2021 AAPM Virtual 63rd Annual Meeting
Back to session list

Session Title: Adaptive Radiation Therapy in the Era of Big Data with Artificial Intelligence
Question 1: 1. AI tools demonstrate significant efficiency promotion in following aspects during adaptive therapy:
Reference:(1) The Emergence of Artificial Intelligence within Radiation Oncology Treatment Planning. Tucker J. Netherton, Carlos E. Cardenas, Dong Joo Rhee, Laurence E. Court, Beth M. Beadle. Oncology, 2021;99(2):124-134. (2) Artificial intelligence in radiation oncology. Huynh E, Hosny A, Cuthier C, et al. Rev Clin Oncol. 2020 Dec;17(12):771-781.
Choice A:a) Synthetic CT generation
Choice B:b) Auto-segmentation
Choice C:c) Auto-planning
Choice D:d) Patient specific QA
Choice E:e) All of the above
Question 2: 2. What are advantages of AI-based automated contouring over Atlas-based ones?
Reference:(1) Jackson P, Kron T, Hardcastle N. A future of automated image contouring with machine learning in radiation therapy [published online ahead of print 2019/12/20]. J Med Radiat Sci. 2019;66(4):223-225. (2) Seo H, Badiei Khuzani M, Vasudevan V, et al. Machine learning techniques for biomedical image segmentation: An overview of technical aspects and introduction to state-of-art applications. Med Phys. 2020;47(5):e148-e167.
Choice A:a) High accuracy
Choice B:b) Efficient inference time after model training
Choice C:c) Generalizability across different database
Choice D:d) (a) & (b)
Question 3: What are the challenges of AI-based automated contouring for adaptive radiotherapy application?
Reference:(1) Seo H, Badiei Khuzani M, Vasudevan V, et al. Machine learning techniques for biomedical image segmentation: An overview of technical aspects and introduction to state-of-art applications. Med Phys. 2020;47(5):e148-e167. (2) Liang X, Chen L, Nguyen D, et al. Generating synthesized computed tomography (CT) from cone-beam computed tomography (CBCT) using CycleGAN for adaptive radiation therapy [published online ahead of print 2019/05/21]. Phys Med Biol. 2019;64(12):125002.
Choice A:a) Scarcity of ground-truth labels
Choice B:b) Variation of ground-truth labels
Choice C:c) Image-quality variation on different on-board imaging systems
Choice D:d) (a), (b) & (c)
Question 4: How can RapidPlan DVH estimation model be used in Varian Ethos therapy?
Reference:Y Archambault, C Boylan, D Bullock, T Morgas, JPeltola, ERuokokoski, A Genghi, B Haas, P Suhonen, S Thompson. Making on-line adaptive radiotherapy possible using artificial intelligence and machine learning for efficient daily re-planning. Medical Physics International Journal, Vol 8(2), page 77-86, 2020.
Choice A:a) Can be used alone as the primary engine for both initial planning and online adaptation, similar to how it is used in Eclipse treatment planning system.
Choice B:b) Can be used initial planning in Ethos therapy in a similar manner to Eclipse treatment planning system but cannot be used yet for online adaptation.
Choice C:c) Can be used as a quality monitor to remind the planner if the automated plan generation cannot achieve a result within the DVH predictions.
Choice D:d) The Intelligent Optimization Engine (IOE) can effectively monitor and modify the strength of the line objective derived from the Rapid plan model.
Question 5: 5. Which of the following machine learning models have been investigated for patient specific VMAT QA?
Reference:M F Chan, A Witztum, G Valdes. Integration of AI and Machine Learning in Radiotherapy QA. Frontiers in Artificial Intelligence, 29(9), Article 577620, 2020.
Choice A:a) Convolutional neural network
Choice B:b) Random forest
Choice C:c) Support vector classifier
Choice D:d) All the above
Question 6: 6. If a radiation treatment plan is Pareto optimal with respect to local tumor control and a set of specific toxicities, then
Reference:https://en.wikipedia.org/wiki/Multi-objective_optimization
Choice A:a) Local tumor control cannot be improved without increasing the probability of one or more of the specified toxicities;
Choice B:b) The variance of the tumor dose distribution will increase if the dose to any of the relevant organs at risk (OARs) is reduced;
Choice C:c) The root-mean-square deviation between the voxel-wise prescription dose and the measured or calculated absorbed dose is at a (local) minimum.
Choice D:d) All of the above
Question 7: 7. Is any of the following statement true regarding radiomics and delta radiomics?
Reference:Nasief, H., Zheng, C., Schott, D., Hall W., Tsai S., Erickson B., Li X.A. A machine learning based delta-radiomics process for early prediction of treatment response of pancreatic cancer. npj Precis. Onc. 3, 25 (2019). https://doi.org/10.1038/s41698-019-0096-z
Choice A:a) Radiomics deals with extracting higher dimensional data from images through advanced imaging processing and analysis tools.
Choice B:b) Radiomics involves in mining imaging data, that extend beyond what is visible to human eye, for improved diagnosis and/or prognosis decision.
Choice C:c) Delta radiomics studies the change of radiomic features by analyzing longitudinal image sets.
Choice D:d) All of above
Question 8: What is adaptive radiation therapy?
Reference:XA Li, ed. Adaptive Radiation Therapy. Phys. Taylor & Francis, 2011.
Choice A:a) Creating a boost plan
Choice B:b) Modifying patient contours mid-treatment
Choice C:c) Modifying radiation plans during treatment to account for inter-fraction variations
Choice D:d) Adapting imaging procedures to improve anatomical visualization
Question 9: 9. What is the classical mathematical framework for modeling decision making?
Reference:A Guide to Outcome Modeling In Radiotherapy and Oncology Listening to the Data. Edited By Issam El Naqa, 2018.
Choice A:a) Poisson Process
Choice B:b) Moran Process
Choice C:c) Markov Process
Choice D:d) None of above.
Question 10: A machine learning algorithm typically used for optimizing decision making is based on:
Reference:A Guide to Outcome Modeling In Radiotherapy and Oncology Listening to the Data. Edited By Issam El Naqa, 2018.
Choice A:a) Random forest
Choice B:b) Reinforcement learning
Choice C:c) Active learning
Choice D:d) Community learning
Back to session list