Realistic CT data augmentation for accurate deep-learning based segmentation of head and neck tumors in kV images acquired during radiation therapy.
cGAN
radiotherapy
segmentation
Journal
Medical physics
ISSN: 2473-4209
Titre abrégé: Med Phys
Pays: United States
ID NLM: 0425746
Informations de publication
Date de publication:
Jul 2023
Jul 2023
Historique:
revised:
15
03
2023
received:
13
05
2022
accepted:
20
03
2023
medline:
11
7
2023
pubmed:
9
4
2023
entrez:
8
4
2023
Statut:
ppublish
Résumé
Using radiation therapy (RT) to treat head and neck (H&N) cancers requires precise targeting of the tumor to avoid damaging the surrounding healthy organs. Immobilisation masks and planning target volume margins are used to attempt to mitigate patient motion during treatment, however patient motion can still occur. Patient motion during RT can lead to decreased treatment effectiveness and a higher chance of treatment related side effects. Tracking tumor motion would enable motion compensation during RT, leading to more accurate dose delivery. The purpose of this paper is to develop a method to detect and segment the tumor in kV images acquired during RT. Unlike previous tumor segmentation methods for kV images, in this paper, a process for generating realistic and synthetic CT deformations was developed to augment the training data and make the segmentation method robust to patient motion. Detecting the tumor in 2D kV images is a necessary step toward 3D tracking of the tumor position during treatment. In this paper, a conditional generative adversarial network (cGAN) is presented that can detect and segment the gross tumor volume (GTV) in kV images acquired during H&N RT. Retrospective data from 15 H&N cancer patients obtained from the Cancer Imaging Archive were used to train and test patient-specific cGANs. The training data consisted of digitally reconstructed radiographs (DRRs) generated from each patient's planning CT and contoured GTV. Training data was augmented by using synthetically deformed CTs to generate additional DRRs (in total 39 600 DRRs per patient or 25 200 DRRs for nasopharyngeal patients) containing realistic patient motion. The method for deforming the CTs was a novel deformation method based on simulating head rotation and internal tumor motion. The testing dataset consisted of 1080 DRRs for each patient, obtained by deforming the planning CT and GTV at different magnitudes to the training data. The accuracy of the generated segmentations was evaluated by measuring the segmentation centroid error, Dice similarity coefficient (DSC) and mean surface distance (MSD). This paper evaluated the hypothesis that when patient motion occurs, using a cGAN to segment the GTV would create a more accurate segmentation than no-tracking segmentations from the original contoured GTV, the current standard-of-care. This hypothesis was tested using the 1-tailed Mann-Whitney U-test. The magnitude of our cGAN segmentation centroid error was (mean ± standard deviation) 1.1 ± 0.8 mm and the DSC and MSD values were 0.90 ± 0.03 and 1.6 ± 0.5 mm, respectively. Our cGAN segmentation method reduced the segmentation centroid error (p < 0.001), and MSD (p = 0.031) when compared to the no-tracking segmentation, but did not significantly increase the DSC (p = 0.294). The accuracy of our cGAN segmentation method demonstrates the feasibility of this method for H&N cancer patients during RT. Accurate tumor segmentation of H&N tumors would allow for intrafraction monitoring methods to compensate for tumor motion during treatment, ensuring more accurate dose delivery and enabling better H&N cancer patient outcomes.
Sections du résumé
BACKGROUND
BACKGROUND
Using radiation therapy (RT) to treat head and neck (H&N) cancers requires precise targeting of the tumor to avoid damaging the surrounding healthy organs. Immobilisation masks and planning target volume margins are used to attempt to mitigate patient motion during treatment, however patient motion can still occur. Patient motion during RT can lead to decreased treatment effectiveness and a higher chance of treatment related side effects. Tracking tumor motion would enable motion compensation during RT, leading to more accurate dose delivery.
