Добавил:
kiopkiopkiop18@yandex.ru t.me/Prokururor I Вовсе не секретарь, но почту проверяю Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Ординатура / Хирургия / Библиотека им академика М.И. Перельмана / Книга_5525_Библиотеки_им_академика_М_И_Перельмана.pdf
Скачиваний:
0
Добавлен:
31.08.2026
Размер:
29 Мб
Скачать
Artificial Intelligence in Adaptive Radiation Therapy
images for dose calculation, image segmentation, registration and PET attenuation correction. While feasibility has been demonstrated in recent studies, addressing challenges remains pivotal prior to clinical implementation.

References

[1] Wang T, Lei Y, Fu Y, Wynne J F, Curran W J, Liu T and Yang X 2020 A review on medical
imaging synthesis using deep learning and its clinical applications J. Appl. Clin. Med. Phys.
22 11–36
[2] Yang X, Lei Y, Shu H-K, Rossi P, Mao H, Shim H, Curran W J and Liu T 2017 Pseudo CT
estimation from MRI using patch-based random forest Proc. SPIE
[3] Lee Y K, Bollet M, Charles-Edwards G, Flower M A, Leach M O, McNair H, Moore E,
Rowbottom C and Webb S 2003 Radiotherapy treatment planning of prostate cancer using magnetic resonance imaging alone Radiother. Oncol.
[4] Jonsson J H, Karlsson M G, Karlsson M and Nyholm T 2010 Treatment planning using
MRI data: an analysis of the dose calculation accuracy for different treatment regions
Radiat. Oncol.
[5] Lambert J et al 2011 MRI-guided prostate radiation therapy planning: Investigation of
dosimetric accuracy of MRI-based dose planning Radiother. Oncol.
[6] Kristensen B H, Laursen F J, Løgager V, Geertsen P F and Krarup-Hansen A 2008
Dosimetric and geometric evaluation of an open low-eld magnetic resonance simulator for radiotherapy treatment planning of brain tumours Radiother. Oncol.
[7] Johansson A, Karlsson M and Nyholm T 2011 CT substitute derived from MRI sequences
with ultrashort echo time Med. Phys.
[8] Hsu S-H, Cao Y, Huang K, Feng M and Balter J M 2013 Investigation of a method for
generating synthetic CT models from MRI scans of the head and neck for radiation therapy
Phys. Med. Biol.
[9] Dowling J A, Lambert J, Parker J, Salvado O, Fripp J, Capp A, Wratten C, Denham J W
and Greer P B 2012 An atlas-based electron density mapping method for magnetic resonance imaging (MRI)-alone treatment planning and adaptive MRI-based prostate radiation therapy Int. J. Radiat. Oncol. Biol. Phys.
[10] Uh J, Merchant T E, Li Y, Li X and Hua C 2014 MRI-based treatment planning with
pseudo CT generated through atlas registration Med. Phys.
[11] Sjölund J, Forsberg D, Andersson M and Knutsson H 2015 Generating patient specic
pseudo-CT of the head from MR using atlas-based regression Phys. Med. Biol.
[12] Keereman V, Fierens Y, Broux T, De Deene Y, Lonneux M and Vandenberghe S 2010
MRI-based attenuation correction for PET/MRI using ultrashort echo time sequences
J. Nucl. Med.
[13] Catana C, van der Kouwe A, Benner T, Michel C J, Hamm M, Fenchel M, Fischl B, Rosen
B, Schmand M and Sorensen A G 2010 Toward implementing an MRI-based PET attenuation-correction method for neurologic studies on the MR-PET brain prototype
J. Nucl. Med.
[14] Kops E R and Herzog H 2007 Alternative methods for attenuation correction for PET
images in MR-PET scanners 2007 IEEE Nuclear Science Symp. Conf. Record (Piscataway, NJ: IEEE) pp 4327–30
5 62
38 2708–14
58 8419
83 e5–e11
51 812–8
51 1431–8
66 203–16
10133 101332Q
98 330–4
87 100–9
41 051711
60 825
8-25
Artificial Intelligence in Adaptive Radiation Therapy
[15] Hofmann M, Steinke F, Scheel V, Charpiat G, Farquhar J, Aschoff P, Brady M, Scholkopf
B and Pichler B J 2008 MRI-based attenuation correction for PET/MRI: a novel approach combining pattern recognition and atlas registration J. Nucl. Med.
