During the 109th Scientific Assembly and Annual Meeting of RSNA 2023, Fang organized and hosted an educational session titled “Deep Learning in MRI“. The session featured two speakers who spoke about recent research and clinical updates on AI in MRI. Dr. Susie Huang from MGH/HST Martinos Center discussed the use of AI in rapid MRI, while Dr. Richard Kijowski from NYU Langone talked about the applications of AI in musculoskeletal MRI.
I3T lab is also pleased to introduce RadGPT, a new vision-languageĀ AIĀ LargeLanguageModelĀ (LLM), that enables human-AI interactive radiological diagnosis via image feature learning fromĀ VisionTransformerĀ and language reasoning from locally hostedĀ GPT. We showed the first RadGPT application inĀ radiologyĀ image diagnosis, interpretation, and automated medical report generation through expert-level human-AI conversation. We also developed and demonstrated a comprehensive set of quantitative metrics for medical text assessment focusing on medical keyword accuracy, sentence correctness, conversation concordance, diagnostic insight, and response uniformity. We are Looking forward to seeing RadGPT for large-scale studies with extensive testing and validation.Ā
