Добавил:
Sekretar
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
X
- •Foreword
- •Acknowledgments
- •Editor biographies
- •Yi Wang
- •X. Sharon Qi
- •List of contributors
- •1.2.3 Feature engineering and representation
- •1.2.4 Linear separability
- •1.2.5 Classical models
- •1.1 A brief introduction to AI
- •1.2 Machine learning basics
- •1.2.1 Learning paradigms
- •1.3 Artificial neural networks
- •1.3.1 Feed-forward neural networks
- •1.3.2 Recurrent neural networks
- •1.3.3 Convolutional neural networks
- •1.3.4 Attention
- •1.3.5 Training neural networks
- •1.3.6 Applications and use cases of deep learning
- •1.4 Model training and evaluation
- •1.4.1 Hyperparameters
- •1.4.2 Data split
- •1.4.3 Evaluation metrics
- •1.5 Generative models
- •1.5.1 Generative adversarial networks
- •1.5.2 Diffusion models
- •1.5.3 Applications and use cases
- •1.6 Ethical consideration and bias
- •1.6.1 Transparency and explainability
- •1.6.2 Bias and fairness
- •1.6.3 Data privacy violation
- •1.6.4 Risk and misuse
- •1.7 Summary
- •References
- •2.1 Introduction
- •2.1.2 Staff roles in radiation therapy
- •2.2 Overview of AI in radiation therapy
- •2.2.1 Patient evaluation and dose prescription
- •2.2.2 Treatment simulation
- •2.2.3 Contouring
- •2.2.4 Treatment planning
- •2.2.5 Quality assurance
- •2.2.6 Treatment delivery
- •2.2.7 Response assessment and toxicity management
- •2.3 Summary
- •3.1 Introduction
- •3.1.1 Introduction of clinical decision making and AI
- •3.1.2 The role of AI in clinical decision making
- •3.2 AI algorithms for clinical decision making
- •3.2.2 Radiomics
- •3.2.3 Data integration by AI
- •3.2.4 Interpretability of AI models
- •3.3 Application of AI in clinical decision making
- •3.3.1 Diagnosis and disease phenotyping
- •3.3.2 Personalized treatment
- •3.3.3 Treatment outcome and prognosis prediction
- •3.4 Challenges and future directions of AI in clinical decision making
- •3.4.1 Challenges and concerns
- •3.4.2 Future directions
- •3.5 Summary
- •References
- •4.1 Introduction
- •4.2 Imaging for treatment planning
- •4.2.1 CT simulation
- •4.2.2 4D-CT
- •4.2.3 PET/CT
- •4.2.4 MRI
- •4.3 Imaging for treatment guidance
- •4.3.1 Portal imaging
- •4.3.2 CBCT
- •4.3.3 CT-on-rail and CT-linac
- •4.3.4 MR-linac
- •4.3.5 PET-linac
- •4.4 Imaging for motion management
- •4.4.1 ExacTrac
- •4.4.2 Varian triggered imaging
- •4.4.3 4D-CBCT
- •4.4.4 Cine MRI
- •4.4.5 4D-MRI
- •4.4.6 Surface imaging
- •4.5 Imaging for treatment assessment
- •4.5.1 Contrasted CT
- •4.5.2 PET/CT
- •4.5.3 Functional MRI
- •4.6 Summary
- •5.1 Introduction to big data in radiation oncology
- •5.1.1 Overview of big data
- •5.1.2 Sources of big data in radiation oncology
- •5.1.3 Big data and AI in radiation oncology
- •5.2 Big data lifecycle in radiation oncology
- •5.2.1 Data aggregation and storage
- •5.2.2 Data sharing and security
- •5.4 The application of big data in radiation oncology
- •5.4.1 Medical image segmentation
- •5.4.2 Automatic treatment planning
- •5.4.3 Treatment response prediction
- •5.4.4 Quality assurance and patient safety
- •5.4.5 Clinical decision support
- •5.2.3 Data visualization
- •5.2.4 Knowledge creation and implementation
- •5.2.5 Data archiving and deletion
- •5.3 Big data analytics with AI
- •5.3.1 Data processing and integration
- •5.3.2 AI modeling
- •5.5 Challenges and future perspectives
- •5.6 Summary
- •Reference
- •6.1 The road to ART
- •6.1.1 3D conformal radiotherapy (3DCRT)
- •6.1.2 Intensity modulated radiotherapy (IMRT)
- •6.1.3 Image-guided radiotherapy (IGRT)
- •6.1.4 Adaptive radiotherapy (ART)
- •6.2 ART workflow and implementation
- •6.2.2 Current practice
- •6.2.3 Clinical impact
- •6.3 Considerations for implementing online ART
- •6.3.1 Time as a limiting factor
- •6.3.2 Implications for fast and reliable re-planning
- •6.3.4 Clinical considerations
- •6.4 Summary
- •7.1 Components of ART workflow
- •7.1.1 Simulation
- •7.1.2 Pre-planning
- •7.1.3 Online imaging and daily re-planning
- •7.1.4 Quality assurance
- •7.2 AI-driven ART
- •7.2.1 Simulation
- •7.2.2 Pre-planning
- •7.2.3 AI for delivery
- •7.3 Outlook and future directions
- •7.3.1 Real-time ART with AI
- •7.3.2 Dose escalation and functional adaption with AI
- •7.4 Summary
- •References
- •8.1 Introduction
- •8.2 Synthetic CT: deep learning methods
- •8.2.1 Conventional methods
- •8.2.2 U-Net
- •8.2.3 Generative adversarial networks
- •8.2.4 Denoising diffusion probabilistic model
- •8.3 Synthetic CT from CBCT
- •8.3.1 Noise and artifact reduction
- •8.3.2 Online dose calculation
- •8.3.3 Online image segmentation
- •8.4 Synthetic CT from MRI
- •8.4.1 Synthetic image accuracy
- •8.4.2 Dose calculation in MR-only radiation therapy
- •8.4.3 PET attenuation correction
- •8.4.4 Image registration
- •8.5 Discussion and outlook
- •8.6 Summary
- •References
- •9.1 AI-based image registration and segmentation for ART
- •9.1.1 Adaptive radiation therapy
- •9.2 Artificial intelligence
- •9.2.1 What is machine learning?