[16] Han X 2017 MR-based synthetic CT generation using a deep convolutional neural network
method Med. Phys.
[17] Liu F, Jang H, Kijowski R, Bradshaw T and McMillan A B 2018 Deep learning MR
imaging-based attenuation correction for PET/MR imaging Radiology
[18] Jang H, Liu F, Zhao G, Bradshaw T and McMillan A B 2018 Technical note: deep learning
based MRAC using rapid ultrashort echo time imaging Med. Phys.
[19] Dong X, Wang T, Lei Y, Higgins K, Liu T, Curran W J, Mao H, Nye J A and Yang X 2019
Synthetic CT generation from non-attenuation corrected PET images for whole-body PET imaging Phys. Med. Biol.
[20] Hwang D, Kim K Y, Kang S K, Seo S, Paeng J C, Lee D S and Lee J S 2018 Improving the
accuracy of simultaneously reconstructed activity and attenuation maps using deep learning
J. Nucl. Med.
[21] Fu J, Yang Y, Singhrao K, Ruan D, Chu F I, Low D A and Lewis J H 2019 Deep learning
approaches using 2D and 3D convolutional neural networks for generating male pelvic synthetic computed tomography from magnetic resonance imaging Med. Phys.
[22] Neppl S et al 2019 Evaluation of proton and photon dose distributions recalculated on
2D and 3D Unet-generated pseudoCTs from T1-weighted MR head scans Acta Oncol.
1429–34
[23] Torrado-Carvajal A, Vera-Olmos J, Izquierdo-Garcia D, Catalano O A, Morales M A,
Margolin J, Soricelli A, Salvatore M, Malpica N and Catana C 2019 Dixon-VIBE deep learning (DIVIDE) pseudo-CT synthesis for pelvis PET/MR attenuation correction J. Nucl.
Med.
60 429–35
[24] Leynes A P, Yang J, Wiesinger F, Kaushik S S, Shanbhag D D, Seo Y, Hope T A and
Larson P E Z 2018 Zero-echo-time and Dixon deep pseudo-CT (ZeDD CT): direct generation of pseudo-CT images for pelvic PET/MRI attenuation correction using dep convolutional neural networks with multiparametric MRI J. Nucl. Med.
[25] Chen L, Liang X, Shen C, Jiang S and Wang J 2019 Synthetic CT generation from CBCT
images via deep learning Med. Phys.
[26] Son S J, Park B Y, Byeon K and Park H 2019 Synthesizing diffusion tensor imaging from
functional MRI using fully convolutional networks Comput. Biol. Med.
[27] Spuhler K D, Gardus J 3rd, Gao Y, DeLorenzo C, Parsey R and Huang C 2019 Synthesis of
patient-specic transmission data for PET attenuation correction for PET/MRI neuro­imaging using a convolutional neural network J. Nucl. Med.
[28] Largent A et al 2019 Comparison of deep learning-based and patch-based methods for
pseudo-CT generation in MRI-based prostate dose planning Int. J. Radiat. Oncol. Biol.
Phys.
105 1137–50
[29] Nie D, Trullo R, Lian J, Wang L, Petitjean C, Ruan S, Wang Q and Shen D 2018 Medical
image synthesis with deep convolutional adversarial networks IEEE Trans.Biomed. Eng.
2720–30
[30] Emami H, Dong M, Nejad-Davarani S P and Glide-Hurst C K 2018 Generating synthetic
CTs from magnetic resonance images using generative adversarial networks Med. Phys.
