2017 AAPM Annual Meeting
Back to session list

Session Title: ROC: A New Method in Radiotherapy QA
Question 1: A conservative observer/system will have:
Reference:Metz CE. Receiver operating characteristic analysis: a tool for the quantitative evaluation of observer performance and imaging systems. JACR 2006;3:413-422.
Choice A:High true positive and high false positive rate.
Choice B:High true positive and low false positive rate.
Choice C:Low true positive and low false positive rate.
Choice D:Low true positive and high false positive rate.
Question 2: Which of the following does not require target localization?
Reference:Mansson LG. Methods for the evaluation of image quality. Radiat Prot Dosimetry 2000;90:89-99.
Choice A:AFROC.
Choice B:FROC.
Choice C:LROC.
Choice D:ROC.
Question 3: To generate an ROC curve, the data in question must be compared to:
Reference:Rosner, Bernard. Fundamentals of Biostatistics. Boston: Brooks/Cole, Cengage Learning, 2011. Print.
Choice A:A gold standard.
Choice B:Probability distribution function.
Choice C:True negatives.
Choice D:Clinical experience.
Question 4: If the class distributions of your data are highly skewed (e.g. many more passing cases than failing cases) your ROC curve will:
Reference:Fawcett, Tom. "An introduction to ROC analysis." Pattern recognition letters27.8 (2006): 861-874.
Choice A:Shift up.
Choice B:Shift right.
Choice C:Stay the same.
Choice D:Shift down.
Question 5: What is the goal of using the ROC approach with patient specific QA:
Reference:Carlone, M., Cruje, C., Rangel, a, McCabe, R., Nielsen, M., & Macpherson, M. (2013). ROC analysis in patient specific quality assurance. Med Phys, 40(4), 42103. http://doi.org/10.1118/1.4795757.
Choice A:Assess the capability of a given QA system to detect certain type of errors.
Choice B:Evaluate the clinical impact of a given gamma pass rate.
Choice C:Provide a new QA method that does not depend on gamma analysis.
Choice D:Address the increased used of flattening-filter-free radiotherapy delivery .
Question 6: ROC analysis comes from signal detection theory. When applied to patient specific QA, what constitute the signal that needs to be detected?
Reference:Guillot, M., Gingras, L., Archambault, L., Beddar, S., & Beaulieu, L. (2013). Performance assessment of a 2D array of plastic scintillation detectors for IMRT quality assurance. Physics in Medicine and Biology, 58(13), 4439–4454.
Choice A:The patient plan.
Choice B:The measured dose delivery.
Choice C:The error between the planned dose distribution and measured delivery.
Choice D:The probability of an erroneous MLC motion.
Back to session list