Generative Adversarial Network–based Noncontrast CT Angiography for Aorta and Carotid Arteries

Published Online:Nov 14 2023https://doi.org/10.1148/radiol.230681

See editorial byRan Zhang

Abstract

Generative adversarial network–based CT angiography (CTA) images of the neck or abdomen obtained without iodinated contrast agents had image quality comparable to that of real CTA images, and reasonable to good diagnostic accuracy.

Background

Iodinated contrast agents (ICAs), which are widely used in CT angiography (CTA), may cause adverse effects in humans, and their use is time-consuming and costly.

Purpose

To develop an ICA-free deep learning imaging model for synthesizing CTA-like images and to assess quantitative and qualitative image quality as well as the diagnostic accuracy of synthetic CTA (Syn-CTA) images.

Materials and Methods

A generative adversarial network (GAN)–based CTA imaging model was trained, validated, and tested on retrospectively collected pairs of noncontrast CT and CTA images of the neck and abdomen from January 2017 to June 2022, and further validated on an external data set. Syn-CTA image quality was evaluated using quantitative metrics. In addition, two senior radiologists scored the visual quality on a three-point scale (3 = good) and determined the vascular diagnosis. The validity of Syn-CTA images was evaluated by comparing the visual quality scores and diagnostic accuracy of aortic and carotid artery disease between Syn-CTA and real CTA scans.

Results

CT scans from 1749 patients (median age, 60 years [IQR, 50–68 years]; 1057 male patients) were included in the internal data set: 1137 for training, 400 for validation, and 212 for testing. The external validation set comprised CT scans from 42 patients (median age, 67 years [IQR, 59–74 years]; 37 male patients). Syn-CTA images had high similarity to real CTA images (normalized mean absolute error, 0.011 and 0.013 for internal and external test set, respectively; peak signal-to-noise ratio, 32.07 dB and 31.58 dB; structural similarity, 0.919 and 0.906). The visual quality of Syn-CTA and real CTA images was comparable (internal test set, P = .35; external validation set, P > .99). Syn-CTA showed reasonable to good diagnostic accuracy for vascular diseases (internal test set: accuracy = 94%, macro F1 score = 91%; external validation set: accuracy = 86%, macro F1 score = 83%).

Conclusion

A GAN-based model that synthesizes neck and abdominal CTA-like images without the use of ICAs shows promise in vascular diagnosis compared with real CTA images.

Clinical trial registration no. NCT05471869

Supplemental material is available for this article.

See also the editorial by Zhang and Turkbey in this issue.

