EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Explainable Multitask Transformers for Early Detection of Smoking Behaviors and Lung Cancer Symptoms from Danish Electronic Health Records

Amir Sorayaie Azar, Uffe Kock Wiil, Margrethe Bang Høstgaard Henriksen, Jamshid Bagherzadeh Mohasefi, Ole Hilberg, Amin Naemi, Ali Ebrahimi

Studies in Health Technology and Informatics · 21.05.2026

DOI: 10.3233/shti260310 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Studies in Health Technology and Informatics, erschienen am 21.05.2026. Verfasst von Amir Sorayaie Azar und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Engineering, Urmia University, Urmia, Iran. Hauptthemen laut MEDLINE: Lung Neoplasms, Electronic Health Records, Early Detection of Cancer, Smoking. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Lung cancer is the leading cause of cancer-related mortality, and early detection is challenging due to asymptomatic onset. Accurate detection of smoking behaviors and symptoms is essential for risk evaluation and timely intervention. This study developed multitask transformer-based models to simultaneously detect smoking status, pack-year, and lung cancer symptom categories from Danish electronic health records. Texts were annotated for smoking behaviors, and 16 lung cancer–related symptoms were grouped into six categories. Preprocessing and tokenization were performed, and three transformer models, including BioBERT, XLM-RoBERTa, and Danish BERT, were trained in a multitask framework. Model performance was evaluated, and LIME was applied for explainability. BioBERT performed best for smoking status (Accuracy: 97.93%, F1-score: 86.30%), while XLM-RoBERTa excelled in pack-year (Accuracy: 96.23%, F1-score: 92.02%) and symptom detection (Accuracy: 99.55%, F1-score: 96.67%). LIME highlighted key predictive features, enhancing transparency. These results demonstrate that multitask transformer models can support early risk assessment and improved clinical decision-making, with future work applying large language models.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Lung NeoplasmsElectronic Health RecordsEarly Detection of CancerSmokingNatural Language ProcessingHumansDenmark

Stichwörter der Autor:innen

DetectionExplainable multitask transformersLung cancerSmoking behaviorSymptoms

Autor:innen und Institutionen

  • Department of Computer Engineering, Urmia University, Urmia, Iran.
  • SDU Health Informatics and Technology, The Maersk Mc-Kinney Moller Institute, University of Southern Denmark, Odense, Denmark.
  • Department of Oncology, Lillebaelt Hospital, University Hospital of Southern Denmark, Vejle, Denmark.
  • Department of Internal Medicine, Lillebaelt Hospital, University Hospital of Southern Denmark, Vejle, Denmark.
  • Institute of Regional Health Research, University of Southern Denmark, Odense, Denmark.
  • Nordcee, Department of Biology, University of Southern Denmark, Odense, Denmark.

Zitieren

APA

Sorayaie Azar, A., Wiil, U. K., Henriksen, M. B. H., Bagherzadeh Mohasefi, J., Hilberg, O., Naemi, A., & Ebrahimi, A. (2026). Explainable Multitask Transformers for Early Detection of Smoking Behaviors and Lung Cancer Symptoms from Danish Electronic Health Records. Studies in Health Technology and Informatics. https://doi.org/10.3233/shti260310

Vancouver

Sorayaie Azar A, Wiil UK, Henriksen MBH, Bagherzadeh Mohasefi J, Hilberg O, Naemi A, et al. Explainable Multitask Transformers for Early Detection of Smoking Behaviors and Lung Cancer Symptoms from Danish Electronic Health Records. Studies in Health Technology and Informatics. 2026. doi:10.3233/shti260310

BibTeX
@article{SorayaieAzar2026,
  author = {Sorayaie Azar, Amir and Wiil, Uffe Kock and Henriksen, Margrethe Bang Høstgaard and Bagherzadeh Mohasefi, Jamshid and Hilberg, Ole and Naemi, Amin and Ebrahimi, Ali},
  title = {Explainable Multitask Transformers for Early Detection of Smoking Behaviors and Lung Cancer Symptoms from Danish Electronic Health Records},
  journal = {Studies in Health Technology and Informatics},
  year = {2026},
  doi = {10.3233/shti260310}
}
Kostenlos starten →