Question 1: 1. What are different ways to implement AI in clinical practice |
Reference: | Giger ML, Karssemeijer N, Schnabel J: Breast image analysis for risk assessment, detection, diagnosis, and treatment of cancer. Annual Review of Biomedical Engineering15:327-357, 2013. |
Choice A: | Second reader |
Choice B: | Primary reader for triaging |
Choice C: | Autonomous reader |
Choice D: | All of the above |
Question 2: 2. Conceptually, what is the difference between using human-engineered radiomic features and deep learning in characterizing tumors for assessing treatment response? |
Reference: | Giger ML: Machine Learning in Medical Imaging. J Am Coll Radiol. 2018 Mar;15 (3 Pt B):512-520. doi: 10.1016/j.jacr.2017.12.028. Epub Feb 2, 2018. |
Choice A: | Intuitive understanding |
Choice B: | Benefit to the prognosis |
Choice C: | Calculation time once system is trained |
Choice D: | Need for medical truth for the evaluation |
Question 3: What may be characteristics of therapeutic biomarkers? |
Reference: | Sahiner B, Pezeshk A, Hadjiiski LM, Wang X, Drukker K, Cha KH, Summers RM, Giger ML: Deep learning in medical imaging and radiation therapy. Medical Physics, 2018 |
Choice A: | Single feature characteristic |
Choice B: | Merged characteristic via a tumor signature |
Choice C: | Change in biomarker over treatment |
Choice D: | Correlation with cancer subtypes |
Choice E: | All of the above |
Question 4: The biological evolution of species acts through differences in survival based most directly on: |
Reference: | Wagner GP, Altenberg L. Perspective: complex adaptations and the evolution of evaluability. Evolution. 1996 Jun;50(3):967-76. https://doi.org/10.1111/j.1558-5646.1996.tb02339.x |
Choice A: | The genotype |
Choice B: | The phenotype |
Choice C: | The metabolome |
Question 5: Single gene effect sizes with respect to complex phenotype endpoints have a distribution that is: |
Reference: | Chesmore, K., Bartlett, J. & Williams, S.M. The ubiquity of pleiotropy in human disease. Hum Genet 137, 39–44 (2018). https://doi.org/10.1007/s00439-017-1854-z |
Choice A: | Gaussian |
Choice B: | Linear |
Choice C: | Exponentially decaying |
Question 6: The pre-conditioned random forest regression (PRFR) machine learning method has been used to show that: |
Reference: | Oh JH, Kerns S, Ostrer H, Powell SN, Rosenstein B, Deasy JO. Computational methods using genome-wide association studies to predict radiotherapy complications and to identify correlative molecular processes. Scientific reports. 2017 Feb 24;7:43381. https://doi.org/10.1038/srep43381 |
Choice A: | Large variations in risk between people based on genetic factors are common |
Choice B: | Many genes typically contribute to the probability of a complex phenotype endpoint |
Choice C: | Closely related genes (sub-networks/modules) often appear among the predictors |
Choice D: | All of the above |
Question 7: Which of the following are useful in the management of patients with colorectal cancer? |
Reference: | Irene Yu and Winson Cheung: Metastatic Colorectal cancer in the Era of Personalized Medicine: A more tailored approach to systemic therapy. Canadian Journal of Gastroenterology and Hepatology. Volume 2018, Article ID 9450754, 11 pages https://doi.org/10.1155/2018/9450754 |
Choice A: | Kras mutation |
Choice B: | MSI status |
Choice C: | B-RAF mutation |
Choice D: | all of the above |
Question 8: Radiomics is a brunch of imaging sciences which generally includes the following sequential process: |
Reference: | Chintan Parmar, Patrick Grossmann, Johan Bussink, Philippe Lambin, Hugo J. W. L. Aerts. “Machine learning methods for Quantitative Radiomic Biomarkers.” Scientific Reports | 5:13087 | DOI: 10.1038/srep13087 5 |
Choice A: | Image acquisition, image segmentation, feature extraction, feature analysis |
Choice B: | CT and MRI, image segmentation, deep learning, prediction model |
Choice C: | Artificial intelligence, auto-segmentation, feature analysis, prediction model |
Choice D: | Texture analysis, image segmentation, machine learning, precision medicine |
Choice E: | Feature calculation, auto-segmentation, feature analysis, precision medicine |
Question 9: The following factors affect the calculated values of radiomics features |
Reference: | Chang, Yushi; Lafata, Kyle; Wang, Chunhao; Duan, Xiaoyu; Geng, Ruiqi; Yang, Zhenyu; Yin, Fang-Fang. “Digital Phantoms for Characterizing Inconsistencies among Radiomics Extraction Toolboxes." Biomedical Physics & Engineering Express Biomed. Phys. Eng. Express 6 (2020) 025016 |
Choice A: | Texture pattern in the image |
Choice B: | Image quality |
Choice C: | Algorithm used for parameter calculation |
Choice D: | Algorithm used for radiomics feature calculation |
Choice E: | All above |
Question 10: Treatment outcome prediction using radiomics features extracted from the free-breathing thoracic CT images may differ from that extracted from the breath hold thoracic CT images |
Reference: | K. Lafata, R. Geng, B. Ackerson, J. Hong, C. Kelsey, Z. Zhou, J.G. Liu, F.F. Yin. “Association of Pre-treatment Radiomic Features with Lung Cancer Recurrence Following Stereotactic Body Radiation Therapy.” Phys Med Biol. 2019 Jan 8;64(2):025007. doi: 10.1088/1361-6560/aaf5a5. |
Choice A: | True |
Choice B: | False |