PURPOSE
OBJECTIVE
The purpose of this paper is to develop a method to detect and segment the tumor in kV images acquired during RT. Unlike previous tumor segmentation methods for kV images, in this paper, a process for generating realistic and synthetic CT deformations was developed to augment the training data and make the segmentation method robust to patient motion. Detecting the tumor in 2D kV images is a necessary step toward 3D tracking of the tumor position during treatment.
METHOD
METHODS
In this paper, a conditional generative adversarial network (cGAN) is presented that can detect and segment the gross tumor volume (GTV) in kV images acquired during H&N RT. Retrospective data from 15 H&N cancer patients obtained from the Cancer Imaging Archive were used to train and test patient-specific cGANs. The training data consisted of digitally reconstructed radiographs (DRRs) generated from each patient's planning CT and contoured GTV. Training data was augmented by using synthetically deformed CTs to generate additional DRRs (in total 39 600 DRRs per patient or 25 200 DRRs for nasopharyngeal patients) containing realistic patient motion. The method for deforming the CTs was a novel deformation method based on simulating head rotation and internal tumor motion. The testing dataset consisted of 1080 DRRs for each patient, obtained by deforming the planning CT and GTV at different magnitudes to the training data. The accuracy of the generated segmentations was evaluated by measuring the segmentation centroid error, Dice similarity coefficient (DSC) and mean surface distance (MSD). This paper evaluated the hypothesis that when patient motion occurs, using a cGAN to segment the GTV would create a more accurate segmentation than no-tracking segmentations from the original contoured GTV, the current standard-of-care. This hypothesis was tested using the 1-tailed Mann-Whitney U-test.
RESULTS
RESULTS
The magnitude of our cGAN segmentation centroid error was (mean ± standard deviation) 1.1 ± 0.8 mm and the DSC and MSD values were 0.90 ± 0.03 and 1.6 ± 0.5 mm, respectively. Our cGAN segmentation method reduced the segmentation centroid error (p < 0.001), and MSD (p = 0.031) when compared to the no-tracking segmentation, but did not significantly increase the DSC (p = 0.294).
CONCLUSIONS
CONCLUSIONS
The accuracy of our cGAN segmentation method demonstrates the feasibility of this method for H&N cancer patients during RT. Accurate tumor segmentation of H&N tumors would allow for intrafraction monitoring methods to compensate for tumor motion during treatment, ensuring more accurate dose delivery and enabling better H&N cancer patient outcomes.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
4206-4219Subventions
Organisme : Cancer Australia
ID : APP1180776
Informations de copyright
© 2023 The Authors. Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.
Références
Barton MB, Jacob S, Shafiq J, et al. Estimating the demand for radiotherapy from the evidence: a review of changes from 2003 to 2012. Radiother Oncol. 2014;112(1):140-144. doi:10.1016/j.radonc.2014.03.024
Lee AWM, Ng WT, Chan LLK, et al. Evolution of treatment for nasopharyngeal cancer - success and setback in the intensity-modulated radiotherapy era. Radiother Oncol. 2014;110(3):377-384. doi:10.1016/j.radonc.2014.02.003
Setton J, Caria N, Romanyshyn J, et al. Intensity-modulated radiotherapy in the treatment of oropharyngeal cancer: an update of the memorial sloan-kettering cancer center experience. Int J Radiat Oncol Biol Phys. 2012;82(1):291-298. doi:10.1016/j.ijrobp.2010.10.041
