Diagnosis extraction from unstructured Dutch echocardiogram reports using span- and document-level characteristic classification
We recommend using our published SpanCategorizer and MedRoBERTa.nl models for span- and document-level diagnosis extraction from Dutch echocardiography reports. For settings with limited training data, SetFit may be a promising alternative for document classification. Future research should be aimed at training a RoBERTa based span classifier and applying English based models on translated echocardiogram reports.
Authors
- Pim van der Harst (ORCID: https://orcid.org/0000-0002-2713-686X)
- René van Es (ORCID: https://orcid.org/0000-0001-9950-4388)
- Bauke Arends (ORCID: https://orcid.org/0009-0009-7494-4542)
- Melle Vessies
- Arco J. Teske (ORCID: https://orcid.org/0000-0001-5305-1619)
- Dirk van Osch
- Bram van Es (ORCID: https://orcid.org/0000-0003-4667-1304)
Institutions
- University Medical Center Utrecht (NL)
Publication Details
- Journal
- BMC Medical Informatics and Decision Making
- Published
- 2025-03-07
- DOI
- https://doi.org/10.1186/s12911-025-02897-w
- Citations
- 1
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 2.48