Performance of Open-Source LLMs in Identifying Pediatric Pneumonia From Free-Text Chest Radiograph Reports

Key Points

• The aim is to evaluate the effectiveness of open-source LLMs in classifying pediatric pneumonia from chest radiograph reports.
• Analyzed free-text chest radiograph reports using open-source LLMs for classification.
• Evaluated accuracy in identifying pneumonia cases among pediatric patients.
• Open-source LLMs exhibited high accuracy in classifying pediatric pneumonia cases from CXR reports.
• Findings suggest viable integration of LLMs into clinical decision support systems.