EUVIMEDEUVIMED entdecken
Crossref

Language model-based labeling of German thoracic radiology reports

Alessandro Wollek iD, Philip Haitzer, Thomas Sedlmeyr, Sardi Hyska, Johannes Rueckel, Bastian O Sabel, Michael Ingrisch iD, Tobias Lasser

RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren · 01.01.2025 · 197 (01) · 55-64

DOI: 10.1055/a-2287-5054 ↗

Kurz gefasst

Fachartikel in der Zeitschrift RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, erschienen am 01.01.2025. Verfasst von Alessandro Wollek und 7 weiteren Autor:innen. Beteiligte Einrichtung: Munich Institute of Biomedical Engineering, Technical University of Munich, Garching near Munich, Germany. Hauptthemen laut MEDLINE: Radiography, Thoracic, Deep Learning. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 25 Literaturangaben. Gefördert von: Bundesministerium für Gesundheit.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The aim of this study was to explore the potential of weak supervision in a deep learning-based label prediction model. The goal was to use this model to extract labels from German free-text thoracic radiology reports on chest X-ray images and for training chest X-ray classification models. The proposed label extraction model for German thoracic radiology reports uses a German BERT encoder as a backbone and classifies a report based on the CheXpert labels. For investigating the efficient use of manually annotated data, the model was trained using manual annotations, weak rule-based labels, and both. Rule-based labels were extracted from 66071 retrospectively collected radiology reports from 2017–2021 (DS 0), and 1091 reports from 2020–2021 (DS 1) were manually labeled according to the CheXpert classes. Label extraction performance was evaluated with respect to mention extraction, negation detection, and uncertainty detection by measuring F1 scores. The influence of the label extraction method on chest X-ray classification was evaluated on a pneumothorax data set (DS 2) containing 6434 chest radiographs with associated reports and expert diagnoses of pneumothorax. For this, DenseNet-121 models trained on manual annotations, rule-based and deep learning-based label predictions, and publicly available data were compared. The proposed deep learning-based labeler (DL) performed on average considerably stronger than the rule-based labeler (RB) for all three tasks on DS 1 with F1 scores of 0.938 vs. 0.844 for mention extraction, 0.891 vs. 0.821 for negation detection, and 0.624 vs. 0.518 for uncertainty detection. Pre-training on DS 0 and fine-tuning on DS 1 performed better than only training on either DS 0 or DS 1. Chest X-ray pneumothorax classification results (DS 2) were highest when trained with DL labels with an area under the receiver operating curve (AUC) of 0.939 compared to RB labels with an AUC of 0.858. Training with manual labels performed slightly worse than training with DL labels with an AUC of 0.934. In contrast, training with a public data set resulted in an AUC of 0.720. Our results show that leveraging a rule-based report labeler for weak supervision leads to improved labeling performance. The pneumothorax classification results demonstrate that our proposed deep learning-based labeler can serve as a substitute for manual labeling requiring only 1000 manually annotated reports for training. Wollek A, Haitzer P, Sedlmeyr T et al. Language modelbased labeling of German thoracic radiology reports. Fortschr Röntgenstr 2024; DOI 10.1055/a-2287-5054

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Radiography, ThoracicDeep LearningGermanyHumansRetrospective StudiesNatural Language ProcessingPneumothorax

Autor:innen und Institutionen

  • Munich Institute of Biomedical Engineering, Technical University of Munich, Garching near Munich, Germany.
  • School of Computation, Information and Technology, Technical University of Munich, Garching near Munich, Germany.
  • Department of Radiology, Ludwig-Maximilians-University Hospital Munich, Germany, Munich, Germany.
  • Institute of Neuroradiology, Ludwig-Maximilians-University Hospital Munich, Munich, Germany.

ORCID-Profile: Alessandro Wollek, Michael Ingrisch

Förderung

  • Bundesministerium für Gesundheit

Zitieren

APA

Wollek, A., Haitzer, P., Sedlmeyr, T., Hyska, S., Rueckel, J., Sabel, B. O., Ingrisch, M., & Lasser, T. (2025). Language model-based labeling of German thoracic radiology reports. RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, 197(01), 55-64. https://doi.org/10.1055/a-2287-5054

Vancouver

Wollek A, Haitzer P, Sedlmeyr T, Hyska S, Rueckel J, Sabel BO, et al. Language model-based labeling of German thoracic radiology reports. RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren. 2025;197(01):55-64. doi:10.1055/a-2287-5054

BibTeX
@article{Wollek2025,
  author = {Wollek, Alessandro and Haitzer, Philip and Sedlmeyr, Thomas and Hyska, Sardi and Rueckel, Johannes and Sabel, Bastian O and Ingrisch, Michael and Lasser, Tobias},
  title = {Language model-based labeling of German thoracic radiology reports},
  journal = {RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren},
  year = {2025},
  volume = {197},
  number = {01},
  pages = {55--64},
  doi = {10.1055/a-2287-5054}
}
Kostenlos starten →