ESC TV Today – Your Cardiovascular News Season 3 - Ep.27: 'ChatGPT, MD?': large language models at the bedside - Management decisions in myocarditis
5 snips
Nov 20, 2025 Folkert Asselbergs, a cardiologist and expert in digital health and AI, discusses the implications of large language models (LLMs) in clinical settings, including their role in patient engagement and reducing clinician workload. Massimo Imazio, chair of the upcoming ESC guidelines on myocarditis, delves into the nuances of diagnosing myocarditis, emphasizing cardiac MRI and biopsies when necessary. He also provides critical insights on managing immune-related myocarditis and guidelines for returning patients to physical activity.
AI Snips
Chapters
Transcript
Episode notes
Protect Patient Data From Public LLMs
- Avoid pasting identifiable patient data into public LLMs because those inputs are stored on external servers.
- Build LLM integrations within your EHR in a closed, secure environment to protect patient privacy.
Domain LLMs Deliver Traceable Guideline Answers
- Domain-specific LLMs trained on guideline texts can provide traceable, guideline-based answers.
- The ESC chat links answers to source PDFs and highlights the exact phrases used for transparency.
AI Can Match Empathy And Augment Care
- Patient evaluations rate AI chatbots as at least as empathic as physicians in studies when assessed by patients.
- A hybrid model combining in-person care with 24/7 AI avatars could democratize information for patients and caregivers.