
France's national radiology society (Société Française de Radiologie, SFR) has published guidelines on the use of artificial intelligence (AI) in clinical practice. They cover examination planning, access to patient data, optimization of imaging technique, image quality, detection and characterization of lesions, monitoring, prognosis, and communication of results.
Ethics, patient and data safety, and adherence to regulatory standards remain paramount throughout, according to a statement posted on the SFR website on 2 December. AI tools should be integrated into clinical practice in an adapted and pertinent manner with their limits and systematic bias known, the SFR noted, adding that all AI tools must be developed for future use that is in line with legal criteria for medical practice.
The SFR stated that the tools must not be a substitute for radiologists, nor shoulder regional demographic problems, nor allow the acquisition of imaging equipment without there being a team of radiologists behind a defined local project.
The guidelines are aimed at suggesting rules and good practice for the use of AI tools and at highlighting currently underestimated key elements that can directly or indirectly have negative or unforeseen consequences when AI algorithms are deployed in radiology and emergency imaging, the SFR noted. The six guidelines straddle major themes including technical and regulatory prerequisites, stakeholders in conclusions drawn from AI, and traceability and surveillance.












![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnieeurope.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)






