CoChat
Sign Up
Cited by 8Topic ModelingMultimodal Machine Learning ApplicationsArtificial Intelligence in Healthcare and Education

The DOI matches its Crossref record; 31 of the 31 references it cites were found in a registry or an index.

Fine-tuning BERT Models for Summarizing German Radiology Findings

Open access
Liang, Siting, Kades, Klaus, Fink, Matthias, Full, Peter, Weber, Tim, Kleesiek, Jens, Strube, Michael, Maier-Hein, Klaus·2022·Proceedings of the 4th Clinical Natural Language Processing Workshop
DOI 10.18653/v1/2022.clinicalnlp-1.4
Open this paper in CoChat — free accountOpen PDF

Abstract

Writing the conclusion section of radiology reports is essential for communicating the radiology findings and its assessment to physician in a condensed form.In this work, we employ a transformer-based Seq2Seq model for generating the conclusion section of German radiology reports.The model is initialized with the pretrained parameters of a German BERT model and fine-tuned in our downstream task on our domain data.We proposed two strategies to improve the factual correctness of the model.In the first method, next to the abstractive learning objective, we introduce an extraction learning objective to train the decoder in the model to both generate one summary sequence and extract the key findings from the source input.The second approach is to integrate the pointer mechanism into the transformer-based Seq2Seq model.The pointer network helps the Seq2Seq model to choose between generating tokens from the vocabulary or copying parts from the source input during generation.The results of the automatic and human evaluations show that the enhanced Seq2Seq model is capable of generating human-like radiology conclusions and that the improved models effectively reduce the factual errors in the generations despite the small amount of training data.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.18653/v1/2022.clinicalnlp-1.4

Cites

31 of 31 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Standardised Layout of SOR (Weber et al., 2020). Each report has a uniform organization: the general section expresses background information on imaging and clinical data, the next section (Findings) describes oncology and nononcology findings, and the Conclusion section gives oncological and non-oncological impressions.
Figure 1 Standardised Layout of SOR (Weber et al., 2020). Each report has a uniform organization: the general section expresses background information on imaging and clinical data, the next section (Findings) describes oncology and nononcology findings, and the Conclusion section gives oncological and non-oncological impressions.CC-BY · From the PDF
Figure 2 BERT2BERT model adding Extraction Loss. In order to train the decoder to extract the key findings through generation, we supply an additional target sequence ('Background + Key Findings'), which consists of the key findings selected from the source sequence.
Figure 2 BERT2BERT model adding Extraction Loss. In order to train the decoder to extract the key findings through generation, we supply an additional target sequence ('Background + Key Findings'), which consists of the key findings selected from the source sequence.CC-BY · From the PDF
Figure 3 BERT2BERT model incorporating the Pointer Mechanism.
Figure 3 BERT2BERT model incorporating the Pointer Mechanism.CC-BY · From the PDF
Figure 4 Number of reports for the three data partitions after matching to patient degree categories. We have significantly more reports in the Without Evidence and Stable Disease categories than in the other two categories, and the fewest reports are found in the Partial Response category.
Figure 4 Number of reports for the three data partitions after matching to patient degree categories. We have significantly more reports in the Without Evidence and Stable Disease categories than in the other two categories, and the fewest reports are found in the Partial Response category.CC-BY · From the PDF
Figure 5 Average scores with standard deviation for the three criteria: Oncological correctness, non-oncological correctness and comprehensibility.
Figure 5 Average scores with standard deviation for the three criteria: Oncological correctness, non-oncological correctness and comprehensibility.CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 31 references

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Liang, S., Kades, K., Fink, M., Full, P., Weber, T., Kleesiek, J., Strube, M., & Maier-Hein, K.
Affiliations
  • German Research Centre for Artificial Intelligence (DE)
  • German Cancer Research Center (DE)
  • University Hospital Heidelberg (DE)
  • Deutsches Konsortium für Translationale Krebsforschung (DE)
  • Heidelberg Institute for Theoretical Studies (DE)
ORCID
Venue
Proceedings of the 4th Clinical Natural Language Processing Workshop
Publisher
Association for Computational Linguistics
Published
2022
DOI
10.18653/v1/2022.clinicalnlp-1.4
OpenAlex
W4287888487
Cited by
8
References
31
Citations by year
2025: 1 · 2024: 3 · 2023: 4
FWCI
1.11
Type
conference paper
Language
en
Open access
Licence
cc-by
Locations
Clinical Natural Language Processing Workshop (ClinicalNLP) (publishedVersion, cc-by)
Funders
  • Deutsches Krebsforschungszentrum
  • Deutschen Konsortium für Translationale Krebsforschung
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
radiology report summarization, German language, BERT fine-tuning, sequence-to-sequence models, pointer networks, factual accuracy, abstractive summarization, clinical text generation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review