โจทย์ ปรนัย
What is the primary goal of the article according to its introduction?

The last paragraph of the introduction gives a brief overview of the paper's purpose,

The last paragraph of the introduction explains that the paper is meant to explore give a "comprehensive overview of synethetic data in medical imaging" as well as claiminng to "critically analyse the advancements, applications, and challenges of this field." These different points are summarized by the answer chosen.

โจทย์ ปรนัย
How do generative AI models differ from traditional discriminative models in healthcare applications?

The first paragraph of the introduction displays the difference between generative models and traditional discriminative models.

Not only does "Generative" mean "with the ability to create," which is only properly represented by the answer chosen, but the first paragraph of the introduction how Generative models differ from discriminative models, where the former is built around the production of data, whereas the latter is based around the understanding and classifying of data.

โจทย์ ปรนัย
What is meant by the term “model as a dataset”?

The paper refers "a model as a dataset" as basing model weights on other pre-established model weights.

The paper argues that sharing already-trained model weights to other models is an efficient and effective method of mimicking true sources, and refers to these model-made datasets as "models as a dataset."

โจทย์ ปรนัย
Which statement correctly distinguishes physics-informed and statistical models?

The paper states that "physics informed models are primarily rule-based approaches that incorporate domain-specific knowledge and physics principles through mathematical equations and plausible data," whereas statistical models "learn from data patterns and distributions."

The question asks us to determine the distinguishing factor between physics and statistics-based models. The paper explains that physics based models utilize the laws of physics known to itself to simulate a result, which directly matches with the answer chosen.

โจทย์ ปรนัย
According to the article, what does the “image generation trilemma” describe?

Figure 2 in the paper shows a triangular graph with each point of the triangle representing a different factor that must be compromised on for a model to be able to reasonably represent the two other factors. These three points include Quality, Diversity, and Speed.

A trilemma describes a situation in which three options are presented, but not all can be achieved. This represents a three-way trade-off as described by the answer chosen. The three points which are described as having to be compromised between in the paper (Diversity, Quality, and Speed) are the same as the three points displayed in the answer chosen.

โจทย์ ปรนัย
What is the Human Turing Test used for in medical image synthesis?

The paper describes how the data synthetically generated by models must be validated as plausible and accurate before use in the medical field. One of the suggested methods to assess the quality of the data is through a Human Turing Test.

The Human Turing Test is an examination which asks peoples to differentiate synthetically generated data from real data. By using experts to discriminate between the two, the paper means to describe how synthetic images can be compared to real ones, offering valuable data to experts involved.

โจทย์ ปรนัย
Which of the following is NOT mentioned as a potential benefit of synthetic data in healthcare?

The paper describes how, especially if trained on datasets with low geographic diversity, certain biases may be enhanced when generating images.

As denoted by the paper, the result of an image-generating model relies on the dataset given to it. Populations with low amounts of representation may be improperly represented by the model. Older models especially struggle with this issue, and while newer models are able to vastly mitigate the indiscriminant nature of their predecessors, it is still extremely difficult to completely eliminate all medical biases permanently as described in the chosen answer.

โจทย์ ปรนัย
What is one major ethical concern associated with generative AI in medical imaging?

While the use of image generation can vastly mitigate the risks of exposing a patient's medical history by only focusing on specific issues as denoted by the paper, generative models are still able to accidentally recreate images that closely resemble the original dataset of the patient, potentially revealing private information.

As stated above, while generative models are able to create anonymous images that are not directly linked to the patient involved, if the dataset that the model is trained upon is too specific, then sensitive data may be recreated and spoil the anonymity of the generated image.

โจทย์ ปรนัย
What regulatory precedent did the article cite for synthetic data technologies?

The second-to-last paragraph of the paper denotes the "FDA's clearance of synthetic MRI technologies" as a framework for evaluating synthetic medical imaging.

While the paper recognises the importance of analyzing concerns with generative models in the medical field, it also mentions how it is important for large organizational bodies such as the FDA or European medical agencies to set standards and frameworks for the future use of generative models in medicine, citing the FDA's approval of synethetic MRI technologies as an example of the sort of framework that should be set out, as denoted by the chosen answer.

โจทย์ ปรนัย
What is the main purpose of the article?

The paper specifically focuses on Asian, and especially East Asian persons and different models used in risk-modelling of relevant persons. This is only properly represented in the chosen answer.

Throughout the entire paper, the focus is always on the risk that East Asians face against ASCVD. Western countries are mostly only mentioned to include East Asian subpopulations that have emmigrated from their places of origin, and the economic issues presented by cardiovascular disease are not elaborated on, thus leaving only the chosen answer by processs of elimination.

