Paying Per-label Attention for Multi-label Extraction from Radiology Reports
This addresses the problem of expensive expert annotation for medical image analysis by enabling efficient label extraction from radiology reports, though it is incremental as it extends existing models.
The paper tackles automated extraction of structured labels from head CT reports for stroke patients using deep learning, achieving robust extraction of 31 labels with a single model through a label-dependent attention mechanism and synthetic data augmentation.
Training medical image analysis models requires large amounts of expertly annotated data which is time-consuming and expensive to obtain. Images are often accompanied by free-text radiology reports which are a rich source of information. In this paper, we tackle the automated extraction of structured labels from head CT reports for imaging of suspected stroke patients, using deep learning. Firstly, we propose a set of 31 labels which correspond to radiographic findings (e.g. hyperdensity) and clinical impressions (e.g. haemorrhage) related to neurological abnormalities. Secondly, inspired by previous work, we extend existing state-of-the-art neural network models with a label-dependent attention mechanism. Using this mechanism and simple synthetic data augmentation, we are able to robustly extract many labels with a single model, classified according to the radiologist's reporting (positive, uncertain, negative). This approach can be used in further research to effectively extract many labels from medical text.