Nutting C, A'Hern R, Rogers MS, et al. First results of a phase III multicenter randomized controlled trial of intensity modulated (IMRT) versus conventional radiotherapy (RT) in head and neck cancer (PARSPORT: ISRCTN48243537; CRUK/03/005). J Clin Oncol. 2009;27(18 suppl):LBA6006-LBA6006. doi:10.1200/jco.2009.27.18_suppl.lba6006
Fang FM, Chien CY, Tsai WL, et al. Quality of life and survival outcome for patients with nasopharyngeal carcinoma receiving three-dimensional conformal radiotherapy vs. intensity-modulated radiotherapy-a longitudinal study. Int J Radiat Oncol Biol Phys. 2008;72(2):356-364. doi:10.1016/j.ijrobp.2007.12.054
Eisbruch A, Kim HM, Terrell JE, Marsh LH, Dawson LA, Ship JA. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. Int J Radiat Oncol Biol Phys. 2001;50(3):695-704. doi:10.1016/S0360-3016(01)01512-7
Leech M, Coffey M, Mast M, et al. ESTRO ACROP guidelines for positioning, immobilisation and position verification of head and neck patients for radiation therapists. Tech Innov Patient Support Radiat Oncol. 2017;1:1-7. doi:10.1016/j.tipsro.2016.12.001
Marsh R, Balter J, Evans VL, Eisbruch A. Design and analysis of an immobilization and repositioning system for treatment of neck malignancies. Med Dosim. 1997;22(4):293-297. doi:10.1016/S0958-3947(97)00102-7
Amelio D, Winter M, Habermehl D, Jäkel O, Debus J, Combs SE. Analysis of inter- and intrafraction accuracy of a commercial thermoplastic mask system used for image-guided particle radiation therapy. J Radiat Res. 2013;54(1):i69-i76. doi:10.1093/jrr/rrt038
Juan-Senabre XJ, López-Tarjuelo J, Conde-Moreno A, et al. Uncertainties and CTV to PTV margins quantitative assessment using cone-beam CT technique in clinical application for prostate, and head and neck irradiation tumours. Clin Transl Oncol. 2011;13(11):819-825. doi:10.1007/s12094-011-0740-8
Bruijnen T, Stemkens B, Terhaard CHJ, Lagendijk JJW, Raaijmakers CPJ, Tijssen RHN. Intrafraction motion quantification and planning target volume margin determination of head-and-neck tumors using cine magnetic resonance imaging. Radiother Oncol. 2019;130:82-88. doi:10.1016/j.radonc.2018.09.015
Neubauer E, Dong L, Followill DS, et al. Assessment of shoulder position variation and its impact on IMRT and VMAT doses for head and neck cancer. Radiat Oncol. 2012;7(1):19. doi:10.1186/1748-717X-7-19
Keall PJ, Ng JA, Juneja P, et al. Real-time 3D image guidance using a standard LINAC: measured motion, accuracy, and precision of the first prospective clinical trial of kilovoltage intrafraction monitoring-guided gating for prostate cancer radiation therapy. Int J Radiat Oncol Biol Phys. 2016;94(5):1015-1021. doi:10.1016/j.ijrobp.2015.10.009
Ng JA, Booth JT, Poulsen PR, et al. Kilovoltage intrafraction monitoring for prostate intensity modulated arc therapy: first clinical results. Int J Radiat Oncol Biol Phys. 2012;84(5):e655-e661. doi:10.1016/J.IJROBP.2012.07.2367
Keall PJ, Aun Ng J, O'Brien R, et al. The first clinical treatment with kilovoltage intrafraction monitoring (KIM): a real-time image guidance method. Med Phys. 2014;42(1):354-358. doi:10.1118/1.4904023
Mueller M, Zolfaghari R, Briggs A, et al. The first prospective implementation of markerless lung target tracking in an experimental quality assurance procedure on a standard linear accelerator. Phys Med Biol. 2020;65(2):025008. doi:10.1088/1361-6560/ab5d8b
