Share：
Share this content in WeChat
X
[Chinese] [PDF] 191 9
• Review •
Research progress of Transformer in MRI image segmentation of brain tumors
CHEN Lei  LI Guangyu  YANG Feng  CAI Jingxin  GAO Mengyao 

DOI:10.12015/issn.1674-8034.2025.08.027.


[Abstract] Accurate segmentation of brain tumors is crucial, but traditional convolutional neural networks are difficult to model long-range dependencies in magnetic resonance imaging (MRI) due to local receptive field limitations, which affects the segmentation accuracy of tumors with high heterogeneity and blurred boundaries. Transformer provides a new approach for this through its global self-attention mechanism. This article reviews the progress of Transformer in brain tumor MRI segmentation, focusing on analyzing the improvements of Transformer models in key technologies such as hierarchical attention, encoder-decoder structures, and residual connections, and exploring innovative strategies for multimodal fusion, handling missing modalities, lightweight design, and the attention mechanism itself; although Transformers have significantly improved accuracy, they still face challenges such as data scarcity, robustness to modal loss, class imbalance, high computational costs, and insufficient interpretability, necessitating future focus on efficient data utilization, modal elasticity modeling, topology-aware optimization, lightweight and interpretability enhancement, and other directions. This article systematically reviews the current research status of Transformer in the field of brain tumor MRI image segmentation, summarizes the limitations of current research, and points out the future research directions. The aim is to provide a systematic reference for a deeper understanding of its technological evolution, core challenges, and development trends.
[Keywords] Transformer model；segmentation of brain tumors；magnetic resonance imaging；multimodal；attention mechanism；lightweight design

CHEN Lei1*   LI Guangyu2   YANG Feng1   CAI Jingxin2   GAO Mengyao2  

1 Department of Assets and Equipment, the Affiliated Hospital of Shandong University of Chinese Medicine, Jinan 250014, China

2 School of Medical Information Engineering, Shandong University of Chinese Medicine, Jinan 250355, China

Corresponding author: CHEN L, E-mail: szysbc@163.com

Conflicts of interest   None.

Received  2025-06-20
Accepted  2025-08-05
DOI: 10.12015/issn.1674-8034.2025.08.027
DOI:10.12015/issn.1674-8034.2025.08.027.