3627–36
44 1408–19
64 215016
59 1624–9
47 1115–25
49 1875–83
286 676–84
45 3697–704
46 3788–98
58
59 852–8
115 103528
60 555–60
65
47
8-26
Artificial Intelligence in Adaptive Radiation Therapy
[31] Isola P, Zhu J-Y, Zhou T and Efros A A 2016 Image-to-image translation with conditional
adversarial networks arXiv:
[32] Liang X, Chen L, Nguyen D, Zhou Z, Gu X, Yang M, Wang J and Jiang S 2019 Generating
synthesized computed tomography (CT) from cone-beam computed tomography (CBCT) using CycleGAN for adaptive radiation therapy Phys. Med. Biol.
[33] Harms J, Lei Y, Wang T, Zhang R, Zhou J, Tang X, Curran W J, Liu T and Yang X 2019
Paired cycle-GAN-based image correction for quantitative cone-beam computed tomogra­phy Med. Phys.
[34] Dong X, Lei Y, Wang T, Higgins K, Liu T, Curran W J, Mao H, Nye J A and Yang X 2019
Deep learning-based attenuation correction in the absence of structural information for whole-body PET imaging Phys. Med. Biol.
[35] Lei Y, Dong X, Wang T, Higgins K, Liu T, Curran W J, Mao H, Nye J A and Yang X 2019
Whole-body PET estimation from low count statistics using cycle-consistent generative adversarial networks Phys. Med. Biol.
[36] Wang T, Lei Y, Tian Z, Dong X, Liu Y, Jiang X, Curran W J, Liu T, Shu H K and Yang X
2019 Deep learning-based image quality improvement for low-dose computed tomography simulation in radiation therapy J. Med. Imaging
[37] Liu Y, Lei Y, Wang T, Fu Y, Tang X, Curran W J, Liu T, Patel P and Yang X 2020 CBCT-
based synthetic CT generation using deep-attention cycleGAN for pancreatic adaptive radiotherapy Med. Phys.
[38] Dong X, Lei Y, Tian S, Wang T, Patel P, Curran W J, Jani A B, Liu T and Yang X 2019
Synthetic MRI-aided multi-organ segmentation on male pelvic CT using cycle consistent deep attention network Radiother. Oncol.
[39] Lei Y, Harms J, Wang T, Liu Y, Shu H K, Jani A B, Curran W J, Mao H, Liu T and Yang
X 2019 MRI-only based synthetic CT generation using dense cycle consistent generative adversarial networks Med. Phys.
[40] Liu Y, Lei Y, Wang T, Kayode O, Tian S, Liu T, Patel P, Curran W J, Ren L and Yang X
2019 MRI-based treatment planning for liver stereotactic body radiotherapy: validation of a deep learning-based synthetic CT generation method Br. J. Radiol.
[41] Liu Y et al 2019 Evaluation of a deep learning-based pelvic synthetic CT generation
technique for MRI-based prostate proton treatment planning Phys. Med. Biol.
205022
[42] Liu Y et al 2019 MRI-based treatment planning for proton radiotherapy: dosimetric
validation of a deep learning-based liver synthetic CT generation method Phys. Med. Biol.
64 145015
[43] Kim K H, Do W J and Park S H 2018 Improving resolution of MR images with an
adversarial network incorporating images with different contrast Med. Phys.
[44] Olberg S et al 2019 Synthetic CT reconstruction using a deep spatial pyramid convolutional
framework for MR-only breast radiotherapy Med. Phys.
[45] Yang Q, Yan P, Zhang Y, Yu H, Shi Y, Mou X, Kalra M K, Zhang Y, Sun L and Wang G
2018 Low-dose CT image denoising using a generative adversarial network with wasserstein distance and perceptual loss IEEE Trans. Med. Imaging
[46] Ouyang J, Chen K T, Gong E, Pauly J and Zaharchuk G 2019 Ultra-low-dose PET
reconstruction using generative adversarial network with feature matching and task-specic perceptual loss Med. Phys.
