VIZYON FRANCE

Je serai au colloque organisé par l'AUEG et MEDICALPS sur l'intelligence artificielle dans le domaine de la santé. Je participerai à la table ronde de 16h10 sur le sujet "Comment les industriels participent au progrès de l’exercice médical avec l’IA ? " pour partager l'expérience de VIZYON FRANCE dans l'intégration de l'IA dans le flux de téléradiologie et les prochains axes d'évolution

9 products from 8 AI vendors 2 use cases 14 datasets from 10 different centers +30 radiologist 'readers' reading a total of 1470 studies. Four years ago there were still very little studies on commercial AI products for radiology. While the evidence base has significantly grown the past years, it remains hard to compare similar products as different study protocols and datasets make it a comparison between apples and pears. With Project AIR we compare apples with apples on the level of stand-alone performance. And I know. There is more to it. Real-life situations may be somewhat different. And yes other aspects than performance are also important. But I believe this is a good start on further increasing transparency in the field and pushing the bar for vendors to keep improving their products. When we started many people responded skeptical: will vendors do this when the results will not be anonymous? A big thanks to the vendors who trusted us with their algorithms and trusted us in doing this in a fair and transparent way. Visiana, VUNO Inc., Annalise.ai, Infervision, Lunit Global, MILVUE, oxipit, Siemens Healthineers. To all the co-authors, your help has been invaluable, thank you! Steven Schalekamp, Matthieu Rutten, Merel Huisman, MD, PhD, Cornelia Schaefer-Prokop, Maarten de Rooij, Bram van Ginneken. I don't think Linkedin allows to tag that many people, so I am listing the Project AIR Working Group in the comments. Project AIR was, or I should say is, a project by many. Because we will continue! Taking over the honors from me is Marlie Besouw who will start her PhD journey at the Radboudumc soon. To be continued... Leaderboard for bone age prediction: https://lnkd.in/eCcF_HMA Leaderboard for lung nodule detection: https://lnkd.in/eNuxcTPy Full paper: https://lnkd.in/eTJkDX3d Editorial by Radiology: https://lnkd.in/eZktZVsB Radboudumc wetenschap, #AI #radiology #medtech #validation