Clinically Significant Information Extraction from Radiology Reports

Abstract

Radiology reports are one of the most important medical documents that a diagnostician looks into, especially in the emergency situations. They provide the emergency physicians with critical information regarding the condition of the patient and help the physicians take immediate action on urgent conditions. However, the reports are complex and unstructured.

We developed a machine learning system to efficiently extract the clinically significant parts and their level of importance in radiology reports. The system also classifies the overall report into critical or non-critical which help radiologists in identifying potential high priority reports. As a starting point, the system uses Chest X-RAY reports of adults (de-identified) and provides the doctors with 3 levels of medical phrases namely high-critical conditions, critical conditions and non-critical conditions. We used Conditional Random Field to identify clinically significant phrases with an average F1-score of 0.75.

The CRF Model is used as a filter with the web interface which highlights the medical phrases and their criticality level to the emergency physician. The overall classification of the report is identified using Stochastic Gradient Descent and features used are phrases extracted from the CRF model which provides an average accuracy of 0.85.