46 3998–4009
1611.07004
64 125002
65 055011
64 215017
6 043504
47 2472–83
141 192–9
46 3565–81
92 20190067
64
45 3120–31
46 4135–47
37 1348–57
46 3555–64
8-27
Artificial Intelligence in Adaptive Radiation Therapy
[47] Müller-Franzes G et al 2023 A multimodal comparison of latent denoising diffusion
probabilistic models and generative adversarial networks for medical image synthesis Sci.
Rep.
13 12098
[48] Lyu Q and Wang G 2022 Conversion between CT and MRI images using diffusion and
score-matching models arXiv:
[49] Pan S et al 2023 Synthetic CT generation from MRI using 3D transformer-based denoising
diffusion model arXiv:
[50] Fu L, Li X, Cai X, Miao D, Yao Y and Shen Y 2023 Energy-guided diffusion model for
CBCT-to-CT synthesis arXiv:
[51] Peng J et al 2023 CBCT-based synthetic CT image generation using conditional denoising
diffusion probabilistic model arXiv:
[52] Kida S, Nakamoto T, Nakano M, Nawa K, Haga A, Kotoku J, Yamashita H and
Nakagawa K 2018 Cone beam computed tomography image quality improvement using a deep convolutional neural network Cureus
[53] Xie S, Yang C, Zhang Z and Li H 2018 Scatter artifacts removal using learning-based
method for CBCT in IGRT system IEEE Access
[54] Nomura Y, Xu Q, Shirato H, Shimizu S and Xing L 2019 Projection-domain scatter
correction for cone beam computed tomography using a residual convolutional neural network Med. Phys.
[55] Yuan N, Dyer B, Rao S, Chen Q, Benedict S, Shang L, Kang Y, Qi J and Rong Y 2020
Convolutional neural network enhancement of fast-scan low-dose cone-beam CT images for head and neck radiotherapy Phys. Med. Biol.
[56] Kida S, Kaji S, Nawa K, Imae T, Nakamoto T, Ozaki S, Ohta T, Nozawa Y and Nakagawa K
2020 Visual enhancement of cone-beam CT by use of CycleGAN Med. Phys.
[57] Kurz C, Maspero M, Savenije M H F, Landry G, Kamp F, Pinto M, Li M, Parodi K, Belka C
and van den Berg C A T 2019 CBCT correction using a cycle-consistent generative adversarial network and unpaired training to enable photon and proton dose calculation Phys. Med. Biol.
64 225004
[58] Hansen D C, Landry G, Kamp F, Li M, Belka C, Parodi K and Kurz C 2018 ScatterNet: a
convolutional neural network for cone-beam CT intensity correction Med. Phys.
[59] Landry G, Hansen D, Kamp F, Li M, Hoyle B, Weller J, Parodi K, Belka C and Kurz C
2019 Comparing Unet training with three different datasets to correct CBCT images for prostate radiotherapy dose calculations Phys. Med. Biol.
[60] Li Y, Zhu J, Liu Z, Teng J, Xie Q, Zhang L, Liu X, Shi J and Chen L 2019 A preliminary
study of using a deep convolution neural network to generate synthesized CT images based on CBCT for adaptive radiotherapy of nasopharyngeal carcinoma Phys. Med. Biol.
145010
[61] Adrian T, Paolo Z, Arturs M, Gabriel G M, Joao S, Roel J H M S, Johannes A L, Stefan B,
Maria Francesca S and Antje-Christin K 2020 Comparison of CBCT based synthetic CT methods suitable for proton dose calculations in adaptive proton therapy Phys. Med. Biol.
095002
[62] Dai X, Lei Y, Wynne J, Janopaul-Naylor J, Wang T, Roper J, Curran W J, Liu T, Patel P
and Yang X 2021 Synthetic CT-aided multiorgan segmentation for CBCT-guided adaptive pancreatic radiotherapy Med. Phys.