[1]
ZHANG X L, CHEN Q, HE H Z, et al. Generative learning-based lightweight MRI brain tumor segmentation with missing modalities[J]. Expert Syst Appl, 2025, 261: 125478 [2025-06-19]. https://www.sciencedirect.com/science/article/abs/pii/S0957417424023455. DOI: 10.1016/j.eswa.2024.125478.
[2]
GHADIMI D J, VAHDANI A M, KARIMI H, et al. Deep learning-based techniques in glioma brain tumor segmentation using multi-parametric MRI: a review on clinical applications and future outlooks[J]. J Magn Reson Imaging, 2025, 61(3): 1094-1109. DOI: 10.1002/jmri.29543.
[3]
BENZORGAT N, XIA K W, BENZORGAT M N E. Enhancing brain tumor MRI classification with an ensemble of deep learning models and transformer integration[J/OL]. PeerJ Comput Sci, 2024, 10: e2425 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/39650528/. DOI: 10.7717/peerj-cs.2425.
[4]
DOSOVITSKIY A, BEYER L, Kolesnikov A, et al. An Image is Worthx 16x16 Words: Transformers for Image Recognition at Scale[J/OL]. arXiv, 2020 [2025-06-19]. https://arxiv.org/abs/2010.11929. DOI: 10.48550/arXiv.2010.11929.
[5]
CAO H, WANG Y Y, CHEN J, et al. Swin-unet: unet-like pure transformer for medical image segmentation[M]//Computer Vision-ECCV 2022 Workshops. Cham: Springer Nature Switzerland, 2023: 205-218. DOI: 10.1007/978-3-031-25066-8_9.
[6]
WANG W X, CHEN C, DING M, et al. TransBTS: multimodal brain tumor segmentation using transformer[C]//Medical Image Computing and Computer Assisted Intervention-MICCAI 2021. Cham: Springer, 2021: 109-119. DOI: 10.1007/978-3-030-87193-2_11.
[7]
APPASAMI G, SAVARIMUTHU N. Federated learning for secure medical MRI brain tumor image classification[J/OL]. Eur Phys J Spec Top, 2025 [2025-06-19]. https://link.springer.com/article/10.1140/epjs/s11734-025-01516-z. DOI: 10.1140/epjs/s11734-025-01516-z.
[8]
HOU J F, KING T Z, CHEN H B, et al. Concurrent brain structural and functional alterations in the thalamus of adult survivors of childhood brain tumors: a multimodal MRI study[J/OL]. Brain Res Bull, 2024, 211: 110937 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/38570077/. DOI: 10.1016/j.brainresbull.2024.110937.
[9]
MUNTAQIM M D Z, SMRITY T A. Federated learning framework for brain tumor detection using MRI images in non-IID data distributions[J/OL]. J Imaging Inform Med, 2025 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/40128502/. DOI: 10.1007/s10278-025-01484-9.
[10]
LI Z H, WANG W. Research progress of medical image segmentation method based on deep learning[J]. Electron Sci Technol, 2024, 37(1): 72-80. DOI: 10.16180/j.cnki.issn1007-7820.2024.01.011.
[11]
LI Z W, CAO H, YANG F, et al. Research progress of brain tumor segmentation based on convolutional neural network[J]. Laser Optoelectron Prog, 2021, 58(24): 2400003. DOI: 10.3788/LOP202158.2400003.
[12]
VASWANI A, Shazeer N, Parmar N. Attention is all you need[J/OL]. arXiv, 2017 [2025-06-19]. https://arxiv.org/abs/1706.03762. DOI: 10.48550/arXiv.1706.03762.
[13]
LIU Z, LIN Y T, CAO Y, et al. Swin transformer: hierarchical vision transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). October 10-17, 2021, Montreal, QC, Canada. IEEE, 2021: 9992-10002. DOI: 10.1109/ICCV48922.2021.00986.
[14]
ZHOU H Y, GUO J, ZHANG Y, et al. nnFormer: Interleaved Transformer for Volumetric Segmentation[J/OL]. arXiv, 2021 [2025-06-19]. https://arxiv.org/abs/2109.03201. DOI: 10.48550/arXiv.2109.03201.
[15]
A P M, EVURI N, PAKALA S R, et al. 3D swin-res-SegNet: a hybrid transformer and CNN model for brain tumor segmentation using MRI scans[J/OL]. J Inst Eng Ind Ser B, 2024 [2025-06-19]. https://link.springer.com/article/10.1007/s40031-024-01166-0. DOI: 10.1007/s40031-024-01166-0.
[16]
HATAMIZADEH A, NATH V, TANG Y C, et al. Swin UNETR: Swin Transformers forSemantic Segmentation ofBrain Tumors inMRI Images[C]//Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Cham: Springer, 2022: 272-284. DOI: 10.1007/978-3-031-08999-2_22.
[17]
LI Z R, SILAMU W, WANG Y Z, et al. DenseTrans: multimodal brain tumor segmentation using swin transformer[J/OL]. IEEE Access, 2023, 11: 42895-42908 [2025-06-19]. https://ieeexplore.ieee.org/abstract/document/10113635. DOI: 10.1109/ACCESS.2023.3272055.
[18]
DAI A, SONG Y N, XU R H, et al. Study on brain tumor segmentation algorithm based on improved TransBTS[J]. Exp Technol Manag, 2023, 40(8): 13-19, 39. DOI: 10.16791/j.cnki.sjg.2023.08.003.
[19]
