Reconstruction of Complete Cerebral Arterial Anatomy from Non-Contrast CT Using Deep Learning

doi:10.3877/cma.j.issn.2095-5782.2026.01.009

Abstract

Abstract:

Objective

To evaluate the feasibility and clinical potential of deep learning–based cerebral artery reconstruction from non-contrast CT(NCCT).

Methods

Paired NCCT/CTA data from two stroke centers were retrospectively analyzed. An internal dataset of 140 patients without large vessel occlusion was split into training(n=100), validation(n=20),and test(n=20)sets, with 20 additional cases for external validation. CTA-derived segmentations served as reference.nnU-Net 3D-fullres, 3D cascade U-Net, and 2D U-Net were compared using DSC, ASSD, and HD95.

Results

nnU-Net 3D-fullres achieved the best performance(DSC 0.78±0.04, 0.75±0.04, and 0.73±0.06 across validation, test, and external sets), with consistent trends in ASSD and HD95.

Conclusion

Deep learning enables accurate,contrast-free reconstruction of cerebral arteries from NCCT and may support path planning for mechanical thrombectomy.

Key words: non-contrast ct, cerebral arteries, vessel segmentation, artificial intelligence, deep learning

Yifu Liu, Jiulou Zhang, Linbo Zhao, Yuezhou Cao, Sheng Liu, Zhenyu Jia, Haibin Shi. Reconstruction of Complete Cerebral Arterial Anatomy from Non-Contrast CT Using Deep Learning[J]. Chinese Journal of Interventional Radiology(Electronic Edition), 2026, 14(01): 60-65.

/ / Recommend

Add to citation manager EndNote|Ris|BibTeX

URL: https://zhjrfsxdzzz.cma-cmc.com.cn/EN/10.3877/cma.j.issn.2095-5782.2026.01.009

https://zhjrfsxdzzz.cma-cmc.com.cn/EN/Y2026/V14/I01/60

Figures/Tables 5

References 14

[1]	Powers WJ, Rabinstein AA, Ackerson T, et al. Guidelines for the Early Management of Patients With Acute Ischemic Stroke: 2019 Update to the 2018 Guidelines for the Early Management of Acute Ischemic Stroke: A Guideline for Healthcare Professionals From the American Heart Association/American Stroke Association[J]. Stroke, 2019, 50(12): e344-e418.
[2]	Turc G, Bhogal P, Fischer U, et al. European Stroke Organisation (ESO)-European Society for Minimally Invasive Neurological Therapy (ESMINT) Guidelines on Mechanical Thrombectomy in Acute Ischemic Stroke[J]. J NeuroIntervent Surg, 2023, 15(8): e8.
[3]	Happi Ngankou E, Gory B, Marnat G, et al. Thrombectomy Complications in Large Vessel Occlusions: Incidence, Predictors, and Clinical Impact in the ETIS Registry[J]. Stroke, 2021, 52(12): e764-e768.
[4]	Schulze-Zachau V, Brehm A, Ntoulias N, et al. Incidence and Outcome of Perforations During Medium Vessel Occlusion Compared With Large Vessel Occlusion Thrombectomy[J]. J Neurointerv Surg, 2024, 16(8): 775-780.
[5]	Lyu J, Fu Y, Yang M, et al. Generative Adversarial Network–Based Noncontrast CT Angiography for Aorta and Carotid Arteries[J]. Radiology, 2023, 309(2): e230681-e230681.
[6]	Klimont M, Oronowicz-Jaśkowiak A, Flieger M, et al. Deep Learning for Cerebral Angiography Segmentation From Non-Contrast Computed Tomography[J]. PLoS One, 2020, 15(7): e0237092-e0237092.
[7]	Isensee F, Jaeger PF, Kohl SAA, et al. nnU-Net: A Self-Configuring Method for Deep Learning-Based Biomedical Image Segmentation[J]. Nat Methods, 2021, 18(2): 203-211.
[8]	World Medical Association. World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects[J]. JAMA, 2013, 310(20): 2191-2194.
[9]	Klein S, Staring M, Murphy K, et al. Elastix: A Toolbox for Intensity-Based Medical Image Registration[J]. IEEE Trans Med Imaging, 2010, 29(1): 196-205.
[10]	Çiçek Ö, Abdulkadir A, Lienkamp SS, et al. 3D U-Net: Learning Dense Volumetric Segmentation From Sparse Annotation [C]// Ourselin S, Joskowicz L, Sabuncu M R, et al. Medical Image Computing and Computer-Assisted Intervention-MICCAI 2016. Cham: Springer, 2016: 424-432.
[11]	Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation [C]// Navab N, Hornegger J, Wells W, et al. Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Cham: Springer, 2015: 234-241.
[12]	Chen L, Platzer P, Reschl C, et al. Validation of a Deep-Learning Segmentation Model for Adult and Pediatric Head and Neck Radiotherapy in Different Patient Positions[J]. Phys Imaging Radiat Oncol, 2024, 29: 100527-100527.
[13]	Hall E, Ullberg T, Andsberg G, et al. Incidence of Intracranial Hemorrhagic Complications After Anterior Circulation Endovascular Thrombectomy in Relation to Occlusion Site: A Nationwide Observational Register Study[J]. J Neurointerv Surg, 2024, 16(11): 1088-1093.
[14]	Faizy TD, Winkelmeier L, Mlynash M, et al. Brain Edema Growth After Thrombectomy Is Associated With Comprehensive Collateral Blood Flow[J]. J Neurointerv Surg, 2024, 16(12): 1334-1340.

