Free Text Radiology

Free-text radiology research focuses on extracting structured information from unstructured radiology reports to improve efficiency and accuracy in medical image analysis. Current efforts leverage large language models (LLMs) like BERT and T5, often combined with convolutional neural networks (CNNs) for image analysis, employing techniques such as prompt learning, self-supervised learning, and knowledge graph construction to enhance information extraction and report generation. This work aims to automate tasks like abnormality detection, structured report generation, and clinical decision support, ultimately improving diagnostic accuracy and workflow efficiency in radiology.