2025

1.
Borchert, F., Llorca, I., Roller, R., Arnrich, B., Schapranow, M.-P.: xMEN: A Modular Toolkit for Cross-Lingual Medical Entity Normalization. JAMIA Open. 8, ooae147 (2025).
[ BibTeX ] [ URL ]
 
@article{borchert2023xmen, author = {Borchert, Florian and Llorca, Ignacio and Roller, Roland and Arnrich, Bert and Schapranow, Matthieu-P.}, journal = {JAMIA Open}, keywords = {nlp sys:relevantfor:dhc gemtex fb-boettinger}, number = 1, pages = {ooae147}, title = {xMEN: A Modular Toolkit for Cross-Lingual Medical Entity Normalization}, volume = 8, year = 2025 }
2.
Borchert, F., Wullenweber, P., Oeser, A., Kreuzberger, N., Karge, T., Langer, T., Skoetz, N., Wieler, L.H., Schapranow, M.-P., Arnrich, B.: High-precision information retrieval for rapid clinical guideline updates. npj Digital Medicine. 8, 227 (2025).
[ BibTeX ] [ URL ]
 
@article{borchert2025nge, author = {Borchert, Florian and Wullenweber, Paul and Oeser, Annika and Kreuzberger, Nina and Karge, Torsten and Langer, Thomas and Skoetz, Nicole and Wieler, Lothar H. and Schapranow, Matthieu-P. and Arnrich, Bert}, journal = {npj Digital Medicine}, keywords = {nlp myown sys:relevantfor:dhc gemtex}, number = 1, pages = 227, title = {High-precision information retrieval for rapid clinical guideline updates}, volume = 8, year = 2025 }

2024

1.
Borchert, F., Llorca, I., Schapranow, M.-P.: Improving biomedical entity linking for complex entity mentions with LLM-based text simplification. Database. 2024, baae067 (2024).
[ Abstract ] [ BibTeX ] [ URL ]
 
Large amounts of important medical information are captured in free-text documents in biomedical research and within healthcare systems, which can be made accessible through natural language processing (NLP). A key component in most biomedical NLP pipelines is entity linking, i.e. grounding textual mentions of named entities to a reference of medical concepts, usually derived from a terminology system, such as the Systematized Nomenclature of Medicine Clinical Terms. However, complex entity mentions, spanning multiple tokens, are notoriously hard to normalize due to the difficulty of finding appropriate candidate concepts. In this work, we propose an approach to preprocess such mentions for candidate generation, building upon recent advances in text simplification with generative large language models. We evaluate the feasibility of our method in the context of the entity linking track of the BioCreative VIII SympTEMIST shared task. We find that instructing the latest Generative Pre-trained Transformer model with a few-shot prompt for text simplification results in mention spans that are easier to normalize. Thus, we can improve recall during candidate generation by 2.9 percentage points compared to our baseline system, which achieved the best score in the original shared task evaluation. Furthermore, we show that this improvement in recall can be fully translated into top-1 accuracy through careful initialization of a subsequent reranking model. Our best system achieves an accuracy of 63.6% on the SympTEMIST test set. The proposed approach has been integrated into the open-source xMEN toolkit, which is available online via https://github.com/hpi-dhc/xmen.
@article{10.1093/database/baae067, abstract = {{Large amounts of important medical information are captured in free-text documents in biomedical research and within healthcare systems, which can be made accessible through natural language processing (NLP). A key component in most biomedical NLP pipelines is entity linking, i.e. grounding textual mentions of named entities to a reference of medical concepts, usually derived from a terminology system, such as the Systematized Nomenclature of Medicine Clinical Terms. However, complex entity mentions, spanning multiple tokens, are notoriously hard to normalize due to the difficulty of finding appropriate candidate concepts. In this work, we propose an approach to preprocess such mentions for candidate generation, building upon recent advances in text simplification with generative large language models. We evaluate the feasibility of our method in the context of the entity linking track of the BioCreative VIII SympTEMIST shared task. We find that instructing the latest Generative Pre-trained Transformer model with a few-shot prompt for text simplification results in mention spans that are easier to normalize. Thus, we can improve recall during candidate generation by 2.9 percentage points compared to our baseline system, which achieved the best score in the original shared task evaluation. Furthermore, we show that this improvement in recall can be fully translated into top-1 accuracy through careful initialization of a subsequent reranking model. Our best system achieves an accuracy of 63.6\\% on the SympTEMIST test set. The proposed approach has been integrated into the open-source xMEN toolkit, which is available online via https://github.com/hpi-dhc/xmen.}}, author = {Borchert, Florian and Llorca, Ignacio and Schapranow, Matthieu-P}, journal = {Database}, keywords = {nlp myown sys:relevantfor:dhc gemtex highmed}, month = {07}, pages = {baae067}, title = {Improving biomedical entity linking for complex entity mentions with {LLM}-based text simplification}, volume = 2024, year = 2024 }
2.
Bressem, K.K., Papaioannou, J.-M., Grundmann, P., Borchert, F., Adams, L.C., Liu, L., Busch, F., Xu, L., Loyen, J.P., Niehues, S.M., Augustin, M., Grosser, L., Makowski, M.R., Aerts, H.J., Löser, A.: medBERT.de: A Comprehensive German BERT Model for the Medical Domain. Expert Systems with Applications. 121598 (2024).
[ Abstract ] [ BibTeX ] [ URL ]
 