- •9.2.2 What is deep learning?
- •9.3 Deep learning: the basic components
- •9.3.1 Convolutional neural networks: looking at the picture
- •9.3.2 Pooling layers: keeping what matters most
- •9.3.3 Fully connected (dense) layers: bringing it all together
- •9.3.4 Activations
- •9.3.5 Loss: driving the model
- •9.3.6 Auto-encoders: remove the noise
- •9.3.7 Supervised versus unsupervised learning
- •9.3.8 Pre-trained convolutional neural networks
- •9.4 Image registration: bringing two images together
- •9.4.1 Registration similarity metrics
- •9.4.2 Types of registrations
- •9.5 AI-based image registration
- •9.5.1 Supervised learning
- •9.5.2 Unsupervised learning
- •9.5.3 Registration in ART
- •9.5.4 Commonalities in architectures
- •9.6 Image segmentation
- •9.6.1 Introduction: coloring by the numbers
- •9.6.2 Segmentation networks
- •9.6.3 Best practices
- •9.7 Summary
- •10.1 Introduction
- •10.1.1 Overview of chapter content
- •10.2 The landscape of AI-assisted dose prediction
- •10.2.1 Traditional machine learning for dose prediction
- •10.2.2 Deep learning-based dose prediction
- •10.2.3 Challenges in AI-assisted dose prediction
- •10.3 Re-planning workflows powered by AI
- •10.3.1 Deep learning for re-planning pipelines
- •10.4 Future directions of AI-assisted dose prediction and re-planning
- •10.5 Summary
- •11.1 Introduction
- •11.2.1 Imaging-based motion monitoring
- •11.2.2 Delivery system actions
- •11.2.3 Challenges for real-time ART implementation
- •11.3 AI in real-time ART workflows
- •11.3.1 Improving intrafraction motion monitoring through AI
- •11.3.2 Mitigating system latency through AI
- •11.4 AI for ART delivery: future directions
- •11.4.1 Management of non-respiratory motion
- •11.4.2 Training AI models with small or unpaired datasets
- •11.4.4 Biology-guided ART delivery
- •11.5 Summary
- •References
- •12.1 Introduction
- •12.2 Patient QA
- •12.2.1 Pre-planning QA
- •12.2.2 Pre-treatment plan QA
- •12.2.3 On-treatment QA
- •12.3 Treatment delivery systems and instruments
- •12.3.1 Machine commissioning
- •12.3.2 Machine QA
- •12.3.3 Dosimetry tool QA
- •12.4 Summary
- •References
- •13.1 Data resources for response modeling in radiotherapy
- •13.1.1 Clinical data
- •13.1.2 Imaging (radiomics)
- •13.1.3 Treatment planning (dosiomics)
- •13.1.4 Multiomics
- •13.2 Radiotherapy treatment outcome modeling
- •13.2.1 TCP/NTCP in radiotherapy
- •13.2.2 Clinical outcomes versus PROs
- •13.2.3 Machine learning response prediction
- •13.2.4 Explainability of ML response models
- •13.2.5 Sample use cases
- •13.3 AI response-based adaptive radiotherapy
- •13.3.1 Requirements and challenges
- •13.3.2 Prediction versus treatment optimization
- •13.3.3 Sample use cases
- •13.4 Challenges and recommendations
- •13.5 Summary
- •Acknowledgments
- •References
- •14.1 Overview of challenges in AI-driven ART
- •14.2 Data challenges
- •14.2.1 Data availability
- •14.2.2 Data quality
- •14.2.3 Data privacy
- •14.3 Technical challenges
- •14.3.2 Model robustness and generalizability
- •14.3.3 Model explainability and interpretability
- •14.4 Challenges associated with online and real-time workflows
- •14.4.1 Image quality
- •14.4.2 Dose calculation
- •14.4.3 Real-time ART
- •14.5 Operational challenges
- •14.5.1 Clinical validation
- •14.5.3 Staff training
- •14.5.4 User experiences
- •14.5.5 Quality management program
- •14.5.6 Financial challenges
- •14.6 Ethical, regulatory, and legal challenges
