Mitigation of motion-induced artifacts in cone beam computed tomography using deep convolutional neural networks.
cone beam computed tomography
deep learning
motion artifacts
Journal
Medical physics
ISSN: 2473-4209
Titre abrégé: Med Phys
Pays: United States
ID NLM: 0425746
Informations de publication
Date de publication:
Oct 2023
Oct 2023
Historique:
revised:
25
02
2023
received:
04
12
2022
accepted:
19
03
2023
medline:
23
10
2023
pubmed:
31
3
2023
entrez:
30
3
2023
Statut:
ppublish
Résumé
Cone beam computed tomography (CBCT) is often employed on radiation therapy treatment devices (linear accelerators) used in image-guided radiation therapy (IGRT). For each treatment session, it is necessary to obtain the image of the day in order to accurately position the patient and to enable adaptive treatment capabilities including auto-segmentation and dose calculation. Reconstructed CBCT images often suffer from artifacts, in particular those induced by patient motion. Deep-learning based approaches promise ways to mitigate such artifacts. We propose a novel deep-learning based approach with the goal to reduce motion induced artifacts in CBCT images and improve image quality. It is based on supervised learning and includes neural network architectures employed as pre- and/or post-processing steps during CBCT reconstruction. Our approach is based on deep convolutional neural networks which complement the standard CBCT reconstruction, which is performed either with the analytical Feldkamp-Davis-Kress (FDK) method, or with an iterative algebraic reconstruction technique (SART-TV). The neural networks, which are based on refined U-net architectures, are trained end-to-end in a supervised learning setup. Labeled training data are obtained by means of a motion simulation, which uses the two extreme phases of 4D CT scans, their deformation vector fields, as well as time-dependent amplitude signals as input. The trained networks are validated against ground truth using quantitative metrics, as well as by using real patient CBCT scans for a qualitative evaluation by clinical experts. The presented novel approach is able to generalize to unseen data and yields significant reductions in motion induced artifacts as well as improvements in image quality compared with existing state-of-the-art CBCT reconstruction algorithms (up to +6.3 dB and +0.19 improvements in peak signal-to-noise ratio, PSNR, and structural similarity index measure, SSIM, respectively), as evidenced by validation with an unseen test dataset, and confirmed by a clinical evaluation on real patient scans (up to 74% preference for motion artifact reduction over standard reconstruction). For the first time, it is demonstrated, also by means of clinical evaluation, that inserting deep neural networks as pre- and post-processing plugins in the existing 3D CBCT reconstruction and trained end-to-end yield significant improvements in image quality and reduction of motion artifacts.
Sections du résumé
BACKGROUND
BACKGROUND
Cone beam computed tomography (CBCT) is often employed on radiation therapy treatment devices (linear accelerators) used in image-guided radiation therapy (IGRT). For each treatment session, it is necessary to obtain the image of the day in order to accurately position the patient and to enable adaptive treatment capabilities including auto-segmentation and dose calculation. Reconstructed CBCT images often suffer from artifacts, in particular those induced by patient motion. Deep-learning based approaches promise ways to mitigate such artifacts.
PURPOSE
OBJECTIVE
We propose a novel deep-learning based approach with the goal to reduce motion induced artifacts in CBCT images and improve image quality. It is based on supervised learning and includes neural network architectures employed as pre- and/or post-processing steps during CBCT reconstruction.
METHODS
METHODS
Our approach is based on deep convolutional neural networks which complement the standard CBCT reconstruction, which is performed either with the analytical Feldkamp-Davis-Kress (FDK) method, or with an iterative algebraic reconstruction technique (SART-TV). The neural networks, which are based on refined U-net architectures, are trained end-to-end in a supervised learning setup. Labeled training data are obtained by means of a motion simulation, which uses the two extreme phases of 4D CT scans, their deformation vector fields, as well as time-dependent amplitude signals as input. The trained networks are validated against ground truth using quantitative metrics, as well as by using real patient CBCT scans for a qualitative evaluation by clinical experts.