[63] Khoo V S and Joon D L 2006 New developments in MRI for target volume delineation in
radiotherapy Br. J. Radiol.
46 3142–55
2209.12104
2305.19467
2308.03354
2303.02649
10 e2548
6 78031–7
65 035003
47 998–1010
45 4916–26
64 035011
64
65
48 7063–73
79 S2–15
8-28
Artificial Intelligence in Adaptive Radiation Therapy
[64] Nyholm T, Nyberg M, Karlsson M G and Karlsson M 2009 Systematisation of spatial
uncertainties for comparison between a MR and a CT-based radiotherapy workow for prostate treatments Radiat. Oncol.
[65] Ulin K, Urie M M and Cherlow J M 2010 Results of a multi-institutional benchmark test for
cranial CT/MR image registration Int. J. Radiat. Oncol. Biol. Phys.
[66] van der Heide U A, Houweling A C, Groenendaal G, Beets-Tan R G and Lambin P 2012
Functional MRI for radiotherapy dose painting Magn. Reson. Imaging [67] Devic S 2012 MRI simulation for radiotherapy treatment planning Med. Phys. 39 6701–11 [68] Lagendijk J J W, Raaymakers B W, Raaijmakers A J E, Overweg J, Brown K J, Kerkhof E M,
van der Put R W, Hårdemark B, van Vulpen M and van der Heide U A 2008 MRI/linac
integration Radiother. Oncol. [69] Fallone B G, Murray B, Rathee S, Stanescu T, Steciw S, Vidakovic S, Blosser E and
Tymochuk D 2009 First MR images obtained during megavoltage photon irradiation from
a prototype integrated linac-MR system Med. Phys. [70] Kinahan P E, Townsend D W, Beyer T and Sashin D 1998 Attenuation correction for a
combined 3D PET/CT scanner Med. Phys. [71] Burger C, Goerres G, Schoenes S, Buck A, Lonn A H and Von Schulthess G K 2002 PET
attenuation coefcients from CT images: experimental evaluation of the transformation of
CT into PET 511-keV attenuation coefcients Eur. J. Nucl. Med. Mol. Imaging [72] Dinkla A M et al 2019 Dosimetric evaluation of synthetic CT for head and neck
radiotherapy generated by a patch-based three-dimensional convolutional neural network
Med. Phys.
[73] Chen S, Qin A, Zhou D and Yan D 2018 Technical note: U-Net-generated synthetic CT
images for magnetic resonance imaging-only prostate intensity-modulated radiation therapy
treatment planning Med. Phys. [74] Gupta D, Kim M, Vineberg K A and Balter J M 2019 Generation of synthetic CT images
from MRI for treatmentplanning and patient positioning using a 3-channel U-Net trained on
sagittal images Front. Oncol. [75] Dinkla A M, Wolterink J M, Maspero M, Savenije M H F, Verhoeff J J C, Seravalli E,
Išgum I, Seevinck P R and van den Berg C A T 2018 MR-only brain radiation therapy:
dosimetric evaluation of synthetic CTs generated by a dilated convolutional neural network
Int. J. Radiat. Oncol. Biol. Phys.
[76] Arabi H, Dowling J A, Burgos N, Han X, Greer P B, Koutsouvelis N and Zaidi H 2018
Comparative study of algorithms for synthetic CT generation from MRI: consequences for
MRI-guided radiation planning in the pelvic region Med. Phys. [77] Freitag M T et al 2017 Improved clinical workow for simultaneous whole-body PET/MRI
using high-resolution CAIPIRINHA-accelerated MR-based attenuation correction Eur.
Radiol.