LI J, WANG W, CHEN C, et al. TransBTSV2: Wider Instead of Deeper Transformer for Medical Image Segmentation[J/OL]. ArXiv, 2022, abs/2201.12785 [2025-06-19]. https://arxiv.org/abs/2201.12785. DOI: 10.48550/arXiv.2201.12785.
[20]
KANG M, F F T, R C, et al. A Multimodal Feature Distillation with CNN-Transformer Network for Brain Tumor Segmentation with Incomplete Modalities[J/OL]. ARXiv,2024 [2025-06-19]. https://arxiv.org/abs/2404.14019. DOI: 10.48550/arxiv.2404.14019.
[21]
LEI X, YU X, WU H, et al. Clinical Knowledge-Based Hybrid Swin Transformer for Brain Tumor Segmentation[J]. Computers, Materials & Continua, 2023, 76(9): 3797-3811. DOI: 10.32604/cmc.2023.042069.
[22]
LYU Y, TIAN X L. MWG-UNet: hybrid deep learning framework for lung fields and heart segmentation in chest X-ray images[J/OL]. Bioengineering (Basel), 2023, 10(9): 1091 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/37760193/. DOI: 10.3390/bioengineering10091091.
[23]
RASOOL N, IQBAL BHAT J, AHMAD WANI N, et al. TransResUNet: revolutionizing glioma brain tumor segmentation through transformer-enhanced residual UNet[J/OL]. IEEE Access, 2024, 12: 72105-72116. [2025-06-19]. https://ieeexplore.ieee.org/document/10534774. DOI: 10.1109/ACCESS.2024.3402947.
[24]
LUO K, DING D R, PANG F X. LFC-UNETR: a full-convolutional network for multimodal brain tumor segmentation[J/OL]. J Chin Comput Syst, 2024: 1-9. (2024-09-26). https://kns.cnki.net/kcms/detail/21.1106.TP.20240926.1151.006.html.
[25]
HUA K W, FANG X J. Fusing dual adversarial learning CNN-transformer model for multi-modal MRI brain tumor image segmentation[J]. J Hubei Minzu Univ Nat Sci Ed, 2023, 41(4): 479-488. DOI: 10.13501/j.cnki.42-1908/n.2023.12.009.
[26]
WANG H Y, ZHU W T, XIAO G, et al. Research on brain tumor segmentation by optimizing a TransUnet based algorithm[J]. Chin J Stereol Image Anal, 2023, 28(1): 98-107. DOI: 10.13505/j.1007-1482.2023.28.01.011.
[27]
MUTHUSIVARAJAN R, CELAYA A, YUNG J P, et al. Evaluating the relationship between magnetic resonance image quality metrics and deep learning-based segmentation accuracy of brain tumors[J]. Med Phys, 2024, 51(7): 4898-4906. DOI: 10.1002/mp.17059.
[28]
CAI J P, HE Z, ZHENG Z W, et al. Learning global dependencies based on hierarchical full connection for brain tumor segmentation[J/OL]. Comput Methods Programs Biomed, 2022, 221: 106925 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/35688765/. DOI: 10.1016/j.cmpb.2022.106925.
[29]
NGUYEN-TAT T B, NGUYEN T T, NGUYEN H N, et al. Enhancing brain tumor segmentation in MRI images: a hybrid approach using UNet, attention mechanisms, and transformers[J/OL]. Egypt Inform J, 2024, 27: 100528 [2025-06-19]. https://www.sciencedirect.com/science/article/pii/S1110866524000914. DOI: 10.1016/j.eij.2024.100528.
[30]
ZHANG M Q, LIU D W, SUN Q L, et al. Augmented Transformer network for MRI brain tumor segmentation[J/OL]. J King Saud Univ Comput Inf Sci, 2024, 36(1): 101917 [2025-06-19]. https://www.sciencedirect.com/science/article/pii/S1319157824000065. DOI: 10.1016/j.jksuci.2024.101917.
[31]
SHEDBALKAR J, PRABHUSHETTY K, R H H. UNet and transformer-based model for multi-modality brain tumor segmentation[J]. Int J Electron Commun Eng, 2023, 10(8): 22-35. DOI: 10.14445/23488549/ijece-v10i8p103.
[32]
LI Z R, SILAMU W, FENG S R, et al. Focal cross transformer: multi-view brain tumor segmentation model based on cross window and focal self-attention[J/OL]. Front Neurosci, 2023, 17: 1192867 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/37250393/. DOI: 10.3389/fnins.2023.1192867.
[33]
SHI J J, YU L, CHENG Q M, et al. M2 FTrans: modality-masked fusion transformer for incomplete multi-modality brain tumor segmentation[J/OL]. IEEE J Biomed Health Inform, 2023 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/37862279/. DOI: 10.1109/JBHI.2023.3326151.
[34]
ZAKARIAH M, AL-RAZGAN M, ALFAKIH T. Dual vision Transformer-DSUNET with feature fusion for brain tumor segmentation[J/OL]. Heliyon, 2024, 10(18): e37804 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/39323802/. DOI: 10.1016/j.heliyon.2024.e37804.
[35]
ANAYA-ISAZA A, MERA-JIMÉNEZ L, FERNANDEZ-QUILEZ A. CrossTransUnet: a new computationally inexpensive tumor segmentation model for brain MRI[J/OL]. IEEE Access, 2023, 11: 27066-27085 [2025-06-19]. https://ieeexplore.ieee.org/document/10073551. DOI: 10.1109/access.2023.3257767.
[36]
LIU W J, JIANG L, QU H C, et al. MRI brain tumor image segmentation by fusing CNN and Transformer[J]. CAAI Trans Intell Syst, 2024, 19(4): 1007-1015. DOI: 10.11992/tis.202301016.
