A comprehensive, machine-interpretable taxonomy of AI performance metrics https://doi.org/10.1148/ryai.260070 #ACR2026 #RadEcon #RadLeaders #radiology
Charles Kahn
@cekahn.bsky.social
Spouse, dad, grandpa, physician, AI researcher, journal editor
Supporting communication in Radiology AI with ROADMAP, a new ontology https://doi.org/10.1148/ryai.260069 #JPR2026 #radiologia #radiology
2 papers published today in @radiology-ai.bsky.social! Medical AI needs more than better models. It needs infrastructure for machine-readable models, datasets and evaluation. 📝ROADMAP: doi.org/10.1148/ryai... 📝Metrics: doi.org/10.1148/ryai... Thanks to @cekahn.bsky.social et al!
New commentary! Random convolutions for improving segmentation across different medical imaging domains https://doi.org/10.1148/ryai.251014 #ASPNR26 #PedsRad #NeuroRad #radiology
Introducing the LumbarDISC Dataset, the largest publicly available adult MRI lumbar spine dataset for degenerative disease https://doi.org/10.1148/ryai.250480 @utah.edu #MedicalImaging #MSKRad #MachineLearning
Breast Cancers Detected and Missed by AI-CAD: Results from the AI-STREAM Trial https://doi.org/10.1148/ryai.250281 #BreastCancer #mammography #ML
It's been great to see @amia.org #informatics friends & colleagues in Atlanta! For those interested in #AI, come check out the Radiology: Artificial Intelligence journal (rsna.org/ai) and the journal blog, The Vasty Deep. #AMIA2025 #medsky
With this new blogging platform, we look forward to creating a lively forum for exchange of ideas https://radiologyai.substack.com/p/the-vasty-deep-we-begin-anew @cekahn.bsky.social @rsnasky.bsky.social #AMIA2025 #AMIA25 #informatics #AI
New blog post! Pouria Rouzrokh discusses the implications of prompt injection attacks for peer review https://radiologyai.substack.com/p/ai-manipulation-of-the-peer-review #AI #ML #MachineLearning
It's been great to see @amia.org #informatics friends & colleagues in Atlanta! For those interested in #AI, come check out the Radiology: Artificial Intelligence journal (rsna.org/ai) and the journal blog, The Vasty Deep. #AMIA2025 #medsky
With this new blogging platform, we look forward to creating a lively forum for exchange of ideas https://radiologyai.substack.com/p/the-vasty-deep-we-begin-anew @cekahn.bsky.social @rsnasky.bsky.social #AMIA2025 #AMIA25 #informatics #AI
Rigor vs. transparency in @US_FDA marketing authorization of AI-enabled #MedicalDevices https://doi.org/10.1148/ryai.250369 @danafarber.bsky.social @fda.gov #AMIA2025 #AMIA25 #informatics #AI
Characterizing the Impact of Training Data on Generalizability: Application in Deep Learning to Estimate Lung Nodule Malignancy Risk https://doi.org/10.1148/ryai.240636 #LungCancer #cancer #MachineLearning
With this new blogging platform, we look forward to creating a lively forum for exchange of ideas https://radiologyai.substack.com/p/the-vasty-deep-we-begin-anew @cekahn.bsky.social @rsnasky.bsky.social #blog #radiology #MachineLearning
New blog post! Pouria Rouzrokh discusses the implications of prompt injection attacks for peer review https://radiologyai.substack.com/p/ai-manipulation-of-the-peer-review #AI #ML #MachineLearning
Looking forward to presenting information about the #ROADMAP initiative at the upcoming #EuSoMII25 meeting! eusomii-annualmeeting2025.com @myesr.bsky.social
As AI in radiology transitions to #foundation models, dependence on expert annotations to prepare “AI-ready” datasets continues to evolve https://doi.org/10.1148/ryai.240631 @monganmd.bsky.social #LLMs #DeepLearning #MachineLearning
Looking forward to presenting information about the #ROADMAP initiative at the upcoming #EuSoMII25 meeting! eusomii-annualmeeting2025.com @myesr.bsky.social
I'm immensely grateful to our Editorial Board, reviewers, authors, and readers for making @radiology-ai.bsky.social a success! 🌟Many thanks to the superb @rsnasky.bsky.social staff -- and a special shout-out to our Deputy Editors: Drs. Aboian, Hsu, Kalpathy-Cramer, Kontos, & Sebro 🌟
🚨@radiology-ai.bsky.social Impact Factor = 13.2 # 3⃣ of 212 journals in #Radiology #NucMed #imaging # 9⃣ of 204 journals in #ComputerScience / #AI @rsnasky.bsky.social
I'm immensely grateful to our Editorial Board, reviewers, authors, and readers for making @radiology-ai.bsky.social a success! 🌟Many thanks to the superb @rsnasky.bsky.social staff -- and a special shout-out to our Deputy Editors: Drs. Aboian, Hsu, Kalpathy-Cramer, Kontos, & Sebro 🌟
🚨@radiology-ai.bsky.social Impact Factor = 13.2 # 3⃣ of 212 journals in #Radiology #NucMed #imaging # 9⃣ of 204 journals in #ComputerScience / #AI @rsnasky.bsky.social
Follow the journal's publications, blog posts, and other activities on Bluesky: https://bsky.app/profile/radiology-ai.bsky.social https://pubs.rsna.org/page/ai/blog/our_presence_on_bluesky @radiology-ai.bsky.social @cekahn.bsky.social #Radsky #ML #MachineLearning
#Dafne, a free, decentralized, collaborative #DeepLearning system for semantic segmentation of images through federated incremental learning https://doi.org/10.1148/ryai.240097 #ASCO25 #cancer #OncoRad #radiology
#Dafne, a free, decentralized, collaborative #DeepLearning system for semantic segmentation of images through federated incremental learning https://doi.org/10.1148/ryai.240097 #ASCO25 #cancer #OncoRad #radiology
Delighted to receive the Editor's Recognition Award with special distinction for @radiology-ai.bsky.social Special thanks to @cekahn.bsky.social @judywawira.bsky.social @merelhuisman.bsky.social @KMagudia and @radrudie.bsky.social Congratulations to all my fellow TEB members who were also awarded
Sham-AI model acting as a placebo control for intracranial aneurysm diagnosis to mitigate biases https://doi.org/10.1148/ryai.240140 #ShamAI #DeepLearning #ML
Special report on #cybersecurity threats and mitigation strategies for #LLMs in health care https://doi.org/10.1148/ryai.240739 @alitejanimd.bsky.social #ASNR25 #NeuroRad #radiology
Another example of opportunistic A.I. applied to medical imaging. @ai.nejm.org Quantifying coronary artery calcium score from chest CT (nongated, not performed for this purpose) ai.nejm.org/doi/full/10....
AI Opportunistic Coronary Calcium Screening at Veterans Affairs Hospitals
Coronary artery calcium (CAC) is highly predictive of cardiovascular events. Although millions of chest computed tomography (CT) scans are performed annually in the United States, CAC is not routin...
ai.nejm.org