- •14.6.1 Ethical issues
- •14.6.2 Regulatory and legal issues
- •14.7 Summary
- •References
- •15.1 Clinical considerations for CT-based offline ART
- •15.1.1 Patient and site selection
- •15.1.2 Re-simulation
- •15.1.3 Re-planning
- •15.1.4 Plan summation and evaluation
- •15.1.6 Limitations and future directions
- •15.2 Clinical considerations for CBCT/CT-based online ART
- •15.2.2 Patient and site selection
- •15.2.3 Simulation
- •15.2.4 Pre-planning review
- •15.2.5 Reference planning
- •15.2.9 Limitations and future directions
- •15.3 Summary
- •References
- •16.1 Introduction
- •16.2 Overview of MRI-guided ART systems
- •16.3 MRI-guided ART workflow
- •16.4 AI applications for MRI-guided ART
- •16.4.1 Synthetic CT generation
- •References
- •16.4.2 Auto-segmentation
- •16.4.3 Image registration
- •16.4.4 Others
- •16.4.5 Future AI development and implementation
- •16.5 Summary
- •17.1 Functional PET-guided ART
- •17.1.1 PET-based functional imaging overview
- •17.1.2 From anatomy to function: the power of PET in radiation therapy
- •17.1.5 Conclusions and future prospects
- •17.2 Functional MRI-guided ART
- •17.2.1 From anatomy to function: the power of functional MRI in radiation therapy
- •17.2.4 Conclusion and future prospects
- •17.3 Summary
- •References
- •18.1 Proton ART
- •18.1.1 Clinical context and necessity
- •18.1.2 Patient populations
- •18.1.4 Rationale for AI in proton ART
- •18.2 AI in proton ART
- •18.2.1 Imaging
- •18.2.2 Deformable and rigid registration
- •18.2.3 Contour propagation
- •18.2.4 Dose calculations
- •18.2.5 Plan optimization
- •18.2.6 Other developments
- •18.3 Implementation of adaptive proton therapy
- •18.4 Summary
- •References
- •19.1 Designing clinical trials with AI
- •19.1.1 The essential role of clinical trials
- •19.1.2 Trial protocols and methodologies
- •19.1.3 AI-driven clinical trial design and execution
- •19.1.4 Incorporation of digital twins (DTs) in clinical trials
- •19.2 Implementation of AI in ongoing clinical trials
- •19.2.1 Integration with existing clinical trial frameworks
- •19.2.2 Quality assurance, compliance, and standardization
- •19.3 Case studies of AI in adaptive radiotherapy trials
- •19.3.1 Overview of guidance for advanced radiotherapy in clinical trials
- •19.3.2 AI in the radiotherapy clinical trial quality assurance processes
- •19.4 Ethical and regulatory considerations
- •19.4.1 Patient consent and data privacy
- •19.4.2 Bias, fairness, and transparency
- •19.4.3 Regulatory guidelines and compliance
- •19.5 Future directions and challenges
- •19.5.1 Emerging technologies and techniques
- •19.5.2 Alternative strategies
- •19.6 Conclusion
- •19.7 Summary
- •References
- •20.1 Risk management
- •20.1.1 Prospective risk assessments
- •20.1.2 Root cause analysis

Artificial Intelligence in Adaptive Radiation Therapy
[114] Therasse P et al 2000 New guidelines to evaluate the response to treatment in solid tumors
JNCI J. Natl Cancer Inst.
[115] Campbell A, Davis L M, Wilkinson S K and Hesketh R L 2019 Emerging functional
imaging biomarkers of tumour responses to radiotherapy Cancers
[116] Miles K A 1999 Tumour angiogenesis and its relation to contrast enhancement on
computed tomography: a review Eur. J. Radiol.
[117] Marcus C D, Ladam-Marcus V, Cucu C, Bouché O, Lucas L and Hoeffel C 2009 Imaging
techniques to evaluate the response to treatment in oncology: current standards and
perspectives Crit. Rev. Oncol. Hematol.