[37]
ABOUSSALEH I, RIFFI J, FAZAZY K EL, et al. 3DUV-NetR+: a 3D hybrid semantic architecture using transformers for brain tumor segmentation with MultiModal MR images[J/OL]. Results Eng, 2024, 21: 101892 [2025-06-19]. https://www.sciencedirect.com/science/article/pii/S2590123024001452. DOI: 10.1016/j.rineng.2024.101892.
[38]
TING H, LIU M H. Multimodal transformer of incomplete MRI data for brain tumor segmentation[J]. IEEE J Biomed Health Inform, 2024, 28(1): 89-99. DOI: 10.1109/JBHI.2023.3286689.
[39]
NEHRU V, PRABHU V. Automated multimodal brain tumor segmentation and localization in MRI images using hybrid Res2-UNeXt[J]. J Electr Eng Technol, 2024, 19(5): 3485-3497. DOI: 10.1007/s42835-023-01779-3.
[40]
ZHANG D S, WANG C J, CHEN T H, et al. Scalable Swin Transformer network for brain tumor segmentation from incomplete MRI modalities[J/OL]. Artif Intell Med, 2024, 149: 102788 [2025-06-19]. https://pubmed.ncbi.nlm.nih.gov/38462288/. DOI: 10.1016/j.artmed.2024.102788.
[41]
LIN J W, LIN J T, LU C, et al. CKD-TransBTS: clinical knowledge-driven hybrid transformer with modality-correlated cross-attention for brain tumor segmentation[J]. IEEE Trans Med Imag, 2023, 42(8): 2451-2461. DOI: 10.1109/tmi.2023.3250474.
[42]
WANG C L, ZHENG W B. Missing modality brain tumor segmentation based on graph convolution and transformer[J]. Comput Technol Dev, 2025, 35(7): 173-181. DOI: 10.20165/j.cnki.ISSN1673-629X.2025.0064.
[43]
YANG H Y, ZHOU T, ZHOU Y, et al. Flexible fusion network for multi-modal brain tumor segmentation[J]. IEEE J Biomed Health Inform, 2023, 27(7): 3349-3359. DOI: 10.1109/JBHI.2023.3271808.
[44]
CAO J, LIU J J, CHEN J R. A brain tumor segmentation method based on attention mechanism[J/OL]. Sci Rep, 2025, 15: 15229 [2025-06-19]. https://www.nature.com/articles/s41598-025-98355-8. DOI: 10.1038/s41598-025-98355-8.
[45]
WANG Y, XU J, GUAN Y C, et al. MSegNet: a multi-view coupled cross-modal attention model for enhanced MRI brain tumor segmentation[J/OL]. Int J Comput Intell Syst, 2025, 18(1): 63 [2025-06-19]. https://link.springer.com/article/10.1007/s44196-025-00787-7. DOI: .
[46]
PEIRIS H, HAYAT M, CHEN Z L, et al. A robust volumetric transformer forAccurate 3D tumor segmentation[C]//Medical Image Computing and Computer Assisted Intervention-MICCAI 2022. Cham: Springer, 2022: 162-172. DOI: 10.1007/978-3-031-16443-9_16.
[47]
WANG P Y, YANG Q S, HE Z B, et al. Vision transformers in multi-modal brain tumor MRI segmentation: a review[J/OL]. Meta-Radiology, 2023, 1(1): 100004 [2025-06-19]. https://www.sciencedirect.com/science/article/pii/S2950162823000048. DOI: 10.1016/j.metrad.2023.100004.
[48]
PACAL I. A novel Swin transformer approach utilizing residual multi-layer perceptron for diagnosing brain tumors in MRI images[J]. Int J Mach Learn Cybern, 2024, 15(9): 3579-3597. DOI: 10.1007/s13042-024-02110-w.
[49]
ZHU Z Q, HE X Y, QI G Q, et al. Brain tumor segmentation based on the fusion of deep semantics and edge information in multimodal MRI[J/OL]. Inf Fusion, 2023, 91: 376-387 [2025-06-19]. https://www.sciencedirect.com/science/article/abs/pii/S1566253522001981. DOI: 10.1016/j.inffus.2022.10.022.
[50]
VATANPOUR M, HADDADNIA J. TransDoubleU-Net: dual scale swin transformer with dual level decoder for 3D multimodal brain tumor segmentation[J/OL]. IEEE Access, 2023, 11: 125511-125518 [2025-06-19]. https://www.researchgate.net/publication/375513597_TransDoubleU-Net_Dual_Scale_Swin_Transformer_with_Dual_Level_Decoder_for_3D_Multimodal_Brain_Tumor_Segmentation. DOI: 10.1109/access.2023.3330958.
[51]
MA B, SUN Q, MA Z, et al. DTASUnet: a local and global dual transformer with the attention supervision U-network for brain tumor segmentation[J/OL]. Sci Rep, 2024, 14: 28379 [2025-06-19]. https://www.nature.com/articles/s41598-024-78067-1. DOI: 10.1038/s41598-024-78067-1.
[52]
TAO Y P, BAI S Q, ZHOU Z W. Neural architecture search for multi-tissue segmentation using convolutional and transformer-based networks in glioma segmentation[J]. J Comput Appl, 2025, 45(7): 2378-2386. DOI: 10.11772/j.issn.1001-9081.2024070977.

PREV Research progress in segmentation methods for heterogeneity of the microenvironment in glioma based on multimodal magnetic resonance imaging
NEXT Progress in MRI on brain imaging of patients with cervical spondylotic myelopathy
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn