Related Experiment Video
Updated: May 10, 2026
07:50
A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
Published on: September 20, 2018
Extracting BI-RADS Features from Portuguese Clinical Texts
Houssam Nassif1, Filipe Cunha, Inês C Moreira
1University of Wisconsin, Madison, USA ( nassif@wisc.edu ).
Proceedings. IEEE International Conference on Bioinformatics and Biomedicine
|June 26, 2013
Summary
We developed a new tool to automatically extract Breast Imaging Reporting and Data System (BI-RADS) information from Portuguese mammography reports. This automated parser shows performance comparable to human experts.
Area of Science:
- Medical Informatics
- Natural Language Processing
- Radiology
Background:
- The Breast Imaging Reporting and Data System (BI-RADS) is crucial for standardizing mammography interpretations.
- Extracting BI-RADS features from free-text reports is challenging, especially in languages other than English.
Purpose of the Study:
- To develop and evaluate the first automated BI-RADS parser for Portuguese free-text mammography reports.
- To assess the performance of the developed parser against manual annotation by a mammography specialist.
Main Methods:
- Utilized a semantic grammar approach based on the BI-RADS lexicon.
- Incorporated iterative transferred expert knowledge for concept identification.
- Compared the automated parser's output with manual annotations from a specialist.
Main Results:
- The developed BI-RADS parser demonstrated performance comparable to manual annotation by a specialist.
- The system successfully extracts key BI-RADS features from Portuguese free texts.
Conclusions:
- Automated BI-RADS feature extraction from Portuguese mammography reports is feasible and accurate.
- This tool has the potential to improve efficiency and consistency in mammography reporting and analysis.
