The purpose of this review article was to cover the current state of the art for deep learning approaches and its limitations, and some of the potential impact on the field of radiology, with specific reference to chest imaging.
Deep_Learning_Applications_in_Chest_Radiography.2.pdf (673.8 KB)
The hardest thing is being confident about algorithm performance, and creating a community of practice to optimize everyone’s AI. I want to take a moment to applaud Stanford’s Machine Learning group, who have established a best practice model with CheXpert. It’s a large database from Stanford clincs, openly available for development, along with an open invitation for anyone to test their algorithm and see how it performs. Love the ‘leaderboard’ they have posted and the open source community approach to something that has been largely locked up in companies and research groups.
Here’s a good resource for digital pathology AI development:
With the dearth of pathologists and radiologists in LMICs, AI could play an important role in access to diagnostics, as it could increase efficiency of those in practice and at least open the discussion of task-shifting in certain situations.
And, welcome to the community, Puneet! Thanks for contributing…