[118] Garcia-Barros M et al 2003 Tumor response to radiotherapy regulated by endothelial cell
apoptosis Science
[119] Bellomi M, Petralia G, Sonzogni A, Zampino M G and Rocca A 2007 CT perfusion for the
monitoring of neoadjuvant chemotherapy and radiation therapy in rectal carcinoma: initial
experience Radiology
[120] Sahani D V et al 2005 Assessing tumor perfusion and treatment response in rectal cancer
with multisection CT: initial observations Radiology
[121] Ursino S et al 2016 Role of perfusion CT in the evaluation of functional primary tumour
response after radiochemotherapy in head and neck cancer: preliminary findings Br. J. Radiol.
89 20151070
[122] Coolens C, Driscoll B, Foltz W D, Jaffray D A and Chung C 2015 Early detection of tumor
response using volumetric DCE-CT and DCE-MRI in metastatic brain patients treated
with radiosurgery Int. J. Radiat. Oncol.
[123] Šurlan-Popovič K, Bisdas S, Rumboldt Z, Koh T S and Strojan P 2010 Changes in
perfusion CT of advanced squamous cell carcinoma of the head and neck treated during the
course of concomitant chemoradiotherapy Am. J. Neuroradiol.
[124] Kino A et al 2017 Perfusion CT measurements predict tumor response in rectal carcinoma
Abdom. Radiol.
[125] Patchett N, Furlan A and Marsh J W 2016 Decrease in tumor enhancement on contrast-
enhanced CT is associated with improved survival in patients with hepatocellular carcinoma treated with Sorafenib Jpn. J. Clin. Oncol.
[126] Bussink J, Van Herpen C M, Kaanders J H and Oyen W J 2010 PET-CT for response
assessment and treatment adaptation in head and neck cancer Lancet Oncol.
[127] Juweid M E et al 2007 Use of positron emission tomography for response assessment of
lymphoma: consensus of the imaging subcommittee of International Harmonization Project
in lymphoma J. Clin. Oncol.
[128] Cheson B D et al 2007 Revised response criteria for malignant lymphoma J. Clin. Oncol. 25
579–86
[129] Lin C et al 2007 Early18F-FDG PET for prediction of prognosis in patients with diffuse
large B-cell lymphoma: SUV-based assessment versus visual analysis J. Nucl. Med.
1626–32
[130] Fueger B J et al 2005 Performance of 2-deoxy-2-[F-18]fluoro-d-glucose positron emission
tomography and integrated PET/CT in restaged breast cancer patients Mol. Imaging Biol.
369–76
[131] Radan L, Ben-Haim S, Bar-Shalom R, Guralnik L and Israel O 2006 The role of FDG-
PET/CT in suspected recurrence of breast cancer Cancer
300 1155–9
42 1132–40
92 205–16
11 131
30 198–205
72 217–38
244 486–93
234 785–92
93 S7
31 570–5
46 839–44
11 661–9
25 571–8
48
107 2545–51
7
4-22

Artificial Intelligence in Adaptive Radiation Therapy
[132] Kostakoglu L and Goldsmith S J 200318F-FDG PET evaluation of the response to therapy
for lymphoma and for breast, lung, and colorectal carcinoma J. Nucl. Med. 44 224–39
[133] Eschmann S M et al 2007
18
F-FDG PET for assessment of therapy response and
preoperative re-evaluation after neoadjuvant radio-chemotherapy in stage III non-small
cell lung cancer Eur. J. Nucl. Med. Mol. Imaging
34 463–71
[134] Weber W A et al 2003 Positron emission tomography in non-small-cell lung cancer:
prediction of response to chemotherapy by quantitative assessment of glucose use J. Clin.
Oncol.
21 2651–7
[135] Hoekstra C J et al 2005 Prognostic relevance of response evaluation using [18F]−2-fluoro-2-
deoxy-D-glucose positron emission tomography in patients with locally advanced non-small-cell
lung cancer J. Clin. Oncol.
23 8362–70
[136] Wieder H A et al 2004 Time course of tumor metabolic activity during chemoradiotherapy
of esophageal squamous cell carcinoma and response to treatment J. Clin. Oncol.
22 900–8
[137] Wieder H and Weber W 2009 Prediction of tumour response by FDG-PET in patients with
adenocarcinomas of the oesophagogastric junction Eur. J. Nucl. Med. Mol. Imaging
36
158–9
[138] Lordick F et al 2007 PET to assess early metabolic response and to guide treatment of
adenocarcinoma of the oesophagogastric junction: the MUNICON phase II trial Lancet.
Oncol.
8 797–805
[139] De Geus-Oei L F et al 2008 Chemotherapy response evaluation with FDG–PET in patients
with colorectal cancer Ann. Oncol.
19 348–52
[140] Cascini G L et al 200618F-FDG PET is an early predictor of pathologic tumor response to
preoperative radiochemotherapy in locally advanced rectal cancer J. Nucl. Med. 47 1241–8
[141] Avril N et al 2005 Prediction of response to neoadjuvant chemotherapy by sequential F-18-
fluorodeoxyglucose positron emission tomography in patients with advanced-stage ovarian
cancer J. Clin. Oncol.