[78] Izquierdo-Garcia D, Hansen A E, Förster S, Benoit D, Schachoff S, Fürst S, Chen K T,
Chonde D B and Catana C 2014 An SPM8-based approach for attenuation correction
combining segmentation and nonrigid template formation: application to simultaneous PET/
MR brain imaging J. Nucl. Med. [79] Blanc-Durand P, Khalife M, Sgard B, Kaushik S, Soret M, Tiss A, El Fakhri G, Habert
M-O, Wiesinger F and Kas A 2019 Attenuation correction using 3D deep convolutional
neural network for brain 18F-FDG PET/MR: comparison with Atlas, ZTE and CT based
attenuation correction PLoS One
46 4095–104
96 12–20
4 54
77 1584–9
30 1216–23
86 25–9
36 2084–8
25 2046–53
29 922–7
45 5659–65
9 964
102 801–12
45 5218–33
55 1825–30
14 e0223141
8-29
Artificial Intelligence in Adaptive Radiation Therapy
[80] Ladefoged C N, Marner L, Hindsholm A, Law I, Hojgaard L and Andersen F L 2018 Deep
learning based attenuation correction of PET/MRI in pediatric brain tumor patients:
evaluation in a clinical setting Front. Neurosci. [81] Qi M et al 2020 Multi-sequence MR image-based synthetic CT generation using a generative
adversarial network for head and neck MRI-only radiotherapy Med. Phys. [82] Florkow M C et al 2020 Deep learning-based MR-to-CT synthesis: The inuence of varying
gradient echo-based MR images as input channels Magn. Reson. Med. [83] Tie X, Lam S K, Zhang Y, Lee K H, Au K H and Cai J 2020 Pseudo-CT generation from
multi-parametric MRI using a novel multi-channel multi-path conditional generative
adversarial network for nasopharyngeal carcinoma patients Med. Phys. [84] Gong K, Yang J, Kim K, El Fakhri G, Seo Y and Li Q 2018 Attenuation correction for
brain PET imaging using deep neural network based on Dixon and ZTE MR images Phys.
Med. Biol.
[85] Kazemifar S, McGuire S, Timmerman R, Wardak Z, Nguyen D, Park Y, Jiang S and
Owrangi A 2019 MRI-only brain radiotherapy: assessing the dosimetric accuracy of
synthetic CT images generated using a deep learning approach Radiother. Oncol. [86] Lei Y, Harms J, Wang T, Liu Y, Shu H K, Jani A B, Curran W J, Mao H, Liu T and Yang
X 2019 MRI-only based synthetic CT generation using dense cycle consistent generative
adversarial networks Med. Phys. [87] Xiang L, Wang Q, Nie D, Zhang L, Jin X, Qiao Y and Shen D 2018 Deep embedding
convolutional neural network for synthesizing CT image from T1-weighted MR image Med.
Image Anal.
[88] Maspero M, Savenije M H F, Dinkla A M, Seevinck P R, Intven M P W, Jurgenliemk-
Schulz I M, Kerkmeijer L G W and van den Berg C A T 2018 Dose evaluation of fast
synthetic-CT generation using a generative adversarial network for general pelvis MR-only
radiotherapy Phys. Med. Biol. [89] Liu F, Yadav P, Baschnagel A M and McMillan A B 2019 MR-based treatment planning in
radiation therapy using a deep learning approach J. Appl. Clin. Med. Phys. [90] Shafai-Erfani G et al 2019 MRI-based proton treatment planning for base of skull tumors
Int. J. Part. Ther.
[91] Wang Y, Liu C, Zhang X and Deng W 2019 Synthetic CT generation based on T2 weighted
MRI of nasopharyngeal carcinoma (NPC) using a deep convolutional neural network
(DCNN) Front. Oncol. [92] Koike Y, Akino Y, Sumida I, Shiomi H, Mizuno H, Yagi M, Isohashi F, Seo Y, Suzuki O
and Ogawa K 2020 Feasibility of synthetic computed tomography generated with an
adversarial network for multi-sequence magnetic resonance-based brain radiotherapy
J. Radiat. Res.
[93] Brou Boni K N D, Klein J, Vanquin L, Wagner A, Lacornerie T, Pasquier D and Reynaert
N 2020 MR to CT synthesis with multicenter data in the pelvic area using a conditional
generative adversarial network Phys. Med. Biol. [94] Arabi H, Zeng G, Zheng G and Zaidi H 2019 Novel adversarial semantic structure deep
learning for MRI-guided attenuation correction in brain PET/MRI Eur. J. Nucl. Med. Mol.
