Scalable NLP for Phenotyping of Single Ventricle Physiology from Cardiovascular Imaging Reports

Clicks: 3
ID: 340392
2026
Article Quality & Performance Metrics
Overall Quality
Not rated
Combines reader engagement with the AI quality analysis. This article has not been analysed, so there is no overall score — reader engagement is measured and shown alongside.
AI Quality Assessment
Not analyzed
Readership in this journal
Emerging

Ranked #85 of 87 articles by views in European Heart Journal - Digital Health

Most read Least read

Bar heights use a square-root scale.

Mint this article as an NFT
Not yet minted

Create a permanent, verifiable on-chain record of this article on the Scimatic Network. The NFT is held in your Journament account, and you can withdraw it to your own wallet at any time.

5 SUSD one-off · no wallet required
Abstract
Abstract Aims Single-ventricle physiology (SVP) is a high-risk CHD phenotype that is incompletely captured by diagnostic codes, limiting scalable EHR-based research. We developed and evaluated a natural language processing (NLP) framework using cardiovascular imaging reports for automated identification of patients with a longitudinal history of SVP across CHD-enriched and broader healthcare-system cohorts, benchmarking performance against ICD-based classification and a physician-validated rule-based algorithm. Methods and Results Four distinct cohorts of patients receiving care at UCLA Healthcare System between 2013 and 2023 were constructed: a development cohort (N = 1,020), a fully adjudicated primary CHD test cohort (N = 2,500), a secondary CHD screening cohort (N = 22,500), and a general UCLA EHR screening cohort (N = 58,774). Term frequency–inverse document frequency (TF–IDF) and Word2Vec representations with extreme gradient boosting were compared with ICD-based and rule-based phenotyping. TF–IDF models achieved AUROC values up to 0.99 and generally retained higher AUPRC, precision, sensitivity, and F1 scores than Word2Vec as SVP prevalence declined. Performance exceeded the ICD-only classification and approached that of the expert-derived rule-based algorithm. Conclusion Within a single academic healthcare system, NLP of cardiovascular imaging reports supported accurate and scalable retrospective identification of patients with a history of SVP across CHD-enriched and general EHR cohorts. Multi-center external validation is required before clinical deployment and cross-system testing is needed to establish generalizability.
Reference Key
openalex_W7234937937 Use this key to autocite in the manuscript while using SciMatic Manuscript Manager or Thesis Manager
Authors
Journal European Heart Journal - Digital Health
Year 2026
DOI
10.1093/ehjdh/ztag182
URL
Keywords Keywords not found

Citations

No citations found. To add a citation, contact the admin at info@scimatic.org

No comments yet. Be the first to comment on this article.