References

1. Hansen NJ. Computed tomographic angiography of the abdominal aorta. Radiol Clin North Am 2016;54(1):35–54. Crossref, Medline, Google Scholar
2. Bae KT. Intravenous contrast medium administration and scan timing at CT: considerations and approaches. Radiology 2010;256(1):32–61. Link, Google Scholar
3. Böhm I, Nairz K, Morelli JN, Keller PS, Heverhagen JT. Iodinated contrast media and the alleged “iodine allergy”: an inexact diagnosis leading to inferior radiologic management and adverse drug reactions. Rofo 2017;189(4):326–332. Crossref, Medline, Google Scholar
4. Beckett KR, Moriarity AK, Langer JM. Safe use of contrast media: what the radiologist needs to know. RadioGraphics 2015;35(6):1738–1750. Link, Google Scholar
5. Lee SY, Lim KW, Chang YS. Radiocontrast media hypersensitivity in the Asia Pacific region. Asia Pac Allergy 2014;4(2):119–125. Crossref, Medline, Google Scholar
6. Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv 1409.1556 [preprint] https://arxiv.org/abs/1409.1556. Posted September 4, 2014. Updated April 10, 2015. Accessed November 20, 2021. Google Scholar
7. Fard AS, Reutens DC, Vegh V. CNNs and GANs in MRI-based cross-modality medical image estimation. arXiv 2106.02198 [preprint] https://arxiv.org/abs/2106.02198. Posted June 4, 2021. Accessed November 30, 2021. Google Scholar
8. Hu L, Zhou DW, Zha YF, et al. Synthesizing high-b-value diffusion-weighted imaging of the prostate using generative adversarial networks. Radiol Artif Intell 2021;3(5):e200237. Link, Google Scholar
9. Jayachandran Preetha C, 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. Lancet Digit Health 2021;3(12):e784–e794. Crossref, Medline, Google Scholar
10. Xu C, Zhang D, Chong J, Chen B, Li S. Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver MR images using pixel-level graph reinforcement learning. Med Image Anal 2021;69:101976. Crossref, Medline, Google Scholar
11. Liu H, Jin X, Liu L, Jin X. Low-dose CT image denoising based on improved DD-Net and local filtered mechanism. Comput Intell Neurosci 2022;2022:2692301. Medline, Google Scholar
12. Han Z, Shangguan H, Zhang X, Zhang P, Cui X, Ren H. A dual-encoder-single-decoder based low-dose CT denoising network. IEEE J Biomed Health Inform 2022;26(7):3251–3260. Crossref, Medline, Google Scholar
13. Li J, Wang YH, Zheng FL, et al. Feasibility of utilizing ultra-low-dose contrast medium for pancreatic artery depiction using the combination of advanced virtual monoenergetic imaging and high-concentration contrast medium: an intra-patient study. Insights Imaging 2021;12(1):166. Crossref, Medline, Google Scholar
14. Li W, Li Y, Qin W, et al. Magnetic resonance image (MRI) synthesis from brain computed tomography (CT) images based on deep learning methods for magnetic resonance (MR)-guided radiotherapy. Quant Imaging Med Surg 2020;10(6):1223–1236. Crossref, Medline, Google Scholar
15. Dalmaz O, Yurt M, Çukur T. ResViT: residual vision transformers for multimodal medical image synthesis. IEEE Trans Med Imaging 2022;41(10):2598–2614. Crossref, Medline, Google Scholar
16. Isola P, Zhu JY, Zhou T, Efros AA. Image-to-image translation with conditional adversarial networks. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, 2017; 5967–5976. Google Scholar
17. Zhu JY, Park T, Isola P, Efros AA. Unpaired image-to-image translation using cycle-consistent adversarial networks. In: 2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy, 2017; 2242–2251. Google Scholar
18. Armanious K, Jiang C, Fischer M, et al. MedGAN: medical image translation using GANs. Comput Med Imaging Graph 2020;79:101684. Crossref, Medline, Google Scholar
19. Wang TC, Liu MY, Zhu JY, Tao A, Kautz J, Catanzaro B. High-resolution image synthesis and semantic manipulation with conditional GANs. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, Utah, 2018; 8798–8807. Google Scholar
20. Hu X. Multi-texture GAN: exploring the multi-scale texture translation for brain MR images. arXiv 2102.07225 [preprint] https://arxiv.org/abs/2102.07225. Posted February 14, 2021. Accessed March 30, 2022. Google Scholar
21. Kong L, Lian C, Huang D, Hu Y, Zhou Q. Breaking the dilemma of medical image-to-image translation. In: Advances in Neural Information Processing Systems 34 (NeurIPS 2021). 2021; 1964–1978. Google Scholar
22. Han R, Jones CK, Lee J, et al. Deformable MR-CT image registration using an unsupervised, dual-channel network for neurosurgical guidance. Med Image Anal 2022;75:102292. Crossref, Medline, Google Scholar
23. Zhou T, Fu H, Chen G, Shen J, Shao L. Hi-net: hybrid-fusion network for multi-modal MR image synthesis. IEEE Trans Med Imaging 2020;39(9):2772–2781. Crossref, Medline, Google Scholar
24. Duan C, Xiong Y, Cheng K, et al. Accelerating susceptibility-weighted imaging with deep learning by complex-valued convolutional neural network (ComplexNet): validation in clinical brain imaging. Eur Radiol 2022;32(8):5679–5687. Crossref, Medline, Google Scholar
25. Fu F, Wei J, Zhang M, et al. Rapid vessel segmentation and reconstruction of head and neck angiograms using 3D convolutional neural network. Nat Commun 2020;11(1):4829. Crossref, Medline, Google Scholar
26. Nishii T, Kobayashi T, Tanaka H, et al. Deep learning-based post hoc CT denoising for myocardial delayed enhancement. Radiology 2022;305(1):82–91. Link, Google Scholar
27. Doi M, Takahashi F, Kawasaki Y. Bayesian noninferiority test for 2 binomial probabilities as the extension of Fisher exact test. Stat Med 2017;36(30):4789–4803. Crossref, Medline, Google Scholar
28. Kleesiek J, Morshuis JN, Isensee F, et al. Can virtual contrast enhancement in brain MRI replace gadolinium?: a feasibility study. Invest Radiol 2019;54(10):653–660. Crossref, Medline, Google Scholar
29. Cui H, Chen Y, Li K, et al. Untargeted metabolomics identifies succinate as a biomarker and therapeutic target in aortic aneurysm and dissection. Eur Heart J 2021;42(42):4373–4385. Crossref, Medline, Google Scholar
30. Hajian-Tilaki K. Sample size estimation in diagnostic test studies of biomedical informatics. J Biomed Inform 2014;48:193–204. Crossref, Medline, Google Scholar
31. Benzakoun J, Deslys MA, Legrand L, et al. Synthetic FLAIR as a substitute for FLAIR sequence in acute ischemic stroke. Radiology 2022;303(1):153–159. Link, Google Scholar
32. Conte GM, Weston AD, Vogelsang DC, et al. Generative adversarial networks to synthesize missing T1 and FLAIR MRI sequences for use in a multisequence brain tumor segmentation model. Radiology 2021;299(2):313–323. Link, Google Scholar

Article History

Received: Mar 17 2023
Revision requested: May 15 2023
Revision received: Sept 28 2023
Accepted: Oct 4 2023
Published online: Nov 14 2023

Vol. 309, No. 2

Supplemental Material

Abbreviations

Abbreviations:
CTA	CT angiography
CTA-GAN	GAN-based CTA imaging
GAN	generative adversarial network
ICA	iodinated contrast agent
NCCT	noncontrast CT
Syn-CTA	synthetic CTA

Metrics

Altmetric Score

PDF download