GS_NeXt: Graph theory combining segment anything model for liver and tumor segmentation from CT

Qing Wang; Jinke Wang; Liang Guo; Min Xu; Qing Wang; Jinke Wang; Liang Guo; Min Xu

doi:10.3934/era.2025204

Electronic Research Archive

2025, Volume 33, Issue 8: 4495-4528. doi: 10.3934/era.2025204

Previous Article Next Article

Research article Special Issues

GS_NeXt: Graph theory combining segment anything model for liver and tumor segmentation from CT

1.
School of Automation, Harbin University of Science and Technology, Harbin 150080, China
2.
Weihai Research Institute, Harbin University of Science and Technology, Weihai 264300, China
3.
Weihai Municipal Hospital, Affiliated to Shandong University, Weihai 264299, China

Received: 23 May 2025 Revised: 01 July 2025 Accepted: 28 July 2025 Published: 06 August 2025

The static convolutional network is designed with a restricted sense field, but it limits the global feature extraction. While dynamic convolution addresses the issue of limited static receptive fields, it struggles to perform well in discontinuous liver regions, liver tumor border regions, and microtumor segmentation. To alleviate the above issues, we proposed a network, GS_NeXt, based on graph theory and the Segment Anything Model (SAM) for liver and tumor segmentation from CT scans. First, we employed the feature extraction module of ConvNeXt-v2 to learn features across channels, enabling the network to focus on critical liver and tumor regions. Second, we utilized the SAM with frozen weights to extract more comprehensive global information, thereby enhancing feature representation. Third, we applied graph reasoning to globally model unstructured local features, improving the network's understanding of CT images in discontinuous liver regions, liver-tumor boundaries, and microtumor areas. Finally, we incorporated a deep supervision mechanism to facilitate the learning of multi-scale features throughout the network. We evaluated the proposed segmentation method for two publicly available abdominal liver tumor CT datasets. On the LiTS17 dataset, GS_NeXt achieved 97.74% and 87.25% on the Dice scores, 1.01 and 2.23 mm on the average symmetric surface distance (ASD), and 3.68% and 22.60% on the volume overlap errors (VOE) for liver and tumor segmentation, respectively. On the 3DIRCADb dataset, it achieved 97.31% and 87.36% on the Dice scores, ASD values of 1.01 and 2.12 mm, and VOE scores of 3.56% and 21.56% for liver and tumor segmentation, respectively.
- segmentation,
- deep learning,
- liver,
- graph inference,
- dynamic convolution
Citation: Qing Wang, Jinke Wang, Liang Guo, Min Xu. GS_NeXt: Graph theory combining segment anything model for liver and tumor segmentation from CT[J]. Electronic Research Archive, 2025, 33(8): 4495-4528. doi: 10.3934/era.2025204

Related Papers:

Abstract

The static convolutional network is designed with a restricted sense field, but it limits the global feature extraction. While dynamic convolution addresses the issue of limited static receptive fields, it struggles to perform well in discontinuous liver regions, liver tumor border regions, and microtumor segmentation. To alleviate the above issues, we proposed a network, GS_NeXt, based on graph theory and the Segment Anything Model (SAM) for liver and tumor segmentation from CT scans. First, we employed the feature extraction module of ConvNeXt-v2 to learn features across channels, enabling the network to focus on critical liver and tumor regions. Second, we utilized the SAM with frozen weights to extract more comprehensive global information, thereby enhancing feature representation. Third, we applied graph reasoning to globally model unstructured local features, improving the network's understanding of CT images in discontinuous liver regions, liver-tumor boundaries, and microtumor areas. Finally, we incorporated a deep supervision mechanism to facilitate the learning of multi-scale features throughout the network. We evaluated the proposed segmentation method for two publicly available abdominal liver tumor CT datasets. On the LiTS17 dataset, GS_NeXt achieved 97.74% and 87.25% on the Dice scores, 1.01 and 2.23 mm on the average symmetric surface distance (ASD), and 3.68% and 22.60% on the volume overlap errors (VOE) for liver and tumor segmentation, respectively. On the 3DIRCADb dataset, it achieved 97.31% and 87.36% on the Dice scores, ASD values of 1.01 and 2.12 mm, and VOE scores of 3.56% and 21.56% for liver and tumor segmentation, respectively.

