Page 189 - 磁共振成像2024年7期电子刊

综 述｜｜Reviews 磁共振成像 2024年7月第15卷第7期 Chin J Magn Reson Imaging, Jul, 2024, Vol. 15, No. 7

目前在头颈部动脉粥样斑块上的研究以影像组学为 [10] YAN J, WANG X F. Unsupervised and semi-supervised learning: the
next frontier in machine learning for plant systems biology[J]. Plant J,
主，DL 应用相对较少，两者内部运作机制不可见以 2022, 111(6): 1527-1538. DOI: 10.1111/tpj.15905.
[11] CHEN Y, LUO Y, HUANG W, et al. Machine-learning-based
及大量高维特征与病灶具体临床特征的相关性等问 classification of real-time tissue elastography for hepatic fibrosis in
题，在临床应用转化上存在一定问题 [50-51] 。其次，由 patients with chronic hepatitis B[J/OL]. Comput Biol Med, 2017, 89:
18-23 [2024-02-27]. https://pubmed.ncbi.nlm.nih.gov/28779596/. DOI:
于不同设备扫描、不同扫描参数、不同勾画感兴趣区 10.1016/j.compbiomed.2017.07.012.
[12] SRINIVAS S, YOUNG A J. Machine learning and artificial intelligence
的软件等因素 [52-53] ，都会导致模型的可重复性无法保 in surgical research[J]. Surg Clin North Am, 2023, 103(2): 299-316.
DOI: 10.1016/j.suc.2022.11.002.
证。最后多数研究使用单一的 ML 模型或 DL 算法， [13] MATTEUCCI G, PIASINI E, ZOCCOLAN D. Unsupervised learning
普遍数据集小、单中心，少数存在外部验证集，其模 of mid-level visual representations[J/OL]. Curr Opin Neurobiol, 2024,
84: 102834 [2024-02-27]. https://pubmed.ncbi.nlm.nih.gov/38154417/.
型的泛化能力有待验证。 DOI: 10.1016/j.conb.2023.102834.
[14] XUAN P, ZHANG X W, ZHANG Y, et al. Multi-type neighbors
未来可加入可视化如梯度热力图加强模型内部 enhanced global topology and pairwise attribute learning for
drug-protein interaction prediction[J/OL]. Brief Bioinform, 2022,
可视化或与医学知识图谱相结合推动 AI 的临床转 23(5): bbac120 [2024-02-27]. https://pubmed.ncbi.nlm.nih.gov/35514190/.
化，建立跨设备、跨平台的标准化扫描和数据采集流 DOI: 10.1093/bib/bbac120.
[15] LIN A. Artificial intelligence for high-risk plaque detection on carotid CT
程，确保模型的可重复性和泛化能力，多中心合作或 angiography[J/OL]. Atherosclerosis, 2023, 366: 40-41 [2024-02-27]. https://
pubmed.ncbi.nlm.nih.gov/36682983/. DOI: 10.1016/j.atherosclerosis.2023.
建立头颈动脉粥样硬化疾病数据库尽量加强模型的 01.006.
[16] THOMPSON A C, JAMMAL A A, MEDEIROS F A. A review of deep
泛化能力，以及与基因组学和生物标志物的结合，将 learning for screening, diagnosis, and detection of glaucoma progression
能够为头颈动脉粥样硬化疾病研究提供新方向，从 [J/OL]. Transl Vis Sci Technol, 2020, 9(2): 42 [2024-02-27]. https://
pubmed.ncbi.nlm.nih.gov/32855846/. DOI: 10.1167/tvst.9.2.42.
而为患者提供个性化治疗方案，提高其生活质量。 [17] WAGNER M W, NAMDAR K, BISWAS A, et al. Radiomics, machine
learning, and artificial intelligence-what the neuroradiologist needs to
作者利益冲突声明：全体作者均声明无利益冲突。 know[J]. Neuroradiology, 2021, 63(12): 1957-1967. DOI: 10.1007/s00234-
021-02813-9.
