Large language models for structuring knee ultrasound reports-A comparative evaluation of DeepSeek R1, gemini 2.5 flash, and GPT-4o

Key Points

• This evaluation aims to compare the effectiveness of large language models in structuring knee ultrasound reports.
• Comparative evaluation of GPT-4o, DeepSeek R1, and Gemini 2.5 Flash.
• Assessment of report consistency and clinical utility across models.
• GPT-4o consistently outperformed both DeepSeek R1 and Gemini 2.5 Flash.
• Improved consistency in knee ultrasound reporting was observed with GPT-4o.