RESULTS
RESULTS
The presented novel approach is able to generalize to unseen data and yields significant reductions in motion induced artifacts as well as improvements in image quality compared with existing state-of-the-art CBCT reconstruction algorithms (up to +6.3 dB and +0.19 improvements in peak signal-to-noise ratio, PSNR, and structural similarity index measure, SSIM, respectively), as evidenced by validation with an unseen test dataset, and confirmed by a clinical evaluation on real patient scans (up to 74% preference for motion artifact reduction over standard reconstruction).
CONCLUSIONS
CONCLUSIONS
For the first time, it is demonstrated, also by means of clinical evaluation, that inserting deep neural networks as pre- and post-processing plugins in the existing 3D CBCT reconstruction and trained end-to-end yield significant improvements in image quality and reduction of motion artifacts.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
6228-6242Subventions
Organisme : Innosuisse - Schweizerische Agentur für Innovationsförderung
ID : 35244.1 IP-LS
Informations de copyright
© 2023 The Authors. Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.
Références
Jaffray DA, Siewerdsen JH, Wong JW, Martinez AA. Flat-panel cone-beam computed tomography for image-guided radiation therapy. Int J Radiat Oncol Biol Phys. 2002;53:1337-1349.
Elstrøm UV, Muren LP, Petersen JBB, Grau C. Evaluation of image quality for different kV cone-beam CT acquisition and reconstruction methods in the head and neck region. Acta Oncologica. 2011;50:908-917.
Yoon S, Lin H, Alonso-Basanta M, et al. Initial evaluation of a novel cone-beam CT-based semi-automated online adaptive radiotherapy system for head and neck cancer treatment - a timing and automation quality study. Cureus. 2020;12(8):e9660.
Jarema T, Aland T. Using the iterative kV CBCT reconstruction on the Varian Halcyon linear accelerator for radiation therapy-planning CT datasets: a feasibility study. Int J Radiat Oncol Biol Phys. 2019;68:112-116. Proceedings of the American Society for Radiation Oncology 61st Annual Meeting.
Feldkamp LA, Davis LC, Kress JW. Practical cone-beam algorithm. J Opt Soc Am A. 1984;1:612-619.
Hounsfield GN. Method of and apparatus for examining a body by radiation such as x or gamma radiation. 1975.
Grant K, Raupach R. SAFIRE: sinogram affirmed iterative reconstruction. Siemens Healthcare; 2012.
Thibault JB. Veo: CT model-based iterative reconstruction. GE Healthcare; 2010.
Paysan P, Brehm M, Wang A, Seghers D, Star-Lack J. Iterative image reconstruction in image-guided radiation therapy. 2018. US Patent App. 15/952,996.
Gardner SJ, Mao W, Liu C, et al. Improvements in CBCT image quality using a novel iterative reconstruction algorithm: a clinical evaluation. Adv Radiat Oncol. 2019;4:390-400.
Kim H, Huq MS, Lalonde R, Houser CJ, Beriwal S, Heron DE. Early clinical experience with Varian Halcyon V2 linear accelerator: dual-isocenter IMRT planning and delivery with portal dosimetry for gynecological cancer treatments. J Appl Clin Med Phys. 2019;20(11):111-120.
Mao W, Liu C, Gardner SJ, et al. Evaluation and clinical application of a commercially available iterative reconstruction algorithm for CBCT-based IGRT. Technol Cancer Res Treat. 2019;18.
Washio H, Ohira S, Funama Y, et al. Metal artifact reduction using iterative CBCT reconstruction algorithm for head and neck radiation therapy: a phantom and clinical study. Eur J Radiol. 2020;132:109293.
Schulze RKW, Heil U, Gross D, et al. Artefacts in CBCT: a review. Dentomaxillofac Radiol. 2011;40(5):265-273.
