| 1 |
How does the concept of “model as a dataset” reshape traditional data-sharing practices in medical imaging?
|
3. It enables sharing of learned model weights instead of sensitive raw images. |
|
It allows for multicenter collaborations due to being able to share model weights and not reveal patient information. |
Under "Summary" : "enables privacy-preserving multicentre collaborations" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 2 |
Which analytical conclusion can be drawn about the trade-offs between physics-informed and statistical models?
|
2. Physics-informed models are more interpretable but computationally intensive. |
|
The writers states that physics informed model can make higher fidelity data but needs lots of computational power and thus being expensive |
Under "Generative models" : " physics informed models offer high fidelity" and "require extensive domain expertise and computational resources" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 3 |
Why is “mode collapse” considered a critical problem in GAN-based medical image synthesis?
|
2. It reduces image realism and variety by producing repetitive outputs. |
|
Mode collapse happen when not all data variation are covered and the model outputting similar images over and over again |
Under "Generative models" : "might not always cover all data variation" and "leading to mode collapse" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 4 |
Why are healthcare-specific metrics preferred over general-purpose metrics such as FID or SSIM?
|
2. They better capture clinical accuracy and diagnostic relevance. |
|
Healthcare-specific metrics are purpose built for the job the images are going to do and thus having more relevance in the field |
Under "Health-care-specific metrics" : "tailored to health-care needs" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 5 |
What does the article identify as the key tension between privacy preservation and image fidelity?
|
1. Higher realism may risk reproducing identifiable patient data. |
|
When a model is trained on a set of data and is able to recreate the data with high realism, it could also recreate an identifiable part of a patient. But removing identifiable features could also remove variation and diversity from the training sample |
Under "Patient privacy and data copying" : "trained on a specific data set", "reveal sensitive patient information", "raises concern about the degree of anonymisation" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 6 |
Why is the FDA’s approval of synthetic MRI technology significant for future AI-generated data?
|
1. It establishes a framework for validating synthetic data equivalence in clinical use. |
|
It is a proof of performance that synthetic images in medicine works and does not degrade the performance of the radiologist. |
Under "Future directions" : "proof of performance", "performance of radiologist remained equivalent" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 7 |
Which strategy would best mitigate demographic bias in generative models according to the article?
|
2. Applying diversity-aware training and fairness constraints |
|
The writer states that targeting the minorities is able to close the fairness gap thus raising awareness is able to mitigate the demographic bias. |
Under "Increased dataset size and diversity" : "targeted oversampling of minoritised ... close the fairness gap by 40%" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 8 |
How do DDPMs exemplify versatility in healthcare image synthesis?
|
2. They can perform multiple tasks such as denoising, inpainting, and anomaly detection without retraining. |
|
They can be repurposed for other tasks without needing to retrain the model to that specific task. |
Under Versatility across talks" : "can be adapted and repurposed", "without any further training can also be used for inpainting" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 9 |
What analytical insight does the article provide about integrating AI-generated medical images into education and research?
|
2. It enhances training by providing diverse, realistic datasets without ethical breaches. |
|
It increases the sample size and thus the diversity of medical images for education. They are also more anatomically accurate than their real counterpart |
Under "Modelling complex biological phenomena" : serving as virtual surgical planning tools and educational resources", "depict the potential progression of a brain tumor over time" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 10 |
Why is regional calibration essential when applying risk prediction models across countries?
|
2. To adjust for population-specific incidence and lifestyle differences |
|
The need for calibration is because countries don't have the same lifestyle ignoring factors that the existing model accounted for could lead to over or underestimation of risk. |
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 11 |
What analytical conclusion can be drawn when comparing the China-PAR and Framingham models?
|
1. Both overestimate CVD risk in East Asians. |
|
The article states that both of the models overestimate the risk in china |
Under "ASCVD risk predictions in China" : "poor prediction for chinese men" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 12 |
Based on CVD mortality data, what analytical inference can be made about Japan’s position compared to neighboring countries?
|
4. Japan’s diet increases risk compared to Mongolia. |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 13 |
What analytical limitation arises when using Western-derived coefficients in East Asian models?
|
2. It introduces systematic overestimation of ASCVD probability. |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 14 |
What policy implication can be derived from country-specific risk models?
|
1. They allow for targeted national prevention programs. |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 15 |
If a model excludes socioeconomic variables, what analytical consequence might occur?
|
2. Ignored non-biological determinants of disease |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 16 |
How might AI improve next-generation ASCVD risk prediction in East Asia?
|
2. By integrating multimodal data, including imaging and lifestyle information |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 17 |
What conclusion can be drawn from comparing Mongolia’s and South Korea’s CVD mortality rates?
|
2. Both have identical age-adjusted mortality rates. |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 18 |
What is the most logical future direction for improving ASCVD models across East Asia?
|
3. Removing local variability from analysis |
|
|
|
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 19 |
According to the “image generation trilemma” shown in the figure, what analytical conclusion can be drawn about the relative strengths of VAEs, GANs, and DDPMs in medical image synthesis?
|
2. GANs provide a balance between image quality and diversity but may suffer from mode collapse. |
|
The writer states that GANs is the middle ground for image quality and diversity but could suffer from a mode collapse which reduces the diversity. |
Under "Figure 2" : "strike a balance, providing good quality and diversity but can suffer from a mode collapse" |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|
| 20 |
Based on Figure, what analytical conclusion can be drawn regarding the distribution of cardiovascular disease (CVD) subtypes across East Asian countries?
|
2. Stroke dominates as the primary cause of CVD death in all East Asian countries equally. |
|
The pie chart shows that stroke dominates as the leading cause of death in the cardiovascular disease category in east asian countries |
the stroke part of the pie chart takes up 48%, 39%, 47% ,and 47% respectively showing a high value |
7 |
-.50
-.25
+.25
เต็ม
0
-35%
+30%
+35%
|