A new study finds that GPT-4o still faces significant hurdles in accurately interpreting medical images despite promising capabilities.
Key Details
- 1Researchers evaluated GPT-4o on 377 imaging cases across X-ray, CT, and MRI.
- 2The model did not receive clinical context or prior imaging for analysis.
- 3Three radiologists rated GPT-4o's responses using a 5-point scale.
- 4GPT-4o showed high accuracy in some instances but inconsistent, 'all or nothing' results in others.
- 5Potential applications include improving radiology workflows and expanding access to care in rural settings.
Why It Matters
The inconsistent performance of leading large language models like GPT-4o in medical imaging highlights both the potential value and current limitations of applying general AI models in radiology. Progress in this area could significantly impact radiology workflows and help mitigate specialist shortages, but further development and validation are clearly needed.
Source
Health Imaging
Related News
•AuntMinnie
Multimodal LLMs Show Improved Performance on Japanese Radiology Board Exams
New multimodal large language models (LLMs) like OpenAI o3 and Gemini 2.5 Pro demonstrated significant advancements in answering Japanese radiology board exam questions, particularly with image input.
•Health Imaging
AI Surpasses Radiologists in Predicting Lung Cancer Treatment Response
AI demonstrates higher accuracy than radiologists in predicting lung cancer treatment response from imaging.
•Radiology Business
Medicare Denies Over Half of Radiology AI Claims, Study Finds
Medicare denied $16 million in radiology AI claims between 2018 and 2023, rejecting over half submitted.