References

[1]	F. Bray, M. Laversanne, H. Sung, J. Ferlay, R. Siegel, I. Soerjomataram, et al., Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA: Cancer J. Clin., 74 (2024), 229–263. https://doi.org/10.3322/caac.21834 doi: 10.3322/caac.21834
[2]	R. S. Wang, T. Lei, R. X. Cui, B. T. Zhang, H. Y. Meng, A. K. Nandi, Medical image segmentation using deep learning: A survey, IET Image Proc., 16 (2022), 1243–1267. https://doi.org/10.1049/ipr2.12419 doi: 10.1049/ipr2.12419
[3]	J. Long, E. Shelhamer, T. Darrell, Fully convolutional networks for semantic segmentation, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2015), 3431–3440. https://doi.org/10.1109/CVPR.2015.7298965.
[4]	O. Ronneberger, P. Fischer, T. Brox, U-net: convolutional networks for biomedical image segmentation, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2015, Springer, 9351 (2015), 234–241. https://doi.org/10.1109/CVPR.2015.7298965.
[5]	K. Simonyan, A. Zisserman, Very deep convolutional networks for large-scale image recognition, preprint, arXiv: 1409.1556. https://doi.org/10.48550/arXiv.1409.1556.
[6]	V. Rajinikanth, S. Kadry, R. Damaševičius, D. Sankaran, M. A. Mohammed, S. Chander, et al., Skin melanoma segmentation using VGG-UNet with Adam/SGD optimizer: a study, in 2022 Third International Conference on Intelligent Computing Instrumentation and Control Technologies (ICICICT). IEEE, (2022), 982–986. https://doi.org/10.1109/ICICICT54557.2022.9917848.
[7]	K. M. He, X. Y. Zhang, S. Q. Ren, J. Sun, Deep residual learning for image recognition, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2016), 770–778. https://doi.org/10.1109/CVPR.2016.90.
[8]	F. I. Diakogiannis, F. Waldner, P. Caccetta, C. Wu, ResUNet-a: A deep learning framework for semantic segmentation of remotely sensed data, ISPRS J. Photogramm. Remote Sens., 162 (2020), 94–114. https://doi.org/10.1016/j.isprsjprs.2020.01.013 doi: 10.1016/j.isprsjprs.2020.01.013
[9]	Z. Liu, H. Z. Mao, C. Y. Wu, C. Feichtenhofer, T. Darrell, S. N. Xie, A convnet for the 2020s, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2022), 11976–11986. https://doi.org/10.1109/CVPR52688.2022.01167.
[10]	Z. Z. Han, M. W. Jian, G. G. Wang, ConvUNeXt: An efficient convolution neural network for medical image segmentation, Knowl.-Based Syst., 253 (2022), 109512. https://doi.org/10.1016/j.knosys.2022.109512 doi: 10.1016/j.knosys.2022.109512
[11]	M. M. Rahman, M. Munir, R. Marculescu, Emcad: Efficient multi-scale convolutional attention decoding for medical image segmentation, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2024), 11769–11779. https://doi.org/10.1109/CVPR52733.2024.01118.
[12]	C. Li, A. J. Zhou, A. B. Yao, Omni-dimensional dynamic convolution, preprint, arXiv: 2209.07947. https://doi.org/10.48550/arXiv: 2209.07947.
[13]	Z. H. Xing, T. Ye, Y. J. Yang, G. L, L. Zhu, Segmamba: Long-range sequential modeling mamba for 3d medical image segmentation, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2024, Springer, (2024), 578–588. https://doi.org/10.1007/978-3-031-72111-3_54.
[14]	Y. H. Fu, J. F. Liu, J. Shi, TSCA-Net: Transformer based spatial-channel attention segmentation network for medical images, Comput. Biol. Med., 170 (2024), 107938. https://doi.org/10.1016/j.compbiomed.2024.107938 doi: 10.1016/j.compbiomed.2024.107938
[15]	F. Lyu, A. J. Ma, T. C. F. Yip, G. L. H. Wang, P. C. Yuen, Weakly supervised liver tumor segmentation using couinaud segment annotation, IEEE Trans. Med. Imaging, 41 (2021), 1138–1149. https://doi.org/10.1109/TMI.2021.3132905 doi: 10.1109/TMI.2021.3132905
