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

The primary goal of the article is to provide a comprehensive overview of synthetic data generation in medical imaging using generative AI, while critically examining its underlying generation paradigms, applications, benefits, challenges, and future research directions.

Introduction section, final paragraph (Page 1) “This Viewpoint provides a comprehensive overview of synthetic data in medical imaging and critically analyses the advancements, applications, and challenges of this field.”

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

The article explains that discriminative models are designed to distinguish between classes or make decisions based on existing data. In contrast, generative AI learns the underlying distribution of data and can generate entirely new examples that resemble real-world medical images. This capability enables synthetic dataset creation, data augmentation, and simulation of biological phenomena, which are beyond the scope of traditional discriminative approaches.

Introduction section, first paragraph (Page 1) “Generative artificial intelligence is a class of deep learning models capable of creating content that diverges from traditional discriminative models focused on interpretation or decision making.”

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

Instead of distributing raw medical images, researchers can share trained model weights. These weights encode statistical characteristics and relationships found in the original dataset. Other researchers can then generate synthetic images that preserve important properties of the original data while potentially improving privacy and data-sharing efficiency.

Paragraph beginning: “The advancement of generative artificial intelligence introduces a new concept in data sharing, which we refer to as a model as a dataset.”

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

Physics-informed models rely on domain expertise, mathematical equations, and known biological or physical mechanisms (e.g., tissue biomechanics, blood flow, radiation physics). Statistical models, such as VAEs, GANs, and diffusion models, learn from examples in training datasets and generate new images by modeling the statistical structure of the data rather than explicitly encoding physical rules.

Physics-informed models paragraph: “Physics-informed models are primarily rule-based approaches that incorporate domain-specific knowledge and physics principles through mathematical equations.”

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

The article explains that no current generative model can simultaneously maximize all three properties. VAEs are fast and diverse but may sacrifice image quality. GANs produce high-quality images but can suffer from mode collapse and reduced diversity. Diffusion models provide high-quality and diverse outputs but require slower sampling. Researchers must choose a model depending on which trade-offs best suit their application.

Paragraph beginning: “Statistical models encounter the generative artificial intelligence trilemma…”

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

Although computational metrics can assess image quality, human evaluation remains the gold standard. In the Human Turing Test, domain experts such as radiologists review images and attempt to determine whether they are real or AI-generated. If experts struggle to distinguish them, it suggests the synthetic images have high perceptual realism.

Evaluating image quality → Human evaluation section (Page 3) “The human Turing test involves domain experts who are asked to discern between real and derived medical images.”

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

The article discusses many benefits, including: * Increased dataset size and diversity * Privacy preservation * Data augmentation * Improved fairness for underrepresented groups * Modelling complex biological phenomena * Medical education and training

Potentials and promises section (Pages 4–5)

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

Even though synthetic datasets are intended to anonymize patient information, generative models may inadvertently reproduce images that closely resemble training examples. Medical images may contain identifiable anatomical features, creating a risk that sensitive patient information could be exposed.

Source in the Article Challenges and considerations → Patient privacy and data copying section (Page 7)

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

The authors highlight that regulators have already begun accepting synthetic data approaches in certain contexts. This example demonstrates that synthetic data technologies are becoming increasingly recognized within formal regulatory frameworks.

Source in the Article Challenges and considerations / Future directions section

โจทย์ ปรนัย
What is the main purpose of the article?

The article serves as a viewpoint paper that summarizes the current state of generative AI in medical imaging, evaluates different image-generation paradigms, discusses potential clinical and research applications, examines ethical concerns, and outlines future research priorities.

Source in the Article Abstract and Introduction (final paragraph)

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

The PCE was derived from four large U.S. community-based cohort studies and contains relatively few Asian participants. Therefore, its risk estimates may not accurately reflect cardiovascular risk patterns in East Asian populations.

Source in the ASCVD Article “Current State of ASCVD Risk Calculators for East Asian Populations” section

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

The article explains that coronary heart disease (CHD) incidence is generally lower in East Asian populations, while stroke contributes a larger proportion of cardiovascular disease burden. Western cohorts also differ in cholesterol levels, smoking patterns, and baseline cardiovascular risk. As a result, applying Western models directly to East Asian populations can lead to inflated risk estimates.

Source in the ASCVD Article Future Directions and Conclusions section

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

Unlike Western-based models such as Framingham or PCE, China-PAR incorporates risk-factor distributions and disease patterns specific to the Chinese population. This improves risk prediction accuracy and reduces the miscalibration that occurs when Western models are applied to East Asian populations.

Source in the ASCVD Article ASCVD Risk Prediction in China section

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

Alcohol use is not a core variable included in most ASCVD risk prediction models. The article shows that traditional predictors such as blood pressure, cholesterol, diabetes, and smoking are consistently used, whereas alcohol use is included only in some East Asian calculators

Evidence from the article: Central Illustration: East Asian Cardiovascular Risk Calculators

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

The article explains that Framingham-based models often overestimate cardiovascular risk when applied to East Asian populations. The Suita Score was specifically developed and validated in Japan to better estimate CHD risk in Japanese individuals.

Evidence from the article ASCVD Risk Prediction in Japan”

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

Western risk calculators were developed primarily from White American cohorts and often overestimate ASCVD risk in East Asians. Region-specific models can better reflect local epidemiology and risk factor patterns.

Evidence from the article: Abstract (p. 333) The authors state: “current risk stratification models are not adequate to predict the development of ASCVD in East Asian Americans.” Future Directions and Conclusions The article notes “ASCVD risk is significantly overestimated… when applying calculators developed in the United States including the FRS and PCE.”

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

The article repeatedly emphasizes that China, Japan, and Korea have different patterns of CHD, stroke, mortality, and risk factors, which affect ASCVD risk estimation.

Evidence from the article: Introduction The authors state: “epidemiological characteristics of ASCVD vary among East Asian persons.”

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

The authors conclude that current models require refinement and broader validation. They recommend incorporating imaging findings, biomarkers, and East Asian–specific risk factors.

Future Directions and Conclusions The article states: “The new risk prediction models being developed in East Asian countries should standardize the definition of ASCVD…”

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

VAEs generate images by encoding data into a latent space and then decoding it back into images. * GANs generate images through competition between a generator and a discriminator. * DDPMs generate images by gradually removing noise from random noise through a reverse diffusion process.

This is the fundamental distinction shown in Figure 1.

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?

Crude mortality rates are affected by the proportion of older people in a population. Countries with older populations may have higher crude mortality simply because more people are in age groups at greater risk of CVD. Age-standardization adjusts for these age differences, making comparisons between countries more meaningful.The article also emphasizes that East Asian countries have different demographic and epidemiologic characteristics, which is why age-standardized rates are used for fair comparison.

Evidence from the article Figure 1 (p. 335) “Age-Standardized and Crude Mortality Rates for Major CVD (2019)”