Imaging
[95] Wang T, Manohar N, Lei Y, Dhabaan A, Shu H-K, Liu T, Curran W J and Yang X 2019
MRI-based treatment planning for brain stereotactic radiosurgery: dosimetric validation of a
learning-based pseudo-CT generation method Med. Dosim.
63 125011
46 3565–81
47 31–44
63 185001
6 12–25
9 1333
61 92–103
46 2746–59
12 1005
47 1880–94
83 1429–41
47 1750–62
136 56–63
20 105–14
65 075002
44 199–204
8-30
Artificial Intelligence in Adaptive Radiation Therapy
[96] Li B, Lee H C, Duan X, Shen C, Zhou L, Jia X and Yang M 2017 Comprehensive analysis
of proton range uncertainties related to stopping-power-ratio estimation using dual-energy
CT imaging Phys. Med. Biol. [97] McKenzie E M, Santhanam A, Ruan D, OConnor D, Cao M and Sheng K 2019
Multimodality image registration in the head-and-neck using a deep learning-derived
synthetic CT as a bridge Med. Phys. [98] Schilling K G et al 2019 Synthesized b0 for diffusion distortion correction (Synb0-DisCo)
Magn. Reson. Imaging
62 7056–74
47 1094–104
64 62–70
8-31
IOP Publishing
Artificial Intelligence in Adaptive Radiation Therapy
Yi Wang and X. Sharon Qi
Chapter 9
Artificial intelligence-based image registration
and segmentation
Brian M Anderson and Kristy K Brock
This chapter delves into the integration of articial intelligence (AI) into adaptive radiation therapy (ART), emphasizing AI-based image registration and segmenta­tion. It begins by introducing ARTs three categoriesofine, online, and real-time and the challenges they pose, such as accurately adapting treatment plans to patient-specic changes. The chapter then explains fundamental AI concepts, including machine learning and deep learning, highlighting components such as convolutional neural networks and activation functions. It explores how these AI techniques enhance image registration and segmentation, which are crucial for precise targeting and dose calculation in ART. By addressing both the potential benets and challenges of AI in this context, the chapter provides insights into how AI can optimize ART workows and improve personalized patient care.

9.1 AI-based image registration and segmentation for ART

9.1.1 Adaptive radiation therapy
Historically in radiation therapy, a single radiation treatment plan is generated for an individual patient, which will be used for the duration of care [1]. Adaptive radiation therapy (ART) is a broad umbrella which encompasses any changes from this single radiation treatment plan regiment, based on changes present in the patient. These changes in the patient can be inter-fractional (reduction/increase in tumor burden, inammation, changes in bowel location/lling, bladder lling, etc) or intra-fractional (respiratory motion, cardiac motion, peristalsis, skeletal motion).
ART can be categorized into three groups: ofine ART, online ART, and real­time ART. All these approaches have the potential to dose escalate and/or reduce normal tissue dose [24]. However, they also raise a number of challenges and difculties which might require accurate segmentation of target and tissues at risk (segmentation), recalculation of dose and summation of previously delivered
doi:10.1088/978-0-7503-6119-4ch9 9-1 ª IOP Publishing Ltd 2025. All rights,
including for text and data mining (TDM), artificial intelligence (AI) training, and similar technologies, are reserved.
Artificial Intelligence in Adaptive Radiation Therapy
treatments (via deformable registration) re-planning based on patient geometry, and quality assurance ensuring safe adaptation of a plan.
