Multi-dimensional segmentation model and system development for recognizing small lesions after acute ischemic stroke

Xicheng CHEN; Zeliang WEI; Wei YE; Haojia WANG; Yongjun TAO; Dong YI; Yazhou WU

doi:10.16016/j.2097-0927.202212169

您当前的位置：

首页 >

文章列表页 >

Multi-dimensional segmentation model and system development for recognizing small lesions after acute ischemic stroke

Biomedical Engineering | 更新时间：2026-01-09

- Multi-dimensional segmentation model and system development for recognizing small lesions after acute ischemic stroke
- Journal of Army Medical University Vol. 45, Issue 6, Pages: 570-578(2023)
- 作者机构：
  
  1.400038 重庆，陆军军医大学(第三军医大学)军事预防医学系军队卫生统计学教研室
  2.318000 浙江台州，浙江省台州市立医院神经内科
- 作者简介：
- 基金信息：
  
  the National Natural Science Foundation of China(82173621);the National Natural Science Foundation of China(81872716)
- DOI：10.16016/j.2097-0927.202212169
  CLC： R195.1;R741.04;R743.3
- Received：28 December 2022，
  
  Revised：2023-01-15，
  
  Published：30 March 2023
- 稿件说明：
移动端阅览
Xicheng CHEN, Zeliang WEI, Wei YE, et al. Multi-dimensional segmentation model and system development for recognizing small lesions after acute ischemic stroke[J]. Journal of Army Medical University, 2023, 45(6): 570-578.
DOI：

Xicheng CHEN, Zeliang WEI, Wei YE, et al. Multi-dimensional segmentation model and system development for recognizing small lesions after acute ischemic stroke[J]. Journal of Army Medical University, 2023, 45(6): 570-578. DOI： 10.16016/j.2097-0927.202212169.

摘要

目的

开发一种基于深度学习的多维度分割系统，在磁共振成像(magnetic resonance imaging，MRI)影像中识别微小病灶，为急性缺血性脑卒中(acute ischemic stroke，AIS)的诊治提供决策依据。

方法

提取并融合2D、3D网络的特征，引入了联合损失函数，进而提出了一种新的2.5D方法——多维度多尺度注意力增强网络(multi-dimensional multi-scale attention enhanced network，MMAE-Net)。在AIS分割数据集(训练集171例、测试集43例)上训练和测试了所提方法的性能，并与其他方法进行比较。

结果

相较于2D、3D网络，2.5D网络(MMAE-Net)在各项评价指标上均可取得最优异的分割性能，获得了81.25%的Dice相似系数(dice similarity coefficient，DSC)和84.82%的灵敏度(sensitivity，SEN)。相较于U-Net、ResU-Net、DenseU-Net、AttentionU-Net和SEgmentation TRansformer(SETR)等经典方法和既往研究，MMAE-Net取得了更优异的分割性能。我们还开发了可视化自动分割系统以提升方法的实践和推广能力。

结论

本文基于2D、3D特征融合的理念设计了2.5D多维度分割模型MMAE-Net，其在MRI微小病灶的识别中取得了优异性能，为AIS疾病诊治提供了有效的解决方案。

Abstract

Objective

To develop a deep learning-based multi-dimensional segmentation system to recognize small lesions in magnetic resonance imaging (MRI) images in order to provide a decision-making basis for the diagnosis and treatment of acute ischemic stroke (AIS).

Methods

We extracted and fused the features from 2D and 3D network

introduced the joint loss function

and then proposed a new 2.5D method

a multi-dimensional multi-scale attention enhanced network (MMAE-Net). On AIS segmentation datasets (171 cases in training set and 43 cases in testing set)

the proposed method was trained and tested

and its performance was compared with other methods.

Results

When compared to 2D and 3D networks

our 2.5D network (MMAE-Net) achieved the best segmentation performance in all evaluation indicators

with a dice similarity coefficient (DSC) of 81.25% and a sensitivity of 84.82%. MMAE-Net achieved better segmentation performance when compared to U-Net

ResU-Net

DenseU-Net

AttentionU-Net

Segmentation TRansformer (SETR)

and other classical methods and previous research. In addition

we also created a visual and automated clinical application system to improve the practical and promotive capability of methods.

Conclusion

Based on fusing the features from 2D and 3D network

a 2.5D multi-dimensional segmentation model MMAE-Net is developed

which has achieved excellent performance in the recognition of MRI small lesions and provides an effective solution for the diagnosis and treatment of AIS diseases.

关键词

Keywords

references

AVESTA A, HOSSAIN S, LIN M, et al . Comparing 3D, 2.5D, and 2D approaches to brain image segmentation[J ] . Bioengineering (Basel), 2023, 10(2): 181. DOI: 10.3390/bioengineering10020181.

