Automatic processing of clinical reports

Automatic processing of clinical reports applying deep learning and NLP

There is a large amount of extremely valuable information stored in clinical notes written in plain unstructured text. This project aims to develop NLP tools for automated treatment of clinical reports, structuring the text into medical terminologies such as ICD-10 or SNOMED-CT. UCA Datalab collaborates with the Hospital Puerta del Mar in Cádiz, where Dr. García Rojo, Head of Pathology, provides us with more than 160K clinical notes on biopsies, anotated in SNOMED, in order to train NLP algorithms. The possible outcomes of this project are very rich and include efficient cohort segmentation for clinical trials, information retrieval and faster, reliable coding of structured reports during clinical visits.

Carlos III Health Research Institute
(Healthcare Technical Development)
Team: 9 researchers (UCA-INIBICA)
Duration: oct 2020 – oct 2022
Quantity: 47.850 €
IPs: Marcial Gª Rojo – D. Gómez-Ullate
Ref: DTS20/00027