	训练集（n=100）	内部验证集（n=20）	内部测试集（n=20）	外部验证集（n=20）	P值
年龄[岁，M(Q₁, Q₃)]	56 (40,69)	52 (49,68)	63 (48,74)	70 (58,73)	0.048 6
女性[例(%)]	38 (38.0%)	9 (45.0%)	6 (30.0%)	9 (45.0%)	0.747 3
高血压[例(%)]	53 (53.0%)	10 (50.0%)	9 (45.0%)	12 (60.0%)	0.831 2
糖尿病[例(%)]	20 (20.0%)	5 (25.0%)	3 (15.0%)	4 (20.0%)	0.922 4
血脂异常[例(%)]	17 (17.0%)	5 (25.0%)	3 (15.0%)	5 (25.0%)	0.764 3
房颤[例(%)]	6 (6.0%)	0 (0.0%)	5 (25.0%)	2 (10.0%)	0.021 5

	训练集（n=100）	内部验证集（n=20）	内部测试集（n=20）	外部验证集（n=20）	P值
年龄[岁，M(Q₁, Q₃)]	56 (40,69)	52 (49,68)	63 (48,74)	70 (58,73)	0.048 6
女性[例(%)]	38 (38.0%)	9 (45.0%)	6 (30.0%)	9 (45.0%)	0.747 3
高血压[例(%)]	53 (53.0%)	10 (50.0%)	9 (45.0%)	12 (60.0%)	0.831 2
糖尿病[例(%)]	20 (20.0%)	5 (25.0%)	3 (15.0%)	4 (20.0%)	0.922 4
血脂异常[例(%)]	17 (17.0%)	5 (25.0%)	3 (15.0%)	5 (25.0%)	0.764 3
房颤[例(%)]	6 (6.0%)	0 (0.0%)	5 (25.0%)	2 (10.0%)	0.021 5

模型	数据集	DSC, mean (SD)	IoU, mean (SD)	ASSD, mm, mean (SD)	HD95, mm, mean (SD)
nnU-Net	内部验证集	0.78 (0.04)	0.65 (0.05)	0.64 (0.14)	3.17 (1.01)
	内部测试集	0.75 (0.04)	0.64 (0.06)	0.62 (0.16)	2.80 (1.03)
	外部验证集	0.73 (0.06)	0.62 (0.05)	0.75 (0.20)	3.64 (1.09)
3D级联U-Netu	内部验证集	0.76 (0.05)	0.63 (0.06)	0.70 (0.16)	3.55 (1.12)
	内部测试集	0.72 (0.05)	0.61 (0.06)	0.68 (0.18)	3.15 (1.12)
	外部验证集	0.70 (0.07)	0.59 (0.06)	0.88 (0.24)	4.10 (1.20)
U-Net 2D	内部验证集	0.46 (0.05)	0.30 (0.05)	1.30 (0.19)	6.52 (1.46)
	内部测试集	0.45 (0.05)	0.29 (0.05)	1.32 (0.26)	7.30 (1.45)
	外部验证集	0.42 (0.05)	0.27 (0.04)	1.38 (0.24)	7.38 (1.52)