PREETHA C J, MEREDIG H, BRUGNARA G, et al . Deep-learning-based synthesis of post-contrast T1-weighted MRI for tumour response assessment in neuro-oncology: a multicentre, retrospective cohort study[J ] . Lancet Digit Health, 2021, 3 (12): e784-e794. DOI: 10.1016/S2589-7500(21)00205-3.

DU X, MA K, SONG Y. AGMR-Net: Attention-guided multiscale recovery framework for stroke segmentation[J]. Comput Med Imaging Graph, 2022, 101: 102120. DOI: 10.1016/j.compmedimag.2022.102120.

EL-HARIRI H, SOUTO MAIOR NETO L A, CIMFLOVA P, et al . Evaluating nnU-Net for early ischemic change segmentation on non-contrast computed tomography in patients with Acute Ischemic Stroke[J ] . Comput Biol Med, 2022, 141: 105033. DOI: 10.1016/j.compbiomed.2021.105033.

QIU Q, YANG Z, WU S, et al . Automatic segmentation of hippocampus in hippocampal sparing whole brain radiotherapy: A multitask edge-aware learning[J ] . Med Phys, 2021, 48(4): 1771-1780. DOI: 10.1002/mp.14760.

LIU D, SHENG N, HE T, et al . SGEResU-Net for brain tumor segmentation[J ] . Math Biosci Eng, 2022, 19(6): 5576-5590. DOI: 10.3934/mbe.2022261.

WEI Z, CHEN X, HUANG J, et al . Construction of a medical micro-object cascade network for automated segmentation of cerebral microbleeds in susceptibility weighted imaging[J ] . Front Bioeng Biotechnol, 2022, 10: 937314. DOI: 10.3389/fbioe.2022.937314.

ANWAR S M, MAJID M, QAYYUM A, et al . Medical image analysis using convolutional neural networks: a review[J ] . J Med Syst, 2018, 42(11): 226. DOI: 10.1007/s10916-018-1088-1.

FU Y, LEI Y, WANG T, et al . A review of deep learning based methods for medical image multi-organ segmentation[J ] . Phys Med, 2021, 85: 107-122. DOI: 10.1016/j.ejmp.2021.05.003.

FALK T, MAI D, BENSCH R, et al . U-Net: deep learning for cell counting, detection, and morphometry[J ] . Nat Methods, 2019, 16(1): 67-70. DOI: 10.1038/s41592-018-0261-2.

YANG T, SONG J, LI L, et al . Improving brain tumor segmentation on MRI based on the deep U-net and residual units[J ] . J Xray Sci Technol, 2020, 28(1): 95-110. DOI: 10.3233/XST-190552.

GUAN S, KHAN A A, SIKDAR S, et al . Fully dense UNet for 2-D sparse photoacoustic tomography artifact removal[J ] . J Biomed Health Inform, 2020, 24(2): 568-576. DOI: 10.1109/JBHI.2019.2912935.

SCHLEMPER J, OKTAY O, SCHAAP M, et al . Attention gated networks: Learning to leverage salient region s in medical images[J ] . Med Image Anal, 2019, 53: 197-207. DOI: 10.1016/j.media.2019.01.012.

ZHENG S, LU J, ZHAO H, et al . Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers[C ] //2021 IEEE/CVF conference on computer vision and pattern recognition(CVPR). Nashville: IEEE, 2021: 6877-6886. DOI: 10.1109/CVPR46437.2021.00681 http://dx.doi.org/10.1109/CVPR46437.2021.00681 ..

LIN Q, LI S, WANG R, et al . Research on small target detection technology based on the MPH-SSD algorithm[J ] . Comput Intell Neurosci, 2022, 2022: 1-13. DOI: 10.1155/2022/9654930.

MELVILLE-SMITH A, FINN A, UZAIR M, et al . Exploration of motion inhibition for the suppression of false positives in biologically inspired small target detection algorithms from a moving platform[J ] . Biol Cybern, 2022, 116(5-6): 661-685. DOI: 10.1007/s00422-022-00950-9.

NEMOTO T, FUTAKAMI N, YAGI M, et al . Efficacy evaluation of 2D, 3D U-Net semantic segmentation and atlas-based segmentation of normal lungs excluding the trachea and main bronchi[J ] . J Radiat Res, 2020, 61(2): 257-264. DOI: 10.1093/jrr/rrz086.

HAN L, CHEN Y H, LI J M, et al . Liver segmentation with 2.5D perpendicular UNets[J ] . Comput Electr Eng, 2021, 91: 107118. DOI: 10.1016/j.compeleceng.2021.107118.

ZHANG C, HUA Q Q, CHU Y Y, et al . Liver tumor segmentation using 2.5D UV-Net with multi-scale convolution[J ] . Comput Biol Med, 2021, 133: 104424. DOI: 10.1016/j.compbiomed.2021.104424.

