How medical AI devices are evaluated: limitations and recommendations from an analysis of FDA approvals

By Eric Wu, Kevin Wu, Roxana Daneshjou, David Ouyang, Daniel E. Ho, James Y. Zou. Nature Medicine, 2021

409 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Health Informatics, Public Health, Environmental and Occupational Health

View Paper: https://www.semanticscholar.org/paper/fc1da632130ad49c648c6e55a7726dddf8622e61

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Ce Zhang
2. Weixin Liang
3. Matei A. Zaharia
4. James Y. Zou
5. Li Fei-Fei
6. Daniel E. Ho
7. Girmaw Abebe Tadesse
8. Indra Elizabeth Kumar

In Health Informatics

1. Ce Zhang
2. Matei A. Zaharia
3. Angela Zhang
4. James Y. Zou

In Public Health, Environmental and Occupational Health

1. Ce Zhang
2. Inioluwa Deborah Raji

Citations in the corpus

• Metaverse for Healthcare: A Survey on Potential Applications, Challenges and Future Directions (2022, neutral): “doctors immediate and trustworthy recommendations [116].”
• Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance (2022, background): “largely deferring to company benchmarks on limited, retrospective data [89].”
• Advances, challenges and opportunities in creating data for trustworthy AI (2022, background): “As a consequence, many AI algorithms, especially in the biomedical sciences, are developed using data from only one or a small number of sites, which limits generalizability 21 , and evaluation of subgroup performance can be challenging 92 .”
• The Fallacy of AI Functionality (2022, background): “In fact, healthcare is the field where AI product failures appear to be the most visible, in part due to the rigor of pre-established evaluation processes [20, 119, 160, 205].”
• Shifting machine learning for healthcare from development to deployment and from models to data (2022, background): “Current regulation and approval processes are being adapted to meet the emerging needs; in particular, initiatives are attempting to address data shifts and patient representation in the training datasets 165,180,181 .”
• Medical Image Segmentation Review: The Success of U-Net (2022, background): “Common medical imaging modalities for acquiring data are X-ray, Positron Emission Tomography (PET), Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and Ultrasound (US) [4].”