Shieh CC, Caillet V, Dunbar M, et al. A Bayesian approach for three-dimensional markerless tumor tracking using kV imaging during lung radiotherapy. Phys Med Biol. 2017;62(8):3065-3080. doi:10.1088/1361-6560/aa6393
Mori S. Real-time image-processing algorithm for markerless tumour tracking using x-ray fluoroscopic imaging. Br J Radiol. 2014;87:20140001. doi:10.1259/bjr.20140001
Lewis JH, Li R, Watkins WT, et al. Markerless lung tumor tracking and trajectory reconstruction using rotational cone-beam projections: a feasibility study. Phys Med Biol. 2010;55(9):2505-2522. doi:10.1088/0031-9155/55/9/006
Hugo GD, Liang J, Yan D. Marker-free lung tumor trajectory estimation from a cone beam CT sinogram. Phys Med Biol. 2010;55(9):2637-2650. doi:10.1088/0031-9155/55/9/014
Fu D, Kahn R, Wang B, et al. Xsight lung tracking system: a fiducial-less method for respiratory motion tracking. Treating Tumors that Move with Respiration. Springer; 2007:265-282. doi:10.1007/978-3-540-69886-9_26
Hirai R, Sakata Y, Tanizawa A, Mori S. Real-time tumor tracking using fluoroscopic imaging with deep neural network analysis. Phys Medica. 2019;59:22-29. doi:10.1016/j.ejmp.2019.02.006
Zhao W, Han B, Yang Y, et al. Incorporating imaging information from deep neural network layers into image guided radiation therapy (IGRT). Radiother Oncol. 2019;140:167-174. doi:10.1016/j.radonc.2019.06.027
Mylonas A, Mueller M, Keall PJ, Booth JT, Nguyen DT. Towards real-time markerless prostate IGRT during VMAT treatment. In: American Association of Physicists in Medicine (AAPM). 2021.
Zhao W, Shen L, Han B, et al. Markerless pancreatic tumor target localization enabled by deep learning. Int J Radiat Oncol Biol Phys. 2019;105(2):432-439. doi:10.1016/j.ijrobp.2019.05.071
Xie Y, Xing L, Gu J, Liu W. Tissue feature-based intra-fractional motion tracking for stereoscopic x-ray image guided radiotherapy. Phys Med Biol. 2013;58(11):3615. doi:10.1088/0031-9155/58/11/3615
Lin T, Cervĩo LI, Tang X, Vasconcelos N, Jiang SB. Fluoroscopic tumor tracking for image-guided lung cancer radiotherapy. Phys Med Biol. 2009;54(4):981. doi:10.1088/0031-9155/54/4/011
Shieh CC, Keall PJ, Kuncic Z, Huang CY, Feain I. Markerless tumor tracking using short kilovoltage imaging arcs for lung image-guided radiotherapy. Phys Med Biol. 2015;60(24):9437. doi:10.1088/0031-9155/60/24/9437
Staub D, Murphy MJ. A digitally reconstructed radiograph algorithm calculated from first principles. Med Phys. 2013;40(1):011902. doi:10.1118/1.4769413
Dhont J, Verellen D, Mollaert I, Vanreusel V, Vandemeulebroucke J. RealDRR - rendering of realistic digitally reconstructed radiographs using locally trained image-to-image translation. Radiother Oncol. 2020;153:213-219. doi:10.1016/j.radonc.2020.10.004
Strid KG. Significance of quantum fluctuations in roentgen imaging. Acta Oncol (Madr). 1980;19(2):129-138. doi:10.3109/02841868009130144
Grossberg AJ, Elhalawani H, Mohamed ASR, et al. HNSCC Dataset. Published online 2020. doi:10.7937/k9/tcia.2020.a8sh-7363
Grossberg AJ, Mohamed ASR, Elhalawani H, et al. Imaging and clinical data archive for head and neck squamous cell carcinoma patients treated with radiotherapy. Sci Data. 2018;5(1):1-10. doi:10.1038/sdata.2018.173
Hesham E, Abdallah SRM, Aubrey LW, et al. Matched computed tomography segmentation and demographic data for oropharyngeal cancer radiomics challenges. Sci Data. 2017;4(1):1-14. doi:10.1038/sdata.2017.77
Clark K, Vendt B, Smith K, et al. The cancer imaging archive (TCIA): maintaining and operating a public information repository. J Digit Imaging. 2013;26:1045-1057. doi:10.1007/s10278-013-9622-7