作者贡献声明：曾献军拟定本综述的写作思路， [18] MCBEE M P, AWAN O A, COLUCCI A T, et al. Deep learning in
对稿件内容进行修改，获得国家自然科学基金的资 radiology[J]. Acad Radiol, 2018, 25(11): 1472-1480. DOI: 10.1016/j.
acra.2018.02.018.
助；刘洁设计、起草和撰写稿件，获取、分析并解释本 [19] WANG M Q, ZHANG L Y, YU H X, et al. A deep learning network
based on CNN and sliding window LSTM for spike sorting[J/OL].
综述的参考文献；欧阳烽、吕联江、徐紫荷获取、分析 Comput Biol Med, 2023, 159: 106879 [2024-02-27]. https://pubmed.ncbi.
nlm.nih.gov/37080004/. DOI: 10.1016/j.compbiomed.2023.106879.
或解释本综述的参考文献，对稿件内容进行了修改 [20] LI Z W, LIU F, YANG W J, et al. A survey of convolutional neural
以及校对。全体作者都同意发表最后的修改稿，同 networks: analysis, applications, and prospects[J]. IEEE Trans Neural
Netw Learn Syst, 2022, 33(12): 6999-7019. DOI: 10.1109/TNNLS.2021.
意对本研究的所有方面负责，确保本研究的准确性 3084827.
[21] HATT M, LE REST C C, TIXIER F, et al. Radiomics: data are also
和诚信。 images[J/OL]. J Nucl Med, 2019, 60(suppl 2): 38S-44S [2024-02-27].
https://pubmed.ncbi.nlm.nih.gov/31481588/. DOI: 10.2967/jnumed.118.
220582.
参考文献[References] [22] MAYERHOEFER M E, MATERKA A, LANGS G, et al. Introduction
to radiomics[J]. J Nucl Med, 2020, 61(4): 488-495. DOI: 10.2967/
[1] FEIGIN V L, KRISHNAMURTHI R V, PARMAR P, et al. Update on jnumed.118.222893.
the global burden of ischemic and hemorrhagic stroke in 1990-2013: [23] LAMBIN P, RIOS-VELAZQUEZ E, LEIJENAAR R, et al. Radiomics:
the GBD 2013 study[J]. Neuroepidemiology, 2015, 45(3): 161-176. extracting more information from medical images using advanced
DOI: 10.1159/000441085. feature analysis[J]. Eur J Cancer, 2012, 48(4): 441-446. DOI: 10.1016/j.
[2] XIAO J Y, PADRICK M M, JIANG T, et al. Acute ischemic stroke ejca.2011.11.036.
versus transient ischemic attack: differential plaque morphological [24] CHEN Y F, CHEN Z J, LIN Y Y, et al. Stroke risk study based on deep
features in symptomatic intracranial atherosclerotic lesions[J/OL]. learning-based magnetic resonance imaging carotid plaque automatic
Atherosclerosis, 2021, 319: 72-78 [2024-02-27]. https://pubmed. ncbi. segmentation algorithm[J/OL]. Front Cardiovasc Med, 2023, 10:
nlm.nih.gov/33486353/. DOI: 10.1016/j.atherosclerosis.2021.01.002. 1101765 [2024-02-27]. https://pubmed. ncbi. nlm. nih. gov/36910524/.
