2017 AAPM Annual Meeting
Back to session list

Session Title: Big Data 1: Current Big Data Resources and Technology in Radiation Oncology
Question 1: Why are doctors bad at predicting outcomes?
Reference:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4555846/ Predicting outcomes in radiation oncology —multifactorial decision support systems Philippe Lambin, Ruud G. P. M. van Stiphout, Maud H. W. Starmans, et al Nat Rev Clin Oncol. 2013 Jan; 10(1): 27–40.
Choice A:They get confused by too much data.
Choice B:They have too many options to choose from.
Choice C:There is not enough good evidence to make a choice.
Choice D:All of the above.
Question 2: What is the best way to validate a model?
Reference:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4555846/ Predicting outcomes in radiation oncology —multifactorial decision support systems Philippe Lambin, Ruud G. P. M. van Stiphout, Maud H. W. Starmans, et al Nat Rev Clin Oncol. 2013 Jan; 10(1): 27–40.
Choice A:Bootstrap your training set to create a validation set.
Choice B:Send the model to another hospital and have them validate it.
Choice C:Get data from another hospital to validate.
Choice D:Split your own data before modeling into a training and validation set.
Question 3: What is generally the best way to improve models in machine learning?
Reference:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4555846/ Predicting outcomes in radiation oncology —multifactorial decision support systems Philippe Lambin, Ruud G. P. M. van Stiphout, Maud H. W. Starmans, et al Nat Rev Clin Oncol. 2013 Jan; 10(1): 27–40.
Choice A:Remove outlying patients from the training dataset.
Choice B:Extend the training dataset with more patients.
Choice C:Extend the training dataset with more features.
Choice D:Use a better machine learning algorithm.
Question 4: The role of prior knowledge in autonomous and knowledge based treatment planning are, EXCEPT:
Reference:Xing L, Using Population-Based Prior Knowledge to Autopilot Radiation Therapy Treatment Planning, Medical Physics 44, 389-96, 2017.
Choice A:Provides guidance and benchmarks during the solution search for a case under planning.
Choice B:Allows treatment plans for various disease sites to be compared to many other similar plans.
Choice C:Can include adverse effects of radiation on normal tissues.
Choice D:Can include adverse effects of radiation on normal tissues.
Question 5: Radiomics is an emerging field of research and promises to advance significantly patient care. All of the following are true, EXCEPT
Reference:Cui Y, Tha KK, Terasaka S, Yamaguchi S, Wang J, Kudo K, Xing L, Shirato H, Li R. Prognostic Imaging Biomarkers in Glioblastoma: Development and Independent Validation on the Basis of Multiregion and Quantitative Analysis of MR Images. Radiology. 2016 Feb;278(2):546-53.
Choice A:It uses a large amount of imaging data for prognostic prediction of outcome of the patients.
Choice B:It quantifies tumor phenotypic characteristics by applying feature algorithms to imaging data.
Choice C:It is used in current knowledge based treatment planning.
Choice D:It represents a significant component of big data initiatives aimed at drawing inferences from large data sets that are not derived from carefully controlled experiments.
Question 6: Which answer is not a major challenge for the implementation of Learning Health Systems for Radiation Oncology?
Reference:McNutt T., Moore K., Quon H. “Needs and Challenges for Big Data in Radiation Oncology,” Int’l J. of Radiation Oncology, Biology, Physics. Published online: November 27 2015.
Choice A:Clinical culture and documentation to support the quantification of the patient condition.
Choice B:Database technology to enable large scale access to complex RT data.
Choice C:Data science models that can predict outcomes and are consistent with existing knowledge.
Choice D:Data completeness and integrity.
Back to session list