23 7445–53
[142] Brun E et al 2002 FDG PET studies during treatment: prediction of therapy outcome in
head and neck squamous cell carcinoma Head Neck
24 127–35
[143] McCollum A D et al 2004 Positron emission tomography with18F-fluorodeoxyglucose to
predict pathologic response after induction chemotherapy and definitive chemoradiotherapy in head and neck cancer Head Neck
26 890–6
[144] Benz M R et al 2008 Treatment monitoring by18F-FDG PET/CT in patients with
sarcomas: interobserver variability of quantitative parameters in treatment-induced changes
in histopathologically responding and nonresponding tumors J. Nucl. Med.
49 1038–46
[145] Steinert H C, Dellea M M S, Burger C and Stahel R 2005 Therapy response evaluation in
malignant pleural mesothelioma with integrated PET–CT imaging Lung Cancer
[146] Sheikhbahaei S, Mena E, Marcus C, Wray R, Taghipour M and Subramaniam R M 2016
18
F-FDG PET/CT: therapy response assessment interpretation (Hopkins criteria) and
survival outcomes in lung cancer patients J. Nucl. Med.
57 855–60
49 S33–5
[147] Kremer R et al 2016 FDG PET/CT for assessing the resectability of NSCLC patients with
N2 disease after neoadjuvant therapy Ann. Nucl. Med.
30 114–21
[148] De Leyn P et al 2006 Prospective comparative study of integrated positron emission
tomography-computed tomography scan compared with remediastinoscopy in the assessment of residual mediastinal lymph node disease after induction chemotherapy for
mediastinoscopy-proven stage IIIA-N2 non-small-cell lung cancer: a Leuven Lung
Cancer Group Study J. Clin. Oncol.
24 3333–9
4-23

Artificial Intelligence in Adaptive Radiation Therapy
[149] Marchetti L et al 2021 Diagnostic contribution of contrast-enhanced CT as compared
with unenhanced low-dose CT in PET/CT staging and treatment response assessment of
18
F-FDG–avid lymphomas: a prospective study J. Nucl. Med. 62 1372–9
[150] Yassin A, El Sheikh R H and Ali M M 2020 PET/CT vs CECT in assessment of therapeutic
response in lymphoma Egypt J. Radiol. Nucl. Med.
[151] Tawakol A, Abdelhafez Y G, Osama A, Hamada E and El Refaei S 2016 Diagnostic
performance of
18
F-FDG PET/contrast-enhanced CT versus contrast-enhanced CT alone
for post-treatment detection of ovarian malignancy Nucl. Med. Commun.
[152] Bollineni V R, Kramer G M, Jansma E P, Liu Y and Oyen W J G 2016 A systematic review on
18
[
F]FLT-PET uptake as a measure of treatment response in cancer patients Eur. J. Cancer 55
51 238
37 453–60
81–97
[153] Bhatnagar P, Subesinghe M, Patel C, Prestwich R and Scarsbrook A F 2013 Functional
imaging for radiation treatment planning, response assessment, and adaptive therapy in
head and neck cancer RadioGraphics
33 1909–29
[154] Dhermain F G, Hau P, Lanfermann H, Jacobs A H and Van Den Bent M J 2010 Advanced
MRI and PET imaging for assessment of treatment response in patients with gliomas
Lancet Neurol.
9 906–20
[155] Yankeelov T E et al 2007 Integration of quantitative DCE-MRI and ADC mapping to
monitor treatment response in human breast cancer: initial results Magn. Reson. Imaging
25
1–13
[156] Haider M A et al 2008 Dynamic contrast-enhanced magnetic resonance imaging for
localization of recurrent prostate cancer after external beam radiotherapy Int. J. Radiat.
Oncol.
70 425–30
[157] King A D et al 2015 DCE-MRI for pre-treatment prediction and post-treatment assessment
of treatment response in sites of squamous cell carcinoma in the head and neck PLoS One
10 e0144770
[158] Dijkhoff R A P, Beets-Tan R G H, Lambregts D M J, Beets G L and Maas M 2017 Value
of DCE-MRI for staging and response evaluation in rectal cancer: a systematic review Eur.
J. Radiol.
95 155–68
[159] Nelson S J 2011 Assessment of therapeutic response and treatment planning for brain
tumors using metabolic and physiological MRI NMR Biomed.
24 734–49
[160] Chandarana H, Wang H, Tijssen R H N and Das I J 2018 Emerging role of MRI in
radiation therapy J. Magn. Reson. Imaging
48 1468–78
[161] Bains L J, Zweifel M and Thoeny H C 2012 Therapy response with diffusion MRI: an
update Cancer Imaging
12 395–402
[162] Hamstra D A, Rehemtulla A and Ross B D 2007 Diffusion magnetic resonance imaging: a
biomarker for treatment response in oncology J. Clin. Oncol.