This paper presents medBERT.de, a pre-trained German BERT model specifically designed for the German medical domain. The model has been trained on a large corpus of 4.7 Million German medical documents and has been shown to achieve new state-of-the-art performance on eight different medical benchmarks covering a wide range of disciplines and medical document types. In addition to evaluating the overall performance of the model, this paper also conducts a more in-depth analysis of its capabilities. We investigate the impact of data deduplication on the model's performance, as well as the potential benefits of using more efficient tokenization methods. Our results indicate that domain-specific models such as medBERT.de are particularly useful for longer texts, and that deduplication of training data does not necessarily lead to improved performance. Furthermore, we found that efficient tokenization plays only a minor role in improving model performance, and attribute most of the improved performance to the large amount of training data. To encourage further research, the pre-trained model weights and new benchmarks based on radiological data are made publicly available for use by the scientific community.
@article{BRESSEM2023121598, abstract = {This paper presents medBERT.de, a pre-trained German BERT model specifically designed for the German medical domain. The model has been trained on a large corpus of 4.7 Million German medical documents and has been shown to achieve new state-of-the-art performance on eight different medical benchmarks covering a wide range of disciplines and medical document types. In addition to evaluating the overall performance of the model, this paper also conducts a more in-depth analysis of its capabilities. We investigate the impact of data deduplication on the model's performance, as well as the potential benefits of using more efficient tokenization methods. Our results indicate that domain-specific models such as medBERT.de are particularly useful for longer texts, and that deduplication of training data does not necessarily lead to improved performance. Furthermore, we found that efficient tokenization plays only a minor role in improving model performance, and attribute most of the improved performance to the large amount of training data. To encourage further research, the pre-trained model weights and new benchmarks based on radiological data are made publicly available for use by the scientific community.}, author = {Bressem, Keno K. and Papaioannou, Jens-Michalis and Grundmann, Paul and Borchert, Florian and Adams, Lisa C. and Liu, Leonhard and Busch, Felix and Xu, Lina and Loyen, Jan P. and Niehues, Stefan M. and Augustin, Moritz and Grosser, Lennart and Makowski, Marcus R. and Aerts, Hugo JWL. and Löser, Alexander}, journal = {Expert Systems with Applications}, keywords = {nlp myown sys:relevantfor:dhc fb-boettinger}, month = {mar}, pages = 121598, title = {medBERT.de: A Comprehensive German BERT Model for the Medical Domain}, year = 2024 }
3.
Steckhan, N., Ring, R., Borchert, F., Koppold, D.A.: Triangulation of Questionnaires, Qualitative Data and Natural Language Processing: A Differential Approach to Religious Bahá’í Fasting in Germany. Journal of Religion and Health. 63, 3360–3373 (2024).
[ Abstract ] [ BibTeX ]
 
