2019 AAPM Annual Meeting
Back to session list

Session Title: Integrating Radiomics and Genomics for Personalized Cancer Therapy in the Era of AI and Big Data
Question 1: Artificial Intelligence:
Reference:Reference: R. Thompson et al, Artificial intelligence in radiation oncology: A specialty-wide disruptive transformation?, Radiotherapy and Oncology, vol 129, 421-426, (2018)
Choice A:A. Can find biological causes why some patients do not respond to radiation therapy.
Choice B:B. Has been widely used in clinic for target delineation and inverse planning in radiation therapy
Choice C:C. Has the potential for clinical decision support in radiation therapy
Choice D:D. Has no role in radiation therapy
Choice E:E. Does not add any value to those conventional clinical characteristics for treatment outcome prediction
Question 2: To evaluate the performance of a machine learning model, which criterion is preferred?
Reference:Haibo He, and Edwardo A. Garcia, Learning from Imbalanced Data, IEEE Transactions on Knowledge and Data Engineering, VOL. 21, NO. 9, pp. 1263-1284, 2009
Choice A:A. Overall accuracy
Choice B:B. Sensitivity
Choice C:C. Specificity
Choice D:D. Area under the receiver operating characteristic curve (AUC)
Choice E:E. Task specific
Question 3: In random forest models, hundreds of trees are generated and the final decision is determined by averaging out predictions across all the trees. Which of the following is true about tree creation? 1. Each tree is constructed using a subset of the features 2. Each tree is constructed using all the features 3. Each tree is constructed using a subset of samples 4. Each tree is constructed using all the samples
Reference:Oh et al. Computational methods using genome-wide association studies to predict radiotherapy complications and to identify correlative molecular processes. Scientific Reports, 2017, 7:43381
Choice A:1 and 3
Choice B:1 and 4
Choice C:2 and 3
Choice D:2 and 4
Question 4: General population structure (population stratification) can be captured by using the first few principal components of SNPs.
Reference:Brzyski et al. Controlling the rate of GWAS false discoveries. Genetics, 2017, 205: 61-75
Choice A:True
Choice B:False
Question 5: Radiomic features can be extracted from which of the following regions of interest:
Reference:Gillies, Robert J., Paul E. Kinahan, and Hedvig Hricak. "Radiomics: images are more than pictures, they are data." Radiology 278.2 (2015): 563-577
Choice A:A. Gross tumor volume
Choice B:B. Intratumoral subregions.
Choice C:C. Peritumoral areas.
Choice D:D. All of the above
Choice E:E. None of the above
Question 6: Radiogenomic studies that integrate imaging and genomics data may be used to address the following questions:
Reference:Wu, Jia, et al. "Radiomics and radiogenomics for precision radiotherapy." Journal of radiation research 59.suppl_1 (2018): i25-i31
Choice A:A. Identify molecular pathways associated with specific imaging features.
Choice B:B. Identify imaging features associated with specific genetic alterations
Choice C:C. Improve response and prognostic outcome prediction
Choice D:D. All of the above
Choice E:E. None of the above
Back to session list