[42] Gao, S., Van’t Klooster, R., Kitslaar, P.H., et al. (2017) Learning-Based Automated Segmentation of the Carotid Artery Vessel Wall in Dual-Sequence MRI Using Subdivision Surface Fitting. Medical Physics, 44, 5244-5259.
https://doi.org/10.1002/mp.12476
[43] Wu, J., Xin, J., Yang, X., et al. (2023) Segmentation of Carotid Artery Vessel Wall and Diagnosis of Carotid Atherosclerosis on Black Blood Magnetic Resonance Imaging with Multi-Task Learning. Medical Physics, 51, 1775-1797.
https://doi.org/10.1002/mp.16728
[44] Azzopardi, C., Camilleri, K.P. and Hicks, Y.A. (2020) Bimodal Automated Carotid Ultrasound Segmentation Using Geometrically Constrained Deep Neural Networks. IEEE Journal of Biomedical and Health Informatics, 24, 1004-1015.
https://doi.org/10.1109/JBHI.2020.2965088
[45] Liu, J., Cui, X., Wang, D., et al. (2017) Relationship of Thyroid Function with Intracranial Arterial Stenosis and Carotid Atheromatous Plaques in Ischemic Stroke Patients with Euthyroidism. Oncotarget, 8, 46532-46539.
https://doi.org/10.18632/oncotarget.14883
[46] Boyd, C., Brown, G., Kleinig, T., et al. (2021) Machine Learning Quantitation of Cardiovascular and Cerebrovascular Disease: A Systematic Review of Clinical Applications. Diagnostics, 11, Article 551.
https://doi.org/10.3390/diagnostics11030551
[47] Hassan, M., Chaudhry, A., Khan, A., et al. (2014) Robust Information Gain Based Fuzzy C-Means Clustering and Classification of Carotid Artery Ultrasound Images. Computer Methods and Programs in Biomedicine, 113, 593-609.
https://doi.org/10.1016/j.cmpb.2013.10.012
[48] Huang, X., Zhang, Y., Meng, L., et al. (2018) Identification of Ultrasonic Echolucent Carotid Plaques Using Discrete Fréchet Distance between Bimodal Gamma Distributions. IEEE Transactions on Biomedical Engineering, 65, 949-955.
https://doi.org/10.1109/TBME.2017.2676129
[49] Roy-Cardinal, M.H., Destrempes, F., Soulez, G., et al. (2019) Assessment of Carotid Artery Plaque Components with Machine Learning Classification Using Homodyned-K Parametric Maps and Elastograms. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 66, 493-504.
https://doi.org/10.1109/TUFFC.2018.2851846
[50] Golemati, S., Patelaki, E., Gastounioti, A., et al. (2020) Motion Synchronisation Patterns of the Carotid Atheromatous Plaque from B-Mode Ultrasound. Scientific Reports, 10, Article No. 11221.
https://doi.org/10.1038/s41598-020-65340-2
[51] 廖熙妍, 邹佳妮, 黄文才. 冠状动脉CT血管成像联合人工智能在冠状动脉疾病诊疗中的应用进展[J]. 联勤军事医学, 2023, 37(10): 899-902.
[52] Hilbert, A., Ramos, L.A., Van Os, H.J.A., et al. (2019) Data-Efficient Deep Learning of Radiological Image Data for Outcome Prediction after Endovascular Treatment of Patients with Acute Ischemic Stroke. Computers in Biology and Medicine, 115, Article ID: 103516.
https://doi.org/10.1016/j.compbiomed.2019.103516
[53] Nam, Y., Jang, J., Lee, H.Y., et al. (2020) Estimating Age-Related Changes in Vivo Cerebral Magnetic Resonance Angiography Using Convolutional Neural Network. Neurobiology of Aging, 87, 125-131.
https://doi.org/10.1016/j.neurobiolaging.2019.12.008
[54] Johnson, K.M., Johnson, H.E., Zhao, Y., et al. (2019) Scoring of Coronary Artery Disease Characteristics on Coronary CT Angiograms by Using Machine Learning. Radiology, 292, 354-362.
https://doi.org/10.1148/radiol.2019182061
[55] Sarkar, D. and Saha, S. (2019) Machine-Learning Techniques for the Prediction of Protein-Protein Interactions. Journal of Biosciences, 44, Article No. 104.
https://doi.org/10.1007/s12038-019-9909-z