Developing a Manually Annotated Corpus of Clinical Letters for Breast Cancer Patients on Routine Follow-Up.

Journal: Studies in health technology and informatics
Published Date:

Abstract

This paper introduces the annotation schema and annotation process for a corpus of clinical letters describing the disease course and treatment of oestrogen receptor positive breast cancer patients, after completion of primary surgery and radiotherapy treatment. Concepts related to therapy, clinical signs, and recurrence, as well as relationships linking these, are identified and annotated in 200 letters. This corpus will provide the basis for development of natural language processing tools for automatic extraction of key clinical factors from such letters.

Authors

  • Graham Pitson
    Barwon Health, Geelong VIC Australia.
  • Patricia Banks
    Peter MacCallum Cancer Centre, Melbourne VIC Australia.
  • Lawrence Cavedon
    School of Science, RMIT University, Melbourne, Australia.
  • Karin Verspoor
    Dept of Computing and Information Systems, School of Engineering, University of Melbourne, Melbourne, Australia.