Dillon O, Keall PJ, Shieh CC, O'Brien RT. Evaluating reconstruction algorithms for respiratory motion guided acquisition. Phys Med Biol. 2020;65.
Peeters STH, Vaassen F, Hazelaar C, et al. Visually guided inspiration breath-hold facilitated with nasal high flow therapy in locally advanced lung cancer. Acta Oncol. 2021;60:567-574.
Daly M, McWilliam A, Radhakrishna G, Choudhury A, Eccles CL. Radiotherapy respiratory motion management in hepatobiliary and pancreatic malignancies: a systematic review of patient factors influencing effectiveness of motion reduction with abdominal compression. Acta Oncol. 2022;61:833-841.
Mohd Amin AT, Mokri SS, Ahmad R, Ismail F, Abd Rahni AA. Evaluation methodology for respiratory signal extraction from clinical cone-beam CT (CBCT) using data-driven methods. Int J Integr Eng. 2021;13:1-8.
Boas F, Fleischmann D. CT artifacts: causes and reduction techniques. Imaging Med. 2012;4.
Gjesteby L, De Man B, Jin Y, et al. Metal artifact reduction in CT: where are we after four decades? IEEE Access. 2016;4:5826-5849.
Paysan P, Roggen T, Zhu L, et al. Deep learning methods for image guidance in radiation therapy. In: Schilling FP, Stadelmann T, eds. Artificial Neural Networks in Pattern Recognition - 9th IAPR TC3 Workshop, ANNPR 2020, Winterthur, Switzerland, September 2-4, 2020, Proceedings, Lecture Notes in Computer Science. Springer; 2020;12294:3-22.
Hoffmann M, Christlein V, Breininger K. Deep learning computed tomography: learning projection-domain weights from image domain in limited angle problems. IEEE Trans Med Imaging. 2018;37:1454-1463.
Maier A, Syben C, Stimpel B. Learning with known operators reduces maximum error bounds. Nat Mach Intell. 2019;1:373-380.
Wang J, Liang J, Cheng J, Guo Y, Zeng L. Deep learning based image reconstruction algorithm for limited-angle translational computed tomography. PLoS ONE. 2020;15.
Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells WM, Frangi AF, eds. Medical Image Computing and Computer-Assisted Intervention - MICCAI 2015. Springer International Publishing; 2015;234-241.
Andersen A, Kak A. Simultaneous algebraic reconstruction technique (SART): a superior implementation of the ART algorithm. Ultrason Imaging. 1984;6:81-94.
Schnurr AK, Chung K, Russ T, Schad LR, Zöllner FG. Simulation-based deep artifact correction with convolutional neural networks for limited angle artifacts. Zeitschrift fur Medizinische Physik. 2019;29:150-161.
Schmidhuber J. Deep learning in neural networks: an overview. Neural Netw. 2015;61;85-117.
Stadelmann T, Tolkachev V, Sick B, Stampfli J, Dürr O. Beyond ImageNet: deep learning in industrial practice. In: Applied Data Science. Springer; 2019:205-232.
Amirian M, Montoya-Zegarra JA, Gruss J, et al. PrepNet: a convolutional auto-encoder to homogenize CT scans for cross-dataset medical image analysis. In: 2021 14th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI). IEEE;2021:1-7.
Park HS, Lee SM, Kim HP, Seo JK, Chung YE. CT sinogram-consistency learning for metal-induced beam hardening correction. Med Phys. 2018;45:5376-5384.
Zhang Y, Yu H. Convolutional neural network based metal artifact reduction in X-ray computed tomography. IEEE Trans Med Imaging. 2018;37:1370-1381.
Lin WA, Liao H, Peng C, et al. DuDoNet: dual domain network for CT metal artifact reduction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2019.
Yan K, Wang X, Lu L, et al. Deep lesion graphs in the wild: relationship learning and organization of significant radiology image findings in a diverse large-scale lesion database. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2018:9261-9270.