Vercauteren T, Pennec X, Perchant A, Ayache N. Diffeomorphic demons: efficient non-parametric image registration. Neuroimage. 2009;45(1):S61-S72. doi:10.1016/J.NEUROIMAGE.2008.10.040
Eiben B, Bertholet J, Menten MJ, Nill S, Oelfke U, McClelland JR. Consistent and invertible deformation vector fields for a breathing anthropomorphic phantom: a post-processing framework for the XCAT phantom. Phys Med Biol. 2020;65(16):165005. doi:10.1088/1361-6560/ab8533
Papiez BW, Heinrich MP, Fehrenbach J, Risser L, Schnabel JA. An implicit sliding-motion preserving regularisation via bilateral filtering for deformable image registration. Med Image Anal. 2014;18(8):1299-1311. doi:10.1016/J.MEDIA.2014.05.005
Gurney-Champion OJ, McQuaid D, Dunlop A, et al. MRI-based assessment of 3D intrafractional motion of head and neck cancer for radiation therapy. Int J Radiat Oncol Biol Phys. 2018;100(2):306-316. doi:10.1016/j.ijrobp.2017.10.016
Finnegan R, Chlap P. PlatiPy (Processing Library and Analysis Toolkit for Medical Imaging in Python). GitHub. Accessed June 21, 2021. https://github.com/pyplati/platipy
Staring M, Klein S, Pluim JPW. A rigidity penalty term for nonrigid registration. Med Phys. 2007;34(11):4098-4108. doi:10.1118/1.2776236
Klein S, Staring M, Murphy K, Viergever MA, Pluim JPW. Elastix: a toolbox for intensity-based medical image registration. IEEE Trans Med Imaging. 2010;29(1):196-205. doi:10.1109/TMI.2009.2035616
Mirza M, Osindero S. Conditional Generative Adversarial Nets. arXiv preprint arXiv:1411.1784 (2014) http://arxiv.org/abs/1411.1784
Isola P, Zhu JY, Zhou T, Efros AA. Image-to-Image Translation with Conditional Adversarial Networks. Proceedings of 30th IEEE Conf Comput Vis Pattern Recognition, CVPR 2017. 2016; 5967-5976. Accessed April 19, 2021. http://arxiv.org/abs/1611.07004
Yi X, Walia E, Babyn P. Generative adversarial network in medical imaging: a review. Med Image Anal. 2019;58:101552. doi:10.1016/J.MEDIA.2019.101552
Li C, Wand M. Precomputed real-time texture synthesis with Markovian Generative Adversarial Networks. Lecture Notes in Computer Science. 2016:9907:702-716. doi:10.1007/978-3-319-46487-9_43
Ronneberger O, Fischer P, BroxT. U-Net: convolutional networks for biomedical image segmentation. Lecture Notes in Computer Science. 2015;9351:234-241. Accessed August 18, 2021. https://arxiv.org/abs/1505.04597v1
Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets. In: Advances in Neural Information Processing Systems 27 (NIPS 2014). 2014.
Kessler DA, MacKay JW, Crowe VA, et al. The optimisation of deep neural networks for segmenting multiple knee joint tissues from MRIs. Comput Med Imaging Graph. 2020;86:101793. doi:10.1016/J.COMPMEDIMAG.2020.101793
Rit S, Vila Oliva M, Brousmiche S, Labarbe R, Sarrut D, Sharp GC. The Reconstruction Toolkit (RTK), an open-source cone-beam CT reconstruction toolkit based on the Insight Toolkit (ITK). In: Journal of Physics: Conference Series. Vol 489. Institute of Physics Publishing; 2014:012079. doi:10.1088/1742-6596/489/1/012079
Singhrao K, Kirby N, Pouliot J. A three-dimensional head-and-neck phantom for validation of multimodality deformable image registration for adaptive radiotherapy. Med Phys. 2014;41(12):121709. doi:10.1118/1.4901523
Shamir RR, Duchin Y, Kim J, Sapiro G, Harel N. Continuous dice coefficient: a method for evaluating probabilistic segmentations. arXiv preprint arXiv:1906. 2019:11031.
