Why the study?
Does an LLM (ChatGPT) accurately generate echocardiography reports and clinical recommendations compared to standard clinical assessments?
Population
n=21 echocardiographic cases (13 fictional, 8 clinical)
Comparison
Large language model for automated generation of… vs Standard clinical assessments conducted by…
Design
Other
Key result
ChatGPT generated fully acceptable echocardiography reports in 85.7% of cases, with a mean total score of 6.86 and only 5.3% of parameters misinterpreted compared to expert cardiologists.
Authors
Loading...
Should not yet change reporting practices; hypothesis-generating for LLM use in echocardiography.
Cross-Sectional (n=21)
Does an LLM (ChatGPT) accurately generate echocardiography reports and clinical recommendations compared to standard clinical assessments?
ChatGPT demonstrates high accuracy in generating echocardiography reports and clinical recommendations, suggesting potential utility in streamlining clinical workflows.
Syryca et al. (2025) conducted a cross-sectional in Cardiovascular diseases (n=21). ChatGPT (Large Language Model) vs. Standard clinical assessments by experienced cardiologists was evaluated on Mean total score for report generation, diagnostic precision, and recommendations. ChatGPT generated fully acceptable echocardiography reports in 85.7% of cases, with a mean total score of 6.86 and only 5.3% of parameters misinterpreted compared to expert cardiologists.