Medical Imaging
Study finds radiologists and AI models struggle to spot AI-generated “deepfake” X-rays
Raportoinut AI AI:n luoma kuva Faktatarkistettu
A study published March 24, 2026 in *Radiology* reports that AI-generated “deepfake” X-rays can be convincing enough to mislead radiologists and several multimodal AI systems. In testing, radiologists’ average accuracy rose from 41% when they were not told fakes were included to 75% when they were warned, highlighting potential risks for medical imaging security and clinical decision-making.
University of Missouri researchers report that a small antibody fragment targeting the EphA2 protein can be tagged with a radioactive marker to make EphA2-positive tumors stand out on PET scans in mouse experiments, a step they say could help match patients to EphA2-targeted therapies.
Raportoinut AI
Researchers at the University of Michigan have developed an AI system called Prima that interprets brain MRI scans in seconds, identifying neurological conditions with up to 97.5% accuracy. The tool also flags urgent cases like strokes and brain hemorrhages, potentially speeding up medical responses. Findings from the study appear in Nature Biomedical Engineering.