[16]	L. Meng, Y. Y. Tian, S. H. Bu, Liver tumor segmentation based on 3D convolutional neural network with dual scale, J. Appl. Clin. Med. Phys., 21 (2020), 144–157. https://doi.org/10.1002/acm2.12784 doi: 10.1002/acm2.12784
[17]	R. Y. Li, L. Ch. Xu, K. Xie, J. F. Song, X. W. Ma, L. A. Chang, Dht-net: Dynamic hierarchical transformer network for liver and tumor segmentation, IEEE J. Biomed. Health. Inf., 27 (2023), 3443–3454. https://doi.org/10.1109/JBHI.2023.3268218 doi: 10.1109/JBHI.2023.3268218
[18]	Q. Li, H. Song, Z. H. Wei, F. B. Yang, J. F. Fan, D. N. Ai, Densely connected u-net with criss-cross attention for automatic liver tumor segmentation in ct images, IEEE/ACM Trans. Comput. Biol. Bioinf., 20 (2022), 3399–3410. https://doi.org/10.1109/TCBB.2022.3198425 doi: 10.1109/TCBB.2022.3198425
[19]	J. N. Chi, X. Y. Han, C. D. Wu, H. Wang, P. Ji, X-Net: Multi-branch UNet-like network for liver and tumor segmentation from 3D abdominal CT scans, Neurocomputing, 459 (2021), 81–96. https://doi.org/10.1016/j.neucom.2021.06.021 doi: 10.1016/j.neucom.2021.06.021
[20]	A. Gu, T Dao, Mamba: Linear-time sequence modeling with selective state spaces, preprint, arXiv: 2312.00752.https://doi.org/10.48550/arXiv: 2312.00752.
[21]	A. Gu, K. Goel, C. Ré, Efficiently modeling long sequences with structured state spaces, preprint, arXiv: 2111.00396.https://doi.org/10.48550/arXiv: 2111.00396.
[22]	Z. Zhu, Z. Wang, G. Qi, Y. X. Zhao, Y. Liu, Visually Stabilized Mamba U-shaped network with strong inductive bias for 3D brain tumor segmentation, IEEE Trans. Instrum. Meas., 74 (2025), 2518511. https://doi.org/10.1109/TIM.2025.3551581 doi: 10.1109/TIM.2025.3551581
[23]	Y. B. Tang, Y. X. Tang, Y. Y. Zhu, J. Xiao, R. M. Summers, E 2 Net: An edge enhanced network for accurate liver and tumor segmentation on CT scans, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2020, Springer, 12264 (2020), 512–522. https://doi.org/10.1007/978-3-030-59719-1_50.
[24]	H. Seo, C. Huang, M. Bassenne, R. X. Xiao, L. Xing, Modified U-Net (mU-Net) with incorporation of object-dependent high level features for improved liver and liver-tumor segmentation in CT images, IEEE Trans. Med. Imaging, 39 (2019), 1316–1325. https://doi.org/10.1109/TMI.2019.2948320 doi: 10.1109/TMI.2019.2948320
[25]	A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, et al., Attention is all you need, in Advances in Neural Information Processing Systems 30 (NIPS 2017), Curran Associates, Inc., 30 (2017), 1–11. https://doi.org/10.5555/3295222.3295349.
[26]	Y. F. Ni, G. Chen, Z. Feng, H. Cui, D. Metaxas, S. T. Zhang, et al., DA-Tran: Multiphase liver tumor segmentation with a domain-adaptive transformer network, Pattern Recognit., 149 (2024), 110233. https://doi.org/10.1016/j.patcog.2023.110233 doi: 10.1016/j.patcog.2023.110233
[27]	J. T. Hu, S. Y. Chen, Z. Y. Pan, S. Zen, W. M. Yang, Perspective+ unet: Enhancing segmentation with bi-path fusion and efficient non-local attention for superior receptive fields, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2024, Springer, 15009 (2024), 499–509. https://doi.org/10.1007/978-3-031-72114-4_48.
[28]	P. F. Christ, M. E. A. Elshaer, F. Ettlinger, S. Tatavarty, M. Bickel, M. Armbruster, et al., Automatic liver and lesion segmentation in CT using cascaded fully convolutional neural networks and 3D conditional random fields, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2016, Springer, 9901 (2016), 415–423. https://doi.org/10.1007/978-3-319-46723-8_48.