Approaches to integrating mixed methods into medical research are gaining popularity. To get a holistic understanding of the effects of behavioural interventions, we investigated religious fasting using a triangulation of quantitative, qualitative, and natural language analysis. We analysed an observational study of Bahá'í fasting in Germany using a between-method triangulation that is based on links between qualitative and quantitative analyses. Individual interviews show an increase in the mindfulness and well-being categories. Sentiment scores, extracted from the interviews through natural language processing, positively correlate with questionnaire outcomes on quality of life (WHO-5: Spearman correlation r = 0.486, p = 0.048). Five questionnaires contribute to the first principal component capturing the spectrum of mood states (50.1% explained variance). Integrating the findings of the between-method triangulation enabled us to converge on the underlying effects of this kind of intermittent fasting.
@article{steckhan2023triangulation, abstract = {Approaches to integrating mixed methods into medical research are gaining popularity. To get a holistic understanding of the effects of behavioural interventions, we investigated religious fasting using a triangulation of quantitative, qualitative, and natural language analysis. We analysed an observational study of Bahá'í fasting in Germany using a between-method triangulation that is based on links between qualitative and quantitative analyses. Individual interviews show an increase in the mindfulness and well-being categories. Sentiment scores, extracted from the interviews through natural language processing, positively correlate with questionnaire outcomes on quality of life (WHO-5: Spearman correlation r = 0.486, p = 0.048). Five questionnaires contribute to the first principal component capturing the spectrum of mood states (50.1% explained variance). Integrating the findings of the between-method triangulation enabled us to converge on the underlying effects of this kind of intermittent fasting.}, author = {Steckhan, Nico and Ring, Raphaela and Borchert, Florian and Koppold, Daniela A.}, journal = {Journal of Religion and Health}, keywords = {nlp sys:relevantfor:dhc}, pages = {3360–3373}, title = {Triangulation of Questionnaires, Qualitative Data and Natural Language Processing: A Differential Approach to Religious Bahá’í Fasting in Germany}, volume = 63, year = 2024 }

2023

1.
Borchert, F., Llorca, I., Schapranow, M.-P.: Cross-Lingual Candidate Retrieval and Re-ranking for Biomedical Entity Linking. In: Arampatzis, A., Kanoulas, E., Tsikrika, T., Vrochidis, S., Giachanou, A., Li, D., Aliannejadi, M., Vlachos, M., Faggioli, G., en Ferro, N. (reds.) Experimental IR Meets Multilinguality, Multimodality, and Interaction. bll. 135–147. Springer Nature Switzerland, Cham (2023).
[ BibTeX ] [ URL ]
 
@inproceedings{10.1007/978-3-031-42448-9_12, address = {Cham}, author = {Borchert, Florian and Llorca, Ignacio and Schapranow, Matthieu-P.}, booktitle = {Experimental IR Meets Multilinguality, Multimodality, and Interaction}, editor = {Arampatzis, Avi and Kanoulas, Evangelos and Tsikrika, Theodora and Vrochidis, Stefanos and Giachanou, Anastasia and Li, Dan and Aliannejadi, Mohammad and Vlachos, Michalis and Faggioli, Guglielmo and Ferro, Nicola}, keywords = {nlp myown sys:relevantfor:dhc gemtex nephrocage fb-boettinger}, pages = {135–147}, publisher = {Springer Nature Switzerland}, title = {Cross-Lingual Candidate Retrieval and Re-ranking for Biomedical Entity Linking}, year = 2023 }
2.
Borchert, F., Llorca, I., Schapranow, M.-P.: HPI-DHC @ BC8 SympTEMIST Track: Detection and Normalization of Symptom Mentions with SpanMarker and xMEN. In: Islamaj, R., Arighi, C., Campbell, I., Gonzalez-Hernandez, G., Hirschman, L., Krallinger, M., Lima-López, S., Weissenbacher, D., en Lu, Z. (reds.) Proceedings of the BioCreative VIII Challenge and Workshop: Curation and Evaluation in the era of Generative Models. , New Orleans, LA (2023).
[ BibTeX ] [ URL ]
 
@inproceedings{borchert2023hpidhc, address = {New Orleans, LA}, author = {Borchert, Florian and Llorca, Ignacio and Schapranow, Matthieu-P.}, booktitle = {Proceedings of the BioCreative VIII Challenge and Workshop: Curation and Evaluation in the era of Generative Models}, editor = {Islamaj, Rezarta and Arighi, Cecilia and Campbell, Ian and Gonzalez-Hernandez, Graciela and Hirschman, Lynette and Krallinger, Martin and Lima-López, Salvador and Weissenbacher, Davy and Lu, Zhiyong}, keywords = {nlp myown sys:relevantfor:dhc gemtex}, title = {HPI-DHC @ BC8 SympTEMIST Track: Detection and Normalization of Symptom Mentions with SpanMarker and xMEN}, year = 2023 }
3.
Fox, S., Preiß, M., Borchert, F., Rasheed, A., Schapranow, M.-P.: HPIDHC at NTCIR-17 MedNLP-SC: Data Augmentation and Ensemble Learning for Multilingual Adverse Drug Event Detection. NTCIR 17 Conference: Proceedings of the 17th NTCIR Conference on Evaluation of Information Access Technologies. bll. 185–192. , Tokyo, Japan (2023).
[ BibTeX ] [ URL ]
 