Lyu Y, Lin WA, Liao H, Lu J, Zhou SK. Encoding metal mask projection for metal artifact reduction in computed tomography. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer; 2020:147-157.
Liao H, Lin WA, Zhou SK, Luo J. ADN: artifact disentanglement network for unsupervised metal artifact reduction. IEEE Trans Med Imaging. 2020;39:634-643.
Lyu Y, Fu J, Peng C, Zhou SK. U-DuDoNet: unpaired dual-domain network for CT metal artifact reduction. In: de Bruijne M, Cattin PC, Cotin S, Padoy N, Speidel S, Zheng Y, Essert C, eds. Medical Image Computing and Computer Assisted Intervention. 2021:296-306.
Wang H, Li Y, Zhang H, et al. InDuDoNet: an interpretable dual domain network for CT metal artifact reduction. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer; 2021:107-118.
Wang T, Lu Z, Yang Z. IDOL-Net: an interactive dual-domain parallel network for CT metal artifact reduction. IEEE Trans Radiat Plasma Med Sci. 2022;6:874-885.
Han Y, Yoo JJ, Ye JC. Deep residual learning for compressed sensing CT reconstruction via persistent homology analysis. CoRR. 2016;abs/1611.06391.
Jin KH, McCann MT, Froustey E, Unser M. Deep convolutional neural network for inverse problems in imaging. IEEE Trans Image Process. 2017;26:4509-4522.
Zhang Z, Liang X, Dong X, Xie Y, Cao G. A sparse-view CT reconstruction method based on combination of DenseNet and deconvolution. IEEE Trans Med Imaging. 2018;37:1407-1417.
Kofler A, Haltmeier M, Kolbitsch C, Kachelrieß M, Dewey M. A U-Nets cascade for sparse view computed tomography. In: Machine Learning for Medical Image Reconstruction: First International Workshop, MLMIR 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, September 16, 2018, Proceedings, Springer; 2018:91-99.
Chen G, Hong X, Ding Q, et al. AirNet: fused analytical and iterative reconstruction with deep neural network regularization for sparse-data CT. Med Phys. 2020;47:2916-2930.
Chen G, Zhao Y, Huang Q, Gao H. 4D-AirNet: a temporally-resolved CBCT slice reconstruction method synergizing analytical and iterative method with deep learning. Phys Med Biol. 2020;65:175020.
Maier J, Sawall S, Kachelrieß M. Deep scatter estimation (DSE): feasibility of using a deep convolutional neural network for real-time x-ray scatter prediction in cone-beam CT. SPIE Medical Imaging. 2018;10573.
Erath J, Vöth T, Maier J, Kachelrieß M. Forward and cross-scatter estimation in dual source CT using the deep scatter estimation (DSE). In: Medical Imaging 2019: Physics of Medical Imaging. vol 10948. International Society for Optics and Photonics, 2019:24.
Paysan P, Strzelecki A, Arrate F. Convolutional network based motion artifact reduction in cone-beam CT. In: AAPM Annual Meeting 2019, e-Poster. 2019.
Su B, Wen Y, Liu Y, Liao S, Fu J, Quan G, Li Z. A deep learning method for eliminating head motion artifacts in computed tomography. Med Phys. 2022;49:411-419.
Lyu Q, Shan H, Xie Y, et al. Cine cardiac MRI motion artifact reduction using a recurrent neural network. IEEE Trans Med Imaging. 2021;40:2170-2181.
Maier J, Lebedev S, Erath J, et al. Deep learning-based coronary artery motion estimation and compensation for short-scan cardiac CT. Med Phys. 2021;48:3559-3571.
Hansch A, Dicken V, Klein J, Morgasb T, Haas B, Hahn H. Artifact-driven sampling schemes for robust female pelvis CBCT segmentation using deep learning. In: Mori K, Hahn H, eds. Medical Imaging 2019: Computer-Aided Diagnosis, vol 10950; 2019.
