Comparing Large Language Model and Human Reader Accuracy with <i>New England Journal of Medicine</i> Image Challenge Case Image Inputs
Pae Sun Suh, Woo Hyun Shim, Chong Hyun Suh, Hwon Heo, Kye Jin Park, Pyeong Hwa Kim, Se Jin Choi, Yura Ahn, Sohee Park, Ho Young Park, Na Eun Oh, Min Woo Han, Sung Tan Cho, Chang-Yun Woo, Hyung Park
Abstract
< .001). Text input length affected LLM accuracy (odds ratio range, 3.2 [95% CI: 1.9, 5.5] to 6.6 [95% CI: 3.7, 12.0]). Conclusion LLMs demonstrated substantial accuracy with text and image inputs, outperforming a medical student. However, their accuracy decreased with shorter text lengths, regardless of image input. © RSNA, 2024
Topics & Concepts
MedicineImage (mathematics)New englandMedical journalEnglish languageArtificial intelligenceLinguisticsFamily medicineComputer sciencePoliticsPhilosophyLawPolitical scienceArtificial Intelligence in Healthcare and EducationRadiomics and Machine Learning in Medical ImagingRadiology practices and education