25 4104–9
[163] Padhani A R and Koh D-M 2011 Diffusion MR imaging for monitoring of treatment
response Magn. Reson. Imaging Clin. N. Am.
[164] Yoshino E et al 1996 Irradiation effects on the metabolism of metastatic brain tumors:
analysis by positron emission tomography and
Stereotact. Funct. Neurosurg.
66 240–59
19 181–209
1
H-magnetic resonance spectroscopy
[165] Zeng Q-S, Li C-F, Zhang K, Liu H, Kang X-S and Zhen J-H 2007 Multivoxel 3D proton
MR spectroscopy in the distinction of recurrent glioma from radiation injury J. Neurooncol.
84 63–9
4-24

Artificial Intelligence in Adaptive Radiation Therapy
[166] Prat R et al 2010 Relative value of magnetic resonance spectroscopy, magnetic resonance
perfusion, and 2-(
tion of recurrence or grade increase in gliomas J. Clin. Neurosci.
18
F) fluoro-2-deoxy-D-glucose positron emission tomography for detec-
17 50–3
[167] Weybright P et al 2005 Differentiation between brain tumor recurrence and radiation injury
using MR spectroscopy Am. J. Roentgenol.
185 1471
[168] Rock J P et al 2002 Correlations between magnetic resonance spectroscopy and image-
guided histopathology, with special attention to radiation necrosis Neurosurgery
51 912–20
[169] Chen X et al 2024 SC-GAN: Structure-completion generative adversarial network for
synthetic CT generation from MR images with truncated anatomy Comput. Med. Imaging
Graph.
113 102353
[170] Zhao Y et al 2023 Compensation cycle consistent generative adversarial networks (Comp-
GAN) for synthetic CT generation from MR scans with truncated anatomy Med. Phys.
50
4399–414
[171] Zhang Z et al 2018 A predictive model for distinguishing radiation necrosis from tumour
progression after gamma knife radiosurgery based on radiomic features from MR images
Eur. Radiol.
28 2255–63
[172] Beaton L, Bandula S, Gaze M N and Sharma R A 2019 How rapid advances in imaging are
defining the future of precision radiation oncology Br. J. Cancer
120 779–90
4-25

IOP Publishing
Artificial Intelligence in Adaptive Radiation Therapy
Yi Wang and X. Sharon Qi
Chapter 5
Big data for artificial intelligence in
radiation oncology
Jie Fu*, Sunan Cui* and X. Sharon Qi
This chapter explores the role of big data in advancing artificial intelligence (AI)
applications within radiation oncology. With the increasing volume, variety, and velocity
of data in healthcare, radiation oncology has embraced big data to enhance patient care,
streamline workflows, and drive innovations in precision oncology. The chapter begins
by introducing the foundational concept of big data and identifying common data
sources in radiation oncology. It then describes a typical big data lifecycle and illustrates
how AI is normally used for big data analysis. Examples are presented to demonstrate the
use of big data in AI-driven medical image segmentation, treatment planning, treatment
response prediction, quality assurance (QA), and clinical decision support. These
applications have demonstrated the potential to improve accuracy, efficiency, and
personalized treatment in radiation oncology. The chapter concludes with discussion
of the challenges and future perspectives of big data in radiation oncology.
5.1 Introduction to big data in radiation oncology
5.1.1 Overview of big data
With the rapid advancement of digital technologies, more people worldwide are
gaining easier access to the Internet. This expanded connectivity generates vast
amounts of data daily across our digitalized society. The term ‘big data’ refers to
datasets that are too large and complex to be managed using conventional data
management systems and techniques [1].
Big data is often characterized by several critical dimensions, commonly known
as the ‘Vs’ shown in figure 5.1. Initially conceptualized by Doug Laney in 2001, the
three primary characteristics were encapsulated by the terms: volume, velocity, and
variety [1, 2]. Volume refers to the enormous amount of data produced. Velocity
* Both authors contributed equally to this book chapter.
doi:10.1088/978-0-7503-6119-4ch5 5-1 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
Figure 5.1. The 5 Vs (volume, velocity, variety, veracity, and value) of big data.
describes the rapid rate at which data are generated and need to be processed, often
in real-time. Additionally, variety addresses the myriad forms of data, such as text,
images, and videos, adding to the complexity of data management and analysis.
As the field of big data has evolved, additional dimensions have been introduced to
address the broader challenges. Veracity focuses on the reliability and accuracy of
data, recognizing that data can be affected by inconsistencies, biases, and anomalies
[3]. The goal of big data collection and analysis is to derive value, i.e., to distill
meaningful and actionable insights from large and complex datasets [4, 5]. This
expanded framework of the Vs offers a comprehensive understanding of big data.