9.1.1.1 Offline ART
Ofine ART is commonly performed as part of routine clinical practice in many radiation oncology clinics. This involves the evaluation of imaging (daily or weekly) acquired throughout a patients treatment for changes in normal tissue and tumor volumes. Patients with bulky disease that is expected to diminish throughout treatment can often be scheduled proactively for mid-treatment simulation scans, e.g. evaluation at fraction 15, or on an ad hoc basis driven by assessments at time of treatment: such as increased difculty with patient setup. There are several examples of success in dose escalation/normal tissue sparing seen by multiple clinics using ofine ART [58].
9.1.1.2 Online ART
Online ART is focused on adapting a treatment plan based on inter-fractional changes while the patient is on the table/in the treatment position. There are several successful examples with active online ART programs for multiple treatment sites [917]. This has the added difculty of a compressed time frame for many of the steps listed above.
When adapting and recalculating a new treatment plan, a major consideration is the ability to accurately calculate dose on the image acquired in-room during treatment. The Varian Ethos (Varian Medical Systems, Palo Alto, CA) system focuses on CBCT­guided adaptive radiation therapy [18]. The Elekta Unity (Elekta Stockholm, Sweden)
1.5 T MR linac system and the ViewRay MRIdian 0.35 T MR linac system focuses on using an integrated MR linac to visualize changes occurring throughout the treatment process and can be used for both online [19] and real-time ART [20].
9.1.1.3 Real-time ART
Real-time ART aims to account for intra-fractional changes occurring during the patient treatment and enable automatic adjustments to optimize delivery/treat the target. This is not limited to the denition of a complete re-planning for the patient, but can include respiratory gating [21], MLC target tracking via KV monitoring [22] and implanted ducials [23] (CyberKnife system), MR imaging [24] (with both Elekta Unity and ViewRay MRIdian), and external surface tracking.
9.1.1.4 Challenges
As stated previously, ART requires a concert of moving parts to either acquire a new planning image or adapt the daily image for dose accumulation, segmentation of targets and normal tissues for dose evaluation, accumulation of previously delivered radiation on the new image, perform re-planning, and quality assurance. To expedite these time restricted steps, AI is often leveraged to register, deform, and propagate not only previously dened contours, but also accumulate previously delivered dose. In the following sections we will discuss AI in registration (rigid and deformable), image segmentation, and discuss how these solutions can help alleviate some of these challenges, while creating some new ones as well.
9-2
Artificial Intelligence in Adaptive Radiation Therapy

9.2 Artificial intelligence

What denes articial intelligence (AI)? While the phrase AI can sometimes feel like a creation of the twenty-rst century, it is something which has been a part of the user experience with computers for quite some time. AI can be dened as anything which enables computers to mimic human behavior. This behavior can be something as simple as identifying a spam email, or as complex as diagnosing a patients disease. A more extensive explanation of both machine learning and deep learning is beyond the immediate scope of the information presented later in this chapter. For a mathemat­ical explanation of deep learning, particularly as it applies to deep learning in Python, we highly recommend the books Deep Learning by Ian Goodfellow et al [25] and Deep Learning with Python [26].
9.2.1 What is machine learning?
Within the broad scope of AI there exists a smaller distinction known as machine learning. Machine learning can be dened as an AI model which is generated from a loss-minimization or separation. A model lossfunction can be imaged as a measure of how incorrect the model is; hence, we often express the desire to minimize loss. For example, imagine a model is created to separate Honeycrisp apples (a reddish­yellow sweet apple) from Granny Smith apples (a green, sour apple). The user could create a variety of features (size, taste, color) that they feel can be quantied between the two apples. A machine learning model would then try to use these features to separate the apples into two groups. A very basic idea of this is shown on the left in gure 9.1, where the features are used directly to make an output prediction.
Figure 9.1. (Left) Basic level of a machine learning model: combining input features to an output. (Right) Basic level of a shallowlearning model: combining features with some number of hidden layers into an output.
9-3