[29]	J. Jiang, Y. J. Peng, Q. F. Hou, J. Wang, MDCF_Net: A Multi-dimensional hybrid network for liver and tumor segmentation from CT, Biocybern. Biomed. Eng., 43 (2023), 494–506. https://doi.org/10.1016/j.bbe.2023.04.004 doi: 10.1016/j.bbe.2023.04.004
[30]	K. N. Wang, S. X. Li, Z. Y. Bu, F. X. Zhao, G. Q. Zhou, S. J. Zhou, SBCNet: Scale and boundary context attention dual-branch network for liver tumor segmentation, IEEE J. Biomed. Health. Inf., 28 (2024), 2854–2865. https://doi.org/10.1109/JBHI.2024.3370864 doi: 10.1109/JBHI.2024.3370864
[31]	Z. Q. Zhu, Z. M. Zhang, G. Q. Qi, Y. Y. Li, Y. Z. Li, L. Mu, A dual-branch network for ultrasound image segmentation, Biomed. Signal Process. Control, 103 (2025), 107368. https://doi.org/10.1016/j.bspc.2024.107368 doi: 10.1016/j.bspc.2024.107368
[32]	Y. Kaya, E. Akat, S. Yıldırım, Fusion‐Brain‐Net: A Novel Deep Fusion Model for Brain Tumor Classification, Brain Behav., 15(2025), e70520. https://doi.org/10.1002/brb3.70520 doi: 10.1002/brb3.70520
[33]	S. H. Di, Y. Q. Zhao, M. Liao, F. Zhang, X. Li, TD-Net: A hybrid end-to-end network for automatic liver tumor segmentation from CT images, IEEE J. Biomed. Health. Inf., 27 (2022), 1163–1172. https://doi.org/10.1109/JBHI.2022.3181974 doi: 10.1109/JBHI.2022.3181974
[34]	A. Kirillov, E. Mintun, N. Ravi, H. Z. Mao, C. Rolland, L. Gustafson, et al., Segment anything, in Proceedings of the IEEE/CVF International Conference on Computer Vision, (2023), 4015–4026. https://doi.org/10.1109/ICCV51070.2023.00371.
[35]	Y. C. Zhang, Z. R. Shen, R. S. Jiao, Segment anything model for medical image segmentation: Current applications and future directions, Comput. Biol. Med., 171 (2024), 108238. https://doi.org/10.1016/j.compbiomed.2024.108238 doi: 10.1016/j.compbiomed.2024.108238
[36]	C. Qin, J. L. Cao, H. Z. Fu, F. S. Khan, R. M. Anwer, Db-sam: Delving into high quality universal medical image segmentation, in Medical Image Computing and Computer Assisted Intervention-MICCAI 2024, Springer, 15012 (2024), 498–508. https://doi.org/10.1007/978-3-031-72390-2_47.
[37]	C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, Z. Wojna, Rethinking the inception architecture for computer vision, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2016), 2818–2826. https://doi.org/10.1109/CVPR.2016.308.
[38]	N. Ibtehaz, M. S. Rahman, MultiResUNet: Rethinking the U-Net architecture for multimodal biomedical image segmentation, Neural Networks, 121 (2020), 74–87. https://doi.org/10.1016/j.neunet.2019.08.025 doi: 10.1016/j.neunet.2019.08.025
[39]	H. P. Kuang, X. Yang, H. J. Li, J. W. Wei, L. H. Zhang, Adaptive multiphase liver tumor segmentation with multiscale supervision, IEEE Signal Process Lett., 31 (2024), 426–430. https://doi.org/10.1109/LSP.2024.3356414 doi: 10.1109/LSP.2024.3356414
[40]	Z. Liu, Q. Y. Teng, Y. Q. Song, W. Hao, Y. Liu, Y. Zhu, et al., HI-Net: Liver vessel segmentation with hierarchical inter-scale multi-scale feature fusion, Biomed. Signal Process. Control, 96 (2024), 106604. https://doi.org/10.1016/j.bspc.2024.106604 doi: 10.1016/j.bspc.2024.106604
[41]	F. Zhan, W. W. Wang, Q. Chen, Y. N. Guo, L. D. He, L. L. Wang, Three-direction fusion for accurate volumetric liver and tumor segmentation, IEEE J. Biomed. Health. Inf., 28 (2023), 2175–2186. https://doi.org/10.1109/JBHI.2023.3344392 doi: 10.1109/JBHI.2023.3344392
[42]	Y. D. Meng, H. R. Zhang, D. X. Gao, Y. T. Zhao, X. Y. Yang, X. S. Qian, et al., BI-GCN: Boundary-aware input-dependent graph convolution network for biomedical image segmentation, preprint, arXiv: 2110.14775.https://doi.org/10.48550/arXiv: 2110.14775.