Addressing the challenges of big data necessitates the advancement and adoption
of innovative technologies. The key elements of big data analysis include parallel
and distributed computing, which enable data processing across multiple computing
nodes, and scalable AI models that can maintain or improve performance as data
volume expands [6, 7]. Additionally, the capability for real-time querying allows the
instantaneous retrieval and analysis of data, which is particularly important when
time-sensitive decisions are required [8]. The infrastructures supporting the processing of big data include distributed file systems that allow for data to be stored across
multiple locations, computing clusters that aggregate the processing power of
numerous computers, and cloud computing that offers scalable and on-demand
computing resources [9, 10]. Furthermore, it is essential to design scalable and userfriendly workflows that process data consistently and enable the application of
analytical methods across different datasets.
The advent of big data has not only transformed the landscape of data analytics but
also revolutionized how we use information in the twenty-first century. Analysing big
data holds the potential to unlock insights that were previously unattainable and
facilitatepeopleto make data-drivendecisions[11, 12]. Withinthe healthcaresector, big
data has the potential to help physicians personalize treatment for patients through the
integration of patient-specific data such as genomics, medical history, etc [13]. It also
5-2

Artificial Intelligence in Adaptive Radiation Therapy
drivesinnovationby uncovering new treatment regimensand guiding adaptivetherapy.
Furthermore, it can optimize hospital operations by streamlining workflows, staff
allocation,and resource utilization.Big data also enhancespatient engagement through
customized communication, educational materials, and treatment plans, thereby
increasing adherenceto medical advice. Ultimately, big data could revolutionize cancer
care by enhancing precision, fostering innovation,improving efficiency, and increasing
patient involvement [14–16].
5.1.2 Sources of big data in radiation oncology
Radiation oncologystands at the forefrontof accumulating digitalpatient data, pivotal
for enhancing patient care through novel big data initiatives [16]. The discipline’slongstanding tradition of data collection makes it an ideal landscape for exploring big data.
Collectingand analyzingdatafromlarge cohortsof patientscouldopen new avenuesfor
improved safety, more efficient workflow, and precision oncology [17–19].
Radiation oncology uniquely amalgamates various types of data from patients
and relies heavily on various computer systems to operate [20]. In the conventional
workflow shown in figure 5.2, physicians would retrieve patient data from electronic
health records (EHRs) to assist in disease diagnosis and clinical decision-making.
These data include patient demographics, medical notes, diagnostic images, lab and
test results, drug prescriptions, etc. For example, pathology and genetic tests are
sometimes requested for cancer diagnosis and tailoring radiation treatments. Once
physicians decide to treat patients with radiation therapy, medical images such as
CT, MRI, PET, or other modalities are acquired using the imaging device software.
These images are then sent to the picture archiving and communication system
Figure 5.2. Typical workflow for radiation therapy along with the acquired data. Abbreviations: EHRs =
electronic health records; DVHs = dose-volume histograms.
5-3

Artificial Intelligence in Adaptive Radiation Therapy
(PACS) using digital imaging and communications in medicine (DICOM) formats
for data storage and image review. The acquired images are transferred from PACS
to the treatment planning system (TPS), where the contours of treatment targets and
organs-at-risk (OARs) are delineated, and radiation treatment plans are created
based on the contours and clinical goals. Dose distributions of the generated plans
are calculated for plan evaluation.
After the plan is approved and reviewed, its machine parameters are transferred
from TPS to the radiation oncology information system (ROIS), which records and
verifies every treatment fraction. ROIS also supports patient scheduling, charting,
image review, etc. Treatment plans also require patient-specifi c QA before treatment
delivery. The whole treatment may be delivered via a single fraction in one day or
tens of fractions across several weeks. Before each treatment fraction, the treatment
delivery system (TDS) retrieves the plan parameters from ROIS to treat patients.
Daily set-up images, such as x-rays, CBCT, or MRI, are acquired to assist in patient
positioning. These daily images and the delivered treatment data, such as machine
parameters, are sent from TDS to ROIS for recording and verification. Additionally,
follow-up exams could be conducted between treatment fractions or months after
the treatment completion to evaluate treatment response and patient outcome. All of
these acquired data contribute to radiation oncology big data that can be
characterized using the aforementioned five Vs.
Big data in radiation oncology can be categorized into three types. First,
structured data refers to any data that can be stored in a relational database in
table format with rows and columns. These include patient information such as
gender and age, prescriptions, plan parameters, delivery records, billing codes, etc.
Second, semi-structured data have a structure but do not fit into the relational
database. These consist of clinical notes, diagnosis reports, patient feedback, etc.
JSON and XML are common types of semi-structured data. Lastly, unstructured
data are unorganized and do not fit into the relational database system. Examples
include medical images, text files, audio, video, etc. Integrating these various data
types into a cohesive analytical framework presents significant challenges, necessitating advanced computational technologies and sophisticated data management
systems.
