Natural Language Processing (NLP) and Sudden Cardiac Death (SCD)

Status: In progress

Partners: Oslo Universitetssykehus, Akershus University Hospital, University of Oslo, Kunnskapsforlaget

Tablet with medical data

Sudden Cardiac Death (SCD) accounts for approximately 5,000 deaths yearly in Norway. High-risk individuals when identified can be offered life-saving therapy. Selection of patients for this therapy and prediction of SCD is one of the greatest challenges in current cardiology. Use of patient-specific genetic data combined with clinical data is instrumental in risk stratification for SCD.

Activities and deliverables are described below.  

  • SCD Risk Calculator implemented in DIPS Arena 
  • Automatic echo measurements for input to calculator (machine learning) 
  • Natural language processing (NLP) projects: 
    • identification of patients at risk for SCD from electronic medical journal, identification of “syncope” to populate risk calculator. 
    • Interpretation of CT descriptions 
    • Pedigree tool: Extraction from free text - family relations relevant for medical condition
  • Pipeline for text extraction from electronic medical journal 

Figure1. View of the SDS risk Calculator in DIPS.

Figure 2. Illustration of automatic echo measurement.

Figure 3. Pedigree tool for visualisation of family relations.


Partners: OUS cardiology, Ahus, NTNU, UiO IFI Language Technology, Kunnskapsforlaget

Pål Brekke

OUS

+47 467 47 790

Send email