[43]	P. B. Weerakody, K. W. Wong, G. J. Wang, W. Ela, A review of irregular time series data handling with gated recurrent neural networks, Neurocomputing, 441 (2021), 161–178. https://doi.org/10.1016/j.neucom.2021.02.046 doi: 10.1016/j.neucom.2021.02.046
[44]	S. Woo, S. Debnath, R. H. Hu, X. L. Chen, Z. Liu, I. S. Kweon, Convnext v2: Co-designing and scaling convnets with masked autoencoders, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, IEEE, (2023), 16133–16142. https://doi.org/10.1109/CVPR52729.2023.01548.
[45]	Q. Dou, L. Q. Yu, H. Chen, Y. M. Jin, X. Yang, J. Qin, et al., 3D deeply supervised network for automated segmentation of volumetric medical images, Med. Image Anal. , 41 (2017), 40–54. https://doi.org/10.1016/j.media.2017.05.001 doi: 10.1016/j.media.2017.05.001
[46]	J. N. Chen, Y. Y. Lu, Q. H. Yu, X. D. Luo, E. Adeli, Y. Wang, et al., Transunet: Transformers make strong encoders for medical image segmentation, preprint, arXiv: 2102.04306. https://doi.org/10.48550/arXiv: 2102.04306.
[47]	Q. G. Jin, Z. P. Meng, C. M. Sun, H. Cui, R. Su, RA-UNet: A hybrid deep attention-aware network to extract liver and tumor in CT scans, Front. Bioeng. Biotechnol., 8 (2020), 605132. https://doi.org/10.3389/fbioe.2020.605132 doi: 10.3389/fbioe.2020.605132
[48]	Q. H. Gao, M. Almekkawy, ASU-Net++: A nested U-Net with adaptive feature extractions for liver tumor segmentation, Comput. Biol. Med., 136 (2021), 104688. https://doi.org/10.1016/j.compbiomed.2021.104688 doi: 10.1016/j.compbiomed.2021.104688
[49]	H. Cao, Y. Y. Wang, J. Chen, D. S. Jiang, X. P. Zhang, Q. Tian, et al., Swin-unet: Unet-like pure transformer for medical image segmentation, in European Conference on Computer Vision, Springer, Cham, 13803 (2022), 205–218. https://doi.org/10.1007/978-3-031-25066-8_9.
[50]	T. Lei, R. S. Wang, Y. X. Zhang, Y. Wan, C. Liu, A. K. Nandi, et al., DefED-Net: Deformable encoder-decoder network for liver and liver tumor segmentation, IEEE Trans. Radiat. Plasma Med. Sci., 6 (2021), 68–78. https://doi.org/10.1109/TRPMS.2021.3059780 doi: 10.1109/TRPMS.2021.3059780
[51]	L. F. Jiang, J. J. Ou, R. H. Liu, Y. Y. Zou, T. Xie, H. G. Xiao, et al., Rmau-net: Residual multi-scale attention u-net for liver and tumor segmentation in ct images, Comput. Biol. Med., 158 (2023), 106838. https://doi.org/10.1016/j.compbiomed.2023.106838 doi: 10.1016/j.compbiomed.2023.106838
[52]	H. Qi, W. J. Wang, Y. T. Shi, X. H. Wang, AD-DUNet: A dual-branch encoder approach by combining axial Transformer with cascaded dilated convolutions for liver and hepatic tumor segmentation, Biomed. Signal Process., 95 (2024), 106397. https://doi.org/10.1016/j.bspc.2024.106397 doi: 10.1016/j.bspc.2024.106397
[53]	S. J. Yang, Y. B. Liang, S. Wu, P. Sun, Z. C. Chen, SADSNet: A robust 3D synchronous segmentation network for liver and liver tumors based on spatial attention mechanism and deep supervision, J. X-Ray Sci. Technol., 32 (2024), 707–723. https://doi.org/10.3233/XST-230312 doi: 10.3233/XST-230312
[54]	H. Y. Ma, M. Maimaiti, FUF-TransUNet: A Transformer-Based U-Net with fully utilize of features for liver and liver-tumor segmentation in CT images, in Chinese Conference on Pattern Recognition and Computer Vision (PRCV), Springer Nature, Singapore, (2024), 34–47. https://doi.org/10.1007/978-981-97-8499-8_3.
[55]	W. J. Ren, B. Li, H. Peng, J. Wang, Lgma-net: liver and tumor segmentation methods based on local–global feature mergence and attention mechanisms, Signal, Image Video Process., 19 (2025), 43. https://doi.org/10.1007/s11760-024-03731-y doi: 10.1007/s11760-024-03731-y

Reader Comments

Your name:*

Email:*
© 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)