5.1.3 Big data and AI in radiation oncology
The relationship between big data and AI is both symbiotic and integral for
advancing radiation oncology. Big data serves as essential fuel for AI models, in
particular those based on deep learning, to learn and evolve. The more diverse data
these algorithms are exposed to, the more accurate and effective they become. On
the other hand, big data is often too complex for traditional data processing
methods. AI models can process, analyse, and extract meaningful insights from big
data efficiently. For example, natural language processing (NLP) can be leveraged
to extract relevant information from medical notes [21], and AI models can be
trained to identify anomalies and correlations within big data that would be difficult
for humans to identify.
5-4

Artificial Intelligence in Adaptive Radiation Therapy
The integration of big data and AI allows us to improve clinical workflow and
personalize treatment in radiation oncology. For example, AI models could be
trained to improve the quality of diagnostic images and treatment set-up images [22,
23]. AI-based segmentation models could be used to minimize inter-observer
variations and improve the contour delineation efficiency [24]. AI-based automatic
planning also allows us to standardize and shorten the treatment planning workflow
[25]. Furthermore, AI-based predictive modeling could forecast clinical outcomes,
guiding informed clinical decision-making [26].
5.2 Big data lifecycle in radiation oncology
The big data lifecycle refers to the stages through which big data progresses, from its
generation to its utilization and disposal. Figure 5.3 shows the typical big data
lifecycle in radiation oncology, which includes five stages consisting of data
aggregation and storage, data sharing and security, data visualization, knowledge
creation and implementation, and data archive and deletion.
5.2.1 Data aggregation and storage
Data aggregation is a complex process involving the collection of diverse raw data from
various sources, transforming these raw data into a format suitable for advanced
statistical or machine learning analysis. In the realm of radiation oncology, data
originate from multiple channels such as TPS, TDS, PACS, EMR, ROIS, QA software,
simple CSV files, and spreadsheets [27]. Furthermore, breakthroughs in laboratory
technology have amplified the availability of genomic and proteomic information,
adding valuable resources to the field of radiation oncology. There are several
Figure 5.3. Big data lifecycle in radiation oncology: data aggregation and storage, data sharing and security,
data visualization, knowledge creation and implementation, and data archive and deletion. Abbreviations:
HIPAA = health insurance portability and accountability act; PCA = principal component analysis; t-SNE =
t-distributed stochastic neighbor embedding.
5-5

Artificial Intelligence in Adaptive Radiation Therapy
considerations in the process of data collection. Many systems contain a mix of useful
and irrelevant data. Filtering out pertinent information becomes crucial. For instance,
among numerous plans in a TPS, only specific approved or delivered plans might be
relevant for analysis. Data may not always be stored in convenient formats for analysis.
For instance, images exported from PACS systems are often in DICOM format,
necessitating additional processing to obtain 3D images suitable for analysis. Data from
various sources require establishing unified identifiers.Forexample,toderiveanOAR
dose distribution, linking DICOM CT images, RT structure, and RT dose files becomes
essential for future reference and analysis.
The selection of data to collect is tailored to each specific project. Ideally, a
unified data infrastructure should cater to the diverse needs of projects, such as
prospective/retrospective clinical trials, research documentation, and routine clinical
record-keeping. It is crucial to find a balance between making data retrieval easy,
ensuring storage efficiency, and maintaining a comprehensive dataset. This decision
plays a pivotal role in the data collection process.
Once data are gathered, it is important for them undergo further processing to
guarantee their integrity before saving them into the database. Key aspects to focus
on include accuracy, consistency, and completeness. First, the extracted data should
undergo a validation process to confirm their accuracy. Second, if the same variable
exists in different sources, it is prudent to cross-verify for consistency and rectify any
discrepancies encountered. Lastly, ensuring that all essential information is
extracted is vital. In cases where missing data are encountered, exhaustive efforts
should be made to locate them. If unsuccessful, they should be marked as missing
data. This meticulous approach safeguards the reliability of the stored information,
ensuring its usefulness and accuracy for future analyses.
Data storage serves critical functions for archival and future use. It is imperative
to select a storage solution that boasts scalability, enabling seamless accommodation
of escalating data volumes and ensuring efficient data retrieval even with extensive
datasets. Equally vital is the presence of accurate documentation and metadata. This
documentation acts as a guide, aiding in comprehending the stored information for
future applications. Moreover, ensuring data reliability is paramount. This is
achieved through implementing redundancies within the database and adopting
reliable storage solutions. These redundancies serve as safeguards, protecting against
unforeseen database corruption resulting from unexpected hardware failures.
Additionally, regular archiving of the database reinforces these redundancies,
providing an added layer of protection and ensuring data integrity over time.
5.2.2 Data sharing and security
Data sharing plays a pivotal role in maximizing the utility of existing data sources
for several reasons. First, it ensures the verification and reproducibility of research,
fostering transparency within the scientific community. Second, data sharing
facilitates informed decision-making not only within the originating institution
but also across various entities and systems. Additionally, making data accessible to
a wider audience of researchers allows us to accelerate scientific progress and
5-6
Соседние файлы в папке Библиотека им академика М.И. Перельмана
