Automatic and safe extraction of radiology data

Use case

Automatic extraction of useful information from radiology reports

Opportunity

Speed up technological innovation in healthcare to ensure efficient infrastructure and services.

Challenge

Automatically extract information from radiology reports (patient details, medical history, clinical observations from instrumental tests, visit results, and clinician comments) using Natural Language Processing (NLP) techniques, while ensuring:

• data anonymization
• classification of extracted entities according to standard categories
• replicability across multiple clinical settings.

Solution

The solutions make it possible to automatically identify and extract information from radiology reports: using Natural Language Processing techniques, the information is annotated on pseudonymized reports, and the data being searched can be quickly and automatically retrieved, ensuring patient privacy.