โจทย์ ปรนัย
Which of the following models was originally developed for a Western population?

The Framingham study is one of the sources mentioned in the paper to have been used to create a risk calculator for the Chinese population, in which it severely overestimated the absolute CHD risk in the Chinese population.

All other models available to be chosen as an answer were developed in response to the need for an accurate risk calculator for ASCVD in their respective countries and regions, represented by the fact that their overestimations and underestimations for ASCVD risk are minimal. When the Framingham score was used on a Chinese population, it severely overestimated the risk of absolute CHD.

โจทย์ ปรนัย
Why might Western-based risk prediction models overestimate ASCVD risk in East Asian populations?

The paper mentions how due to differences in biology, compared to Western populations, East Asian populations have a different baselines for many diseases, including obesity and ASCVD.

The paper denotes how East Asian populations and Western populations are subject to different environmental stimuli, causing them to have differing baselines to be classified/considered affected by particular phenomenon like non-communicable diseases. While data collection standards are similar in the West and in East Asia, it was also mentioned that East Asians have overall lower cholesterol levels than Western populations, leaving only the chosen answer through process of elimination.

โจทย์ ปรนัย
What is the key advantage of the China-PAR model compared to Western-based models?

The paper mentions that, during a study in China, it was found that the CHD risk for both men and women was severely overestimated. Additionally, it was found that Chinese men were represnted with a low discrimination ability and poor calibration.

โจทย์ ปรนัย
Which of the following variables is not typically included in ASCVD risk prediction models discussed in the article?

In the second paragraph of the section for ASCVD risk prediction in Japan, the paper mentions numerous factors included in the risk calculation model and output using the Suita score.

As mentioned above the Suita score is calculated using a number of factors mentioned. These factors include age, sex, smoking, blood pressure level, HDL, LDL-C, impaired glucose tolerance, and family history. All of which are potential answers except for the chosen answer. Additionally, Serum Cholesterol is seldom mentioned in most other risk-calculation models.

โจทย์ ปรนัย
What is a major difference between the Suita Score and the Framingham Risk Score?

The section for ASCVD risk prediction in Japan mentions the factors going into the calculation of the Suita score, and the ASCVD Risk Prediction in China section mentions the workings of the Framingham score.

The Framingham score was a general ASCVD risk prediction model mainly developed for a western population while the Suita score was chosen from chosen from ten different risk scores published in Japan, displaying how the Suita score was one of many risk calculation models determined to be made specifically for the Japanese population.

โจทย์ ปรนัย
According to the article, what is a potential benefit of developing East Asia–specific risk models?

The Future Directions and Conclusions section mentions potential benefits of developing an East Asia-specific risk calculation model.

The section mentioned above denotes that while Western population-based risk prediction models are reasonably accurate after recalibration when used on East Asian populations, they still vastly overestimate the risk levels of the latter. This implies that the development of an East Asian-specific risk model would greatly increase the accuracy and reduce the overestimation of risk as mentioned by the chosen answer.

โจทย์ ปรนัย
Which factor was highlighted as influencing ASCVD risk differences among East Asian countries?

The Future Directions and Conclusions section mentions sources of differences among East Asian countries.

It is denoted in the section mentioned above that each East Asian country has vastly different geopgraphic influences, whether it be cultural, genetic, or environmental.

โจทย์ ปรนัย
What future direction does the article suggest for improving ASCVD risk prediction?

The Future Directions and Conclusions section's final paragraphs include insights on how ASCVD risk prediction models can be improved.

The section mentioned above denotes that there is a distinct lack of East Asian populations in studies and datasets. By encouraging multimodal AI-based predictions integrated with regional data, the lack of representation for different East Asian subpopulations can be made up for and accuracy may be increased.

โจทย์ ปรนัย
Which statement best explains the key difference in how VAEs, GANs, and DDPMs generate medical images according to the figure?

The figure displays how each model generates images.

All answers except the chosen one improperly represent the workings of each generative model according to the figure. The chosen answer properly displays how DDPMs utilize noise diffusion in order to generate images from an input.

โจทย์ ปรนัย
Which of the following best explains the trend shown in Figure comparing age-standardized and crude CVD mortality rates among East Asian countries?

The figure displays the crude mortality rates and age-adjusted mortality rates of East Asian countries.

A couple of the possible answers reason using the amount of population to explain their points, which is untrue because the graphs represent a rate that is independant of the population size. The third answer is simply wrong, claiming that Japan maintains low mortality rates in both measures despite having relatively high age-adjusted mortality rates. The fifth option fails to properly explain how a high stroke rate corrolates to poor control of infectious diseases, as strokes are caused by non-communicable diseases. Thus, through process of elimination, the only remaining option is the chosen answer.