| Question 1: In a secondary analysis of the TROG 02.02 randomized head and neck cancer trial, major deficiencies in radiation plan quality were associated with: |
| Reference: | Peters et al. JCO 2010 Jun 20;28(18):2996-3001. |
| Choice A: | < 5 enrolled patients at a treatment center |
| Choice B: | >1 involved lymph node |
| Choice C: | <10 staff physicists on staff |
| Choice D: | < 4 years of treating physician experience |
| Question 2: (True/False) An analysis of enrollment on RTOG clinical trials, highest minority accrual clusters in areas of highest minority population density. |
| Reference: | Brunner et al. Int J Radiat Oncol Biol Phys. 2015 Nov 1;93(3):702-9 |
| Choice A: | True |
| Choice B: | False |
| Question 3: Twenty percent of patients enrolled in the VA Larynx Trial (NEJM, 1991) were of non-White race. What was the loss to follow up rate? |
| Reference: | Wolf et al. NEJM 1991, 324 (24): 1685-1690. |
| Choice A: | 2% |
| Choice B: | 7% |
| Choice C: | 15% |
| Choice D: | 30% |
| Question 4: A recent study of proton therapy patients treated between 2010-2018 found which disparity among the patient population? |
| Reference: | Bitterman, D.S., et al., Race Disparities in Proton Radiotherapy Use for Cancer Treatment in Patients Enrolled in Children's Oncology Group Trials. JAMA Oncol, 2020. 6(9): p. 1465-1468. |
| Choice A: | Only adults were treated with proton therapy |
| Choice B: | Black pediatric patients were statistically less likely to receive proton therapy |
| Choice C: | Proton therapy was only administered to celebrities |
| Choice D: | Patients turned down proton therapy in favor of Carbon therapy |
| Question 5: (True/False) Insurance denials prevent many cancer patients from receiving proton therapy. |
| Reference: | Ning, M.S., et al., The Insurance Approval Process for Proton Radiation Therapy: A Significant Barrier to Patient Care. Int J Radiat Oncol Biol Phys, 2019. 104(4): p. 724-733. |
| Choice A: | True |
| Choice B: | False |
| Question 6: A recent study of breast cancer treatment for non-English speaking patients found what tool was underutilized? |
| Reference: | Shukla, U., et al. Disparities in Radiation Therapy: Practice Patterns Analysis of Deep Inspiratory Breath Hold Use in Non-English Speakers. Int J Radiat Oncol Biol Phys, 2022. 113(1): p. 21-25. |
| Choice A: | Step stools |
| Choice B: | BB markers |
| Choice C: | Deep Inspiratory Breath Hold |
| Choice D: | Bolus |
| Question 7: (True/False) Cross-population training and testing of AI models is one method of evaluating for population bias. |
| Reference: | Santosh KC. AI-Driven Tools for Coronavirus Outbreak: Need of Active Learning and Cross-Population Train/Test Models on Multitudinal/Multimodal Data. J Med Syst. 2020 May;44(5):93. |
| Choice A: | True |
| Choice B: | False |
| Question 8: (True/False) Site-specific bias has been observed in related imaging fields, such as digital pathology of breast cancer. |
| Reference: | Howard FM, Dolezal J, Kochanny S, Schulte J, Chen H, Heij L, et al. The impact of site-specific digital histology signatures on deep learning model accuracy and bias. Nat Commun. 2021 Dec;12(1):4423. |
| Choice A: | True |
| Choice B: | False |
| Question 9: Feed-forward neural networks update model weights by using _. |
| Reference: | Goodfellow I, Bengio Y, Courville A. Deep learning. MIT press; 2016 Nov 10. https://www.deeplearningbook.org/ |
| Choice A: | Standardization |
| Choice B: | Normalization |
| Choice C: | Bias |
| Choice D: | Backpropagation |
| Choice E: | All of the above |
| Question 10: Which of the following methods may be utilized to mitigate AI bias in algorithms? |
| Reference: | Bellamy RK, Dey K, Hind M, Hoffman SC, Houde S, Kannan K, Lohia P, Martino J, Mehta S, Mojsilovic A, Nagar S. AI Fairness 360: An extensible toolkit for detecting, understanding, and mitigating unwanted algorithmic bias. arXiv preprint arXiv:1810.01943. 2018 Oct 3. |
| Choice A: | Optimized preprocessing |
| Choice B: | Disparate impact remover |
| Choice C: | Adversarial debiasing |
| Choice D: | Calibrated equalized odds post-processing |
| Choice E: | All of the above |
