Using Neural Networks Algorithm in Ischemic Stroke Diagnosis: A Systematic Review

Figures & data

Table 1 Study Inclusion and Exclusion Criteria

Inclusion	Exclusion
Studies aimed to apply neural networks in ischemic stroke diagnosis/detection Studies included adults with ischemic stroke diagnosis (18 years or older) Studies published in the English language Studies published between January 2018 and December 2022	Studies including children or adolescents <18 years old Conference proceedings, abstracts, review articles, pilot studies, protocol, dissertation, letter to the editor, brief report, or statement paper Involve animal sample Studies where the primary endpoint is not ischemic stroke diagnosis (eg, mortality prediction after stroke, stroke outcomes)

Figure 1 Flow chart diagram displaying the selection method of qualified studies.

Notes: Adapted from Moher D, Liberati A, Tetzlaff J, Altman DG, Prisma Group. Reprint--Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Annals of internal medicine. 2009;151(4):264–269.²²

Table 2 The Characteristics of the Included Studies

Characteristic	Number^a	Percentage (%)
Publication Year
2022	3	33.33%
2021	2	22.22%
2020	3	33.33%
2019	1	11.11%
Country
China	5	55.55%
Japan	2	22.22%
USA	1	11.11%
Finland	1	11.11%
Total Sample Size (n)
1–25	1	11.11%
>25–100	2	22.22%
>100–300	3	33.33%
>300	3	33.33%
Target Population
Acute Ischemic Stroke	7	77.77%
Ischemic Stroke	1	11.11%
Other^b	1	11.11%

Notes: ^aThe number of included studies. ^bIschemic Penumbra (IP) in Early Acute Cerebral Infarction (ACI).

Figure 2 The number of included studies by features.

Abbreviations: DWI, Diffusion-Weighted Imaging; CTA+, 8-second delay after Computed Tomography Angiography; CT, Computed Tomography; MRI, Magnetic Resonance Imaging; CTP, Computed Tomography Perfusion; CTA, Computed Tomography Angiography; NCCT, Non-Contrast Computed Tomography.

Table 3 The Optimal Results of Included Studies by the Algorithm

Algorithm Reference	3D-CNNs^a	Two-Stage DCNNs^b	GL-HOSVD^c	DCNNs^d	AD-CNNnet^e
[24]	X
[25]		X
[26]			X
[27]	X
[28]				X
[29]				X
[30]					X
[31]		X
[23]				X*
Total included studies by algorithm (n, %)	2 (22.22%)	2 (22.22%)	1 (11.11%)	3 (33.33%)	1 (11.11%)

Notes: ^aThree-dimensional convolutional neural networks, ^bTwo-stage deep convolutional neural networks, ^cThe local higher-order singular value decomposition denoising algorithm, ^dDeep Convolutional Neural Networks, ^eA new deconvolution network model based on deep learning. *Combined Faster R-CNN, YOLOV3, and SSD algorithm.

Moher D, Liberati A, Tetzlaff J, Altman DG; Prisma Group. Reprint--Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151(4):264–269. doi:10.7326/0003-4819-151-4-200908180-00135

 PubMed Web of Science ®Google Scholar

Oman O, Makela T, Salli E, Savolainen S, Kangasniemi M. 3D convolutional neural networks applied to CT angiography in the detection of acute ischemic stroke. Eur Radiol Exp. 2019;3(1)8. doi:10.1186/s41747-019-0085-6

 PubMedGoogle Scholar

Nishio M, Koyasu S, Noguchi S, et al. Automatic detection of acute ischemic stroke using non-contrast computed tomography and two-stage deep learning model. Comput Methods Programs Biomed. 2020;196:105711. doi:10.1016/j.cmpb.2020.105711

 PubMed Web of Science ®Google Scholar

Sheng H, Wang XL, Jiang MP, Zhang ZS. Deep learning-based diffusion-weighted magnetic resonance imaging in the diagnosis of ischemic penumbra in early cerebral infarction. Contrast Media Mol Imaging. 2022;2022:6270700. doi:10.1155/2022/6270700

 PubMed Web of Science ®Google Scholar

Wang CY, Shi Z, Yang M, et al. Deep learning-based identification of acute ischemic core and deficit from non-contrast CT and CTA. J Cereb Blood Flow Metab. 2021;41(11):3028–3038. doi:10.1177/0271678x211023660

 PubMed Web of Science ®Google Scholar

Stib MT, Vasquez J. Detecting large vessel occlusion at multiphase CT angiography by using a deep convolutional neural network. Radiology 2020;297(3):640–649. doi:10.1148/radiol.2020200334

 PubMed Web of Science ®Google Scholar

Shinohara Y, Takahashi N, Lee Y, Ohmura T, Kinoshita T. Development of a deep learning model to identify hyperdense MCA sign in patients with acute ischemic stroke. Jpn J Radiol. 2020;38(2):112–117. doi:10.1007/s11604-019-00894-4

 PubMed Web of Science ®Google Scholar

Yang Y, Yang JJ, Feng J, Wang Y. Early diagnosis of acute ischemic stroke by brain computed tomography perfusion imaging combined with head and neck computed tomography angiography on deep learning algorithm. Contrast Media Mol Imaging. 2022;2022:5373585. doi:10.1155/2022/5373585

 PubMed Web of Science ®Google Scholar

Lu J, Zhou YR, Lv WZ, et al. Identification of early invisible acute ischemic stroke in non-contrast computed tomography using two-stage deep-learning model. Theranostics. 2022;12(12):5564–5573. doi:10.7150/thno.74125

 PubMed Web of Science ®Google Scholar

Zhang SJ, Xu SH, Tan LW, Wang HY, Meng JL. Stroke lesion detection and analysis in MRI images based on deep learning. J Healthc Eng. 2021;2021:5524769. doi:10.1155/2021/5524769

 Web of Science ®Google Scholar