Chalana V, Kim Y. A methodology for evaluation of boundary detection algorithms on medical images. IEEE Trans Med Imaging. 1997;16(5):642-652. doi:10.1109/42.640755
Santos JDL, Popple R, Agazaryan N, et al. Image guided radiation therapy (IGRT) technologies for radiation therapy localization and delivery. Int J Radiat Oncol Biol Phys. 2013;87(1):33-45. doi:10.1016/J.IJROBP.2013.02.021
Zhao W, Shen L, Han B, et al. Deep learning approach for markerless pancreatic tumor target localization. Int J Radiat Oncol. 2019;105(1):S202-S203. doi:10.1016/j.ijrobp.2019.06.268
Keall PJ, Mageras GS, Balter JM, et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76a). Med Phys. 2006;33(10):3874-3900. doi:10.1118/1.2349696
Al-Mayah A, Moseley J, Hunter S, et al. Biomechanical-based image registration for head and neck radiation treatment. Phys Med Biol. 2010;55(21):6491-6500. doi:10.1088/0031-9155/55/21/010
Pukala J, Meeks SL, Staton RJ, Bova FJ, Mañon RR, Langen KM. A virtual phantom library for the quantification of deformable image registration uncertainties in patients with cancers of the head and neck. Med Phys. 2013;40(11):111703. doi:10.1118/1.4823467
Varadhan R, Karangelis G, Krishnan K, Hui S. A framework for deformable image registration validation in radiotherapy clinical applications. J Appl Clin Med Phys. 2013;14(1):192-213. doi:10.1120/jacmp.v14i1.4066
Barkley HT, Gilbert Fletcher TH, Richard Jesse TH, Robert Lindberg TD. Management of cervical lymph node metastases in squamous ceil carcinoma of the tonsillar fossa, base of tongue, supraglottic larynx, and hypopharynx. Am J Surg. 1972;124(4):462-467.
Lee N, Harris J, Garden AS, et al. Intensity-modulated radiation therapy with or without chemotherapy for nasopharyngeal carcinoma: radiation therapy oncology group phase II trial 0225. J Clin Oncol. 2009;27(22):3684-3690. doi:10.1200/JCO.2008.19.9109
Poulsen PR, Cho B, Langen K, Kupelian P, Keall PJ. Three-dimensional prostate position estimation with a single x-ray imager utilizing the spatial probability density. Phys Med Biol. 2008;53(16):4331-4353. doi:10.1088/0031-9155/53/16/008
Nguyen DT, O'Brien R, Kim JH, et al. The first clinical implementation of a real-time six degree of freedom target tracking system during radiation therapy based on Kilovoltage Intrafraction Monitoring (KIM). Radiother Oncol. 2017;123(1):37-42. doi:10.1016/j.radonc.2017.02.013
Roggen T, Bobic M, Givehchi N, Scheib SG. Deep learning model for markerless tracking in spinal SBRT. Phys Medica. 2020;74:66-73. doi:10.1016/J.EJMP.2020.04.029
Zhang L, LoSasso T, Zhang P, Hunt M, Mageras G, Tang G. Couch and multileaf collimator tracking: a clinical feasibility study for pancreas and liver treatment. Med Phys. 2020;47(10):4743-4757. doi:10.1002/mp.14438
Clover K, Oultram S, Adams C, Cross L, Findlay N, Ponman L. Disruption to radiation therapy sessions due to anxiety among patients receiving radiation therapy to the head and neck area can be predicted using patient self-report measures. Psychooncology. 2011;20(12):1334-1341. doi:10.1002/pon.1854
Molassiotis A, Rogers M. Symptom experience and regaining normality in the first year following a diagnosis of head and neck cancer: a qualitative longitudinal study. Palliat Support Care. 2012;10(3):197-204. doi:10.1017/S147895151200020X