模型	数据集	DSC, mean (SD)	IoU, mean (SD)	ASSD, mm, mean (SD)	HD95, mm, mean (SD)
nnU-Net	内部验证集	0.78 (0.04)	0.65 (0.05)	0.64 (0.14)	3.17 (1.01)
	内部测试集	0.75 (0.04)	0.64 (0.06)	0.62 (0.16)	2.80 (1.03)
	外部验证集	0.73 (0.06)	0.62 (0.05)	0.75 (0.20)	3.64 (1.09)
3D级联U-Netu	内部验证集	0.76 (0.05)	0.63 (0.06)	0.70 (0.16)	3.55 (1.12)
	内部测试集	0.72 (0.05)	0.61 (0.06)	0.68 (0.18)	3.15 (1.12)
	外部验证集	0.70 (0.07)	0.59 (0.06)	0.88 (0.24)	4.10 (1.20)
U-Net 2D	内部验证集	0.46 (0.05)	0.30 (0.05)	1.30 (0.19)	6.52 (1.46)
	内部测试集	0.45 (0.05)	0.29 (0.05)	1.32 (0.26)	7.30 (1.45)
	外部验证集	0.42 (0.05)	0.27 (0.04)	1.38 (0.24)	7.38 (1.52)

[1]	Zetao Chen, Longshiyu Qiu, Zhuohong Gong, Hengyi Liu, Peisheng Zeng, Mengru Shi. Challenges and strategies for artificial intelligence-based quantitative measurement in oral implant [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 1-8.
[2]	Ting Li, Chao Guo, Chenxi Li, Rouyan Li, Yihan Zhang, Yuhan Zhang, Yan Chen, Gaowa Siqin, Zhongcheng Gong. Deep learning-based automated segmentation and volumetric analysis of impacted maxillary canines [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 9-16.
[3]	Xiaofei Meng, Zhuohong Gong, Yaowen Wan, Peida Li, Qiqi Hu, Longshiyu Qiu, Hengyi Liu, Weili Xie. Preliminary exploration on the application of the classification of dental smile lines via deep learning and multimodal large language model [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 17-24.
[4]	Zhuyu Wang, Jingjing Quan. Artificial intelligence-assisted basic and clinical research in endodontics [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 25-33.
[5]	Siyi Lü, Yanqi Wang, Jun Qiu, Yujiang Chen, Jie Gao. Application of artificial intelligence image processing technology in stomatology [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 40-46.
[6]	Cuijun Li, Gengbin Cai, Xiaoyin Liang, Yongyi Wang, Xinyi Zhan, Peiming Gu. Application of artificial intelligence in dental nursing [J]. Chinese Journal of Stomatological Research(Electronic Edition), 2026, 20(01): 47-50.
[7]	Tinglin Yang, Tao Huang. Progress and challenges in the application of artificial intelligence for evaluating thyroid nodules [J]. Chinese Archives of General Surgery(Electronic Edition), 2026, 20(01): 60-65.
[8]	Haonan Mei, Rui Yang, Xiuheng Liu. Research progress of artificial intelligence-assisted pathological image analysis in the diagnosis of prostate cancer [J]. Chinese Journal of Endourology(Electronic Edition), 2026, 20(01): 1-7.
[9]	Xiaobo Ding, Jie Chen, Yanbo Wang. Progress in the application of artificial intelligence in the diagnosis and treatment of urinary calculi [J]. Chinese Journal of Endourology(Electronic Edition), 2026, 20(01): 15-21.
[10]	Fan Fan, Hao Huang, Lili Fu, Chunmei Zhou, Xuexia Ma, Hai Huang. Construction and effectiveness of intelligent urinary control standardized ward for patients with lower urinary tract dysfunction [J]. Chinese Journal of Endourology(Electronic Edition), 2026, 20(01): 44-50.
[11]	Jiageng Song, Wenhan Yuan, Ying Zheng. The application and prospects of artificial intelligence in minimally invasive gynecological surgery [J]. Chinese Journal of Laparoscopic Surgery(Electronic Edition), 2026, 19(01): 60-64.
[12]	Yufei Deng, Zhixin Wang, Ke Lou, Linxuan Zhang, Guichun Ma, Cuo Gang. Application of radiomics in precise diagnosis, efficacy evaluation, and treatment plan optimization of liver cancer [J]. Chinese Journal of Hepatic Surgery(Electronic Edition), 2026, 15(02): 172-180.
[13]	Chunyuan Yang, Xu Deng, Jingjing Wang, Cairang Yangdan, Wei Pan. Technical progress and clinical application of precise hepatectomy [J]. Chinese Journal of Hepatic Surgery(Electronic Edition), 2026, 15(02): 261-265.
[14]	Yufang Liu, Qing Zhao. MRI in rectal cancer: from precision staging to prognostic prediction [J]. Chinese Journal of Colorectal Diseases(Electronic Edition), 2026, 15(01): 31-36.
[15]	Xiaokun Chen, Shunda Du. Advances and challenges of radiomics in hepatocellular carcinoma [J]. Chinese Journal of Digestion and Medical Imageology(Electronic Edition), 2026, 16(02): 97-100.

Please choose a citation manager

Content to export