DA CRUZ L B, ALVES DIAS JÚNIOR D, DINIZ J O B, et al . Kidney tumor segmentation from computed tomography images using DeepLabv3+ 2.5D model[J ] . Expert Syst Appl, 2022, 192: 116270. DOI: 10.1016/j.eswa.2021.116270.

BAYRAM A F, DERIN A, GURKAN C, et al . Analysis of the effects of segmentation networks and loss functions in ischemic stroke lesion segmentation[J ] . Eur J Scie Tech, 2022 (40): 82-87. DOI: 10.31590/ejosat.1173070.

SHIN H, AGYEMAN R, RAFIQ M, et al . Automated segmentation of chronic stroke lesion using efficient U-Net architecture[J ] . Biocybern Biomed Eng, 2022, 42(1): 285-294. DOI: 10.1016/j.bbe.2022.01.002.

SOLTANPOUR M, GREINER R, BOULANGER P, et al . Improvement of automatic ischemic stroke lesion segmentation in CT perfusion maps using a learned deep neural network[J ] . Comput Biol Med, 2021, 137: 104849. DOI: 10.1016/j.compbiomed.2021.104849.

XUE Y Z, FARHAT F G, BOUKRINA O, et al . A multi-path 2.5 dimensional convolutional neural network system for segmenting stroke lesions in brain MRI images[J ] . Neuroimage Clin, 2020, 25: 102118. DOI: 10.1016/j.nicl.2019.102118.

KUANG H L, MENON B K, QIU W. Automated stroke lesion segmentation in non-contrast CT scans using dense multi-path contextual generative adversarial network[J]. Phys Med Biol, 2020, 65(21): 215013. DOI: 10.1088/1361-6560/aba166.

THIYAGARAJAN S K, MURUGAN K. A systematic review on techniques adapted for segmentation and classification of ischemic stroke lesions from brain MR images[J]. Wirel Pers Commun Int J, 2021, 118(2): 1225-1244. DOI: 10.1007/s11277-021-08069-z.

LIEW S L, LO B P, DONNELLY M R, et al . A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms[J ] . Sci Data, 2022, 9(1): 320. DOI: 10.1038/s41597-022-01401-7.

CUI L Y, FAN Z Y, YANG Y J, et al . Deep learning in ischemic stroke imaging analysis: a comprehensive review[J ] . Biomed Res Int, 2022, 2022: 2456550. DOI: 10.1155/2022/2456550.

OTTESEN J A, YI D, TONG E, et al . 2.5D and 3D segmentation of brain metastases with deep learning on multinational MRI data[J ] . Frontiers in Neuroinformatics, 2023, 16: 1056068. DOI: 10.3389/fninf.2022.1056068.

ZHANG X Y, LI L H, DI D L, et al . SERNet: squeeze and excitation residual network for semantic segmentation of high-resolution remote sensing images[J ] . Remote Sens, 2022, 14(19): 4770. DOI: 10.3390/rs14194770.

TYAGI S, TALBAR S N. CSE-GAN: a 3D conditional generative adversarial network with concurrent squeeze-and-excitation blocks for lung nodule segmentation[J]. Comput Biol Med, 2022, 147: 105781. DOI: 10.1016/j.compbiomed.2022.105781.

WANG J K, LV P Q, WANG H Y, et al . SAR-U-Net: squeeze-and-excitation block and atrous spatial pyramid pooling based residual U-Net for automatic liver segmentation in Computed Tomography[J ] . Comput Methods Programs Biomed, 2021, 208: 106268. DOI: 10.1016/j.cmpb.2021.106268.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Automatic segmentation of female urine control anatomical elements and related structures in MRI images based on deep learning

Integrative model combining deep learning, clinical and radiomic features enhances EGFR mutation prediction in non-small cell lung cancer

Predictive performance of CT images-based 3D ResNet18 model for identifying lung tuberculosis drug resistance

Advances in deep learning and MRI co-assisted genotyping of brain gliomas

Ultrasound thyroid nodule segmentation algorithm based on wavelet transform and CNN-Transformer

Related Author

ZHANG Ziqin

WU Yi

ZHANG Xiaoqin

XU Zhou

LEI Ling

WANG Yanzhou

WANG Yan

YANG Taotao

Related Institution

School of Mathematical Sciences， Chongqing Normal University

Department of Digital Medicine， Faculty of Biomedical Engineering and Imaging Medicine， Army Medical University （Third Military Medical University）

Department of Gynecology， Anshun People’s Hospital

Department of Obstetrics and Gynecology， First Affiliated Hospital， Army Medical University （Third Military Medical University）

Department of Diagnostic Radiology， Eastern Theatre General Hospital

⁰