Zhang Z, Liu J, Yang D, Kamilov US, Hugo GD. Deep learning-based motion compensation for four-dimensional cone-beam computed tomography (4D-CBCT) reconstruction. Med Phys. 2022.
Zhi S, Duan J, Cai J, Mou X. Artifacts reduction method for phase-resolved Cone-Beam CT (CBCT) images via a prior-guided CNN. In: Schmidt TG, Chen GH, Bosmans H, eds. Medical Imaging 2019: Physics of Medical Imaging. vol 10948. International Society for Optics and Photonics, SPIE; 2019:1094828.
Jiang Z, Zhang Z, Chang Y, Ge Y, Yin FF, Ren L. Enhancement of 4-D cone-beam computed tomography (4D-CBCT) using a dual-encoder convolutional neural network (DeCNN). IEEE Trans Radiat Plasma Med Sci. 2022;6;222-230.
Tuy HK. An inversion formula for cone-beam reconstruction. SIAM J Appl Math. 1983;43:546-552.
Buzug TM. Computed Tomography: From Photon Statistics to Modern Cone-Beam CT. Springer; 2008.
Karczmarz S. Angenäherte Auflösung von Systemen linearer Gleichungen. Bull Int Acad Pol Sic Let, Cl Sci Math Nat. 1937;355-357.
Kim D, Ramani S, Fessler JA. Combining ordered subsets and momentum for accelerated X-ray CT image reconstruction. IEEE Trans Med Imaging. 2015;34:167-178.
Nesterov Y. Smooth minimization of non-smooth functions. Math Program. 2005;103;127-152.
Keck B, Hofmann HG, Scherl H, Kowarschik M, Hornegger J. High resolution iterative CT reconstruction using graphics hardware. In: 2009 IEEE Nuclear Science Symposium Conference Record (NSS/MIC). 2009:4035-4040.
Peterlik I, Strzelecki A, Lehmann M, et al. Reducing residual-motion artifacts in iterative 3D CBCT reconstruction in image-guided radiation therapy. Med Phys. 2021;48:6497-6507.
Paysan P, Munro P, Scheib S. CT based simulation framework for motion artifact and ground truth generation of cone-beam CT. In: AAPM Annual Meeting 2019, e-Poster. 2019.
Heinrich MP, Jenkinson M, Brady SM, Schnabel JA. MRF-Based Deformable Registration and Ventilation Estimation of Lung CT. IEEE Trans Med Imaging. 2013;32:1239-1248.
Ulyanov D, Vedaldi A, Lempitsky VS. Instance normalization: the missing ingredient for fast stylization. CoRR. 2016;abs/1607.08022.
Ramachandran P, Zoph B, Le QV. Searching for activation functions. CoRR. 2017;abs/1710.05941.
Woo S, Park J, Lee JY, Kweon IS. CBAM: Convolutional block attention module. In: Proceedings of the European Conference on Computer Vision (ECCV). 2018:3-19.
Zhang Z, Liu Q, Wang Y. Road extraction by deep residual U-Net. CoRR. 2017;abs/1711.10684.
Wang Z, Bovik AC, Sheikh HR, Simoncelli EP. Image quality assessment: from error visibility to structural similarity. IEEE Trans Image Process. 2004;13:600-612.
Paszke A, Gross S, Massa F. PyTorch: an imperative style, high-performance deep learning library. In: Wallach H, Larochelle H, Beygelzimer A, d'Alché-Buc F, Fox E, Garnett R, eds. Advances in Neural Information Processing Systems 32. Curran Associates, Inc; 2019:8024-8035.
Loshchilov I, Hutter F. Decoupled weight decay regularization. In: International Conference on Learning Representations, ICLR 2019, New Orleans, United States, May 6-9, 2019. 2019:1-18.
Sager P, Salzmann S, Burn F, Stadelmann T. Unsupervised domain adaptation for vertebrae detection and identification in 3D CT volumes using a domain sanity loss. J Imaging. 2022;8:222.