[3] WANG Y J, ZHAO X Q, LIU L P, et al. Prevalence and outcomes of DOI: 10.3389/fcvm.2023.1101765.
symptomatic intracranial large artery stenoses and occlusions in China: [25] SABA L, YUAN C, HATSUKAMI T S, et al. Carotid artery wall
the Chinese Intracranial Atherosclerosis (CICAS) Study[J]. Stroke, imaging: perspective and guidelines from the ASNR vessel wall
2014, 45(3): 663-669. DOI: 10.1161/STROKEAHA.113.003508. imaging study group and expert consensus recommendations of the
[4] KIM Y D, CHA M J, KIM J, et al. Increases in cerebral atherosclerosis American society of neuroradiology[J/OL]. AJNR Am J Neuroradiol,
according to CHADS2 scores in patients with stroke with nonvalvular atrial 2018, 39(2): E9-E31 [2024-02-27]. https://pubmed. ncbi. nlm. nih. gov/
fibrillation[J]. Stroke, 2011, 42(4): 930-934. DOI: 10.1161/STROKEAHA. 29326139/. DOI: 10.3174/ajnr.A5488.
110.602987. [26] SHI F, YANG Q, GUO X H, et al. Intracranial vessel wall segmentation
[5] HAMET P, TREMBLAY J. Artificial intelligence in medicine[J/OL]. using convolutional neural networks[J]. IEEE Trans Biomed Eng, 2019,
Metabolism, 2017, 69: S36-S40 [2024-02-27]. https://pubmed.ncbi.nlm. 66(10): 2840-2847. DOI: 10.1109/TBME.2019.2896972.
nih.gov/28126242/. DOI: 10.1016/j.metabol.2017.01.011. [27] WAN L W, LI H X, ZHANG L, et al. Automated morphologic analysis
[6] KANG D W, KIM D Y, KIM J, et al. Emerging concept of intracranial of intracranial and extracranial arteries using convolutional neural
arterial diseases: the role of high resolution vessel wall MRI[J]. J networks[J/OL]. Br J Radiol, 2022, 95(1139): 20210031 [2024-02-27].
Stroke, 2024, 26(1): 26-40. DOI: 10.5853/jos.2023.02481. https://pubmed.ncbi.nlm.nih.gov/36018822/. DOI: 10.1259/bjr.20210031.
[7] SUI Y, SUN J L, CHEN Y, et al. Multimodal MRI study of the [28] WU J Y, XIN J M, YANG X F, et al. Deep morphology aided diagnosis
relationship between plaque characteristics and hypoperfusion in network for segmentation of carotid artery vessel wall and diagnosis of
patients with transient ischemic attack[J/OL]. Front Neurol, 2023, 14: carotid atherosclerosis on black-blood vessel wall MRI[J]. Med Phys,
1242923 [2024-02-27]. https://pubmed.ncbi.nlm.nih.gov/37840913/. DOI: 2019, 46(12): 5544-5561. DOI: 10.1002/mp.13739.
10.3389/fneur.2023.1242923. [29] WU J Y, XIN J M, YANG X F, et al. Segmentation of carotid artery
[8] CHOI Y J, JUNG S C, LEE D H. Vessel wall imaging of the intracranial vessel wall and diagnosis of carotid atherosclerosis on black blood
and cervical carotid arteries[J]. J Stroke, 2015, 17(3): 238-255. DOI: magnetic resonance imaging with multi-task learning[J]. Med Phys,
10.5853/jos.2015.17.3.238. 2024, 51(3): 1775-1797. DOI: 10.1002/mp.16728.
[9] TANDON V, SENTHILVELAN S, SREEDHARAN S E, et al. [30] TIAN X, FANG H, LAN L F, et al. Risk stratification in symptomatic
High-resolution MR vessel wall imaging in determining the stroke aetiology intracranial atherosclerotic disease with conventional vascular risk
and risk stratification in isolated middle cerebral artery disease[J]. factors and cerebral haemodynamics[J]. Stroke Vasc Neurol, 2023,
Neuroradiology, 2022, 64(8): 1569-1577. DOI: 10.1007/s00234-021- 8(1): 77-85. DOI: 10.1136/svn-2022-001606.
02891-9. [31] PRABHAKARAN S, LIEBESKIND D S, COTSONIS G, et al.

·182 · https://www.chinesemri.com