@inproceedings{fox2023hpidhc, address = {Tokyo, Japan}, author = {Fox, Smilla and Preiß, Martin and Borchert, Florian and Rasheed, Aadil and Schapranow, Matthieu-P.}, booktitle = {NTCIR 17 Conference: Proceedings of the 17th NTCIR Conference on Evaluation of Information Access Technologies}, keywords = {nlp myown sys:relevantfor:dhc gemtex}, month = {dec}, pages = {185-192}, title = {HPIDHC at NTCIR-17 MedNLP-SC: Data Augmentation and Ensemble Learning for Multilingual Adverse Drug Event Detection}, year = 2023 }
4.
Hugo, J., Ibing, S., Borchert, F., Sachs, J.P., Cho, J., Ungaro, R.C., Böttinger, E.P.: Machine Learning Based Prediction of Incident Cases of Crohn’s Disease Using Electronic Health Records from a Large Integrated Health System. In: Juarez, J.M., Marcos, M., Stiglic, G., en Tucker, A. (reds.) Artificial Intelligence in Medicine. bll. 293–302. Springer Nature Switzerland, Cham (2023).
[ BibTeX ] [ URL ] [ Download ]
 
@inproceedings{10.1007/978-3-031-34344-5_35, address = {Cham}, author = {Hugo, Julian and Ibing, Susanne and Borchert, Florian and Sachs, Jan Philipp and Cho, Judy and Ungaro, Ryan C. and B{ö}ttinger, Erwin P.}, booktitle = {Artificial Intelligence in Medicine}, editor = {Juarez, Jose M. and Marcos, Mar and Stiglic, Gregor and Tucker, Allan}, keywords = {sys:relevantfor:dhc fb-boettinger}, pages = {293–302}, publisher = {Springer Nature Switzerland}, title = {Machine Learning Based Prediction of Incident Cases of Crohn's Disease Using Electronic Health Records from a Large Integrated Health System}, year = 2023 }
5.
Kämmer, N., and Borchert, F., and Winkler, S., and de Melo, G., and Schapranow, M.-P.: Resolving Elliptical Compounds in German Medical Text. The 22nd Workshop on Biomedical Natural Language Processing and BioNLP Shared Tasks. bll. 292–305. Association for Computational Linguistics, Toronto, Canada (2023).
[ BibTeX ] [ URL ]
 
@inproceedings{kammer2023resolving, address = {Toronto, Canada}, author = {Kämmer, Niklas and and Borchert, Florian and and Winkler, Silvia and and de Melo, Gerard and and Schapranow, Matthieu-P.}, booktitle = {The 22nd Workshop on Biomedical Natural Language Processing and BioNLP Shared Tasks}, keywords = {nlp myown sys:relevantfor:dhc gemtex highmed nephrocage fb-boettinger}, pages = {292-305}, publisher = {Association for Computational Linguistics}, title = {Resolving Elliptical Compounds in German Medical Text}, year = 2023 }
6.
Ladas, N., Borchert, F., Franz, S., Rehberg, A., Strauch, N., Sommer, K.K., Marschollek, M., Gietzelt, M.: Programming techniques for improving rule readability for rule-based information extraction natural language processing pipelines of unstructured and semi-structured medical texts. Health Informatics Journal. 29, 14604582231164696 (2023).
[ BibTeX ] [ URL ]
 
@article{doi:10.1177/14604582231164696, author = {Ladas, Nektarios and Borchert, Florian and Franz, Stefan and Rehberg, Alina and Strauch, Natalia and Sommer, Kim Katrin and Marschollek, Michael and Gietzelt, Matthias}, journal = {Health Informatics Journal}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, note = {PMID: 37068028}, number = 2, pages = 14604582231164696, title = {Programming techniques for improving rule readability for rule-based information extraction natural language processing pipelines of unstructured and semi-structured medical texts}, volume = 29, year = 2023 }
7.
Llorca, I., Borchert, F., Schapranow, M.-P.: A Meta-dataset of German Medical Corpora: Harmonization of Annotations and Cross-corpus NER Evaluation. Proceedings of the 5th Clinical Natural Language Processing Workshop. bll. 171–181. Association for Computational Linguistics, Toronto, Canada (2023).
[ Abstract ] [ BibTeX ] [ URL ]
 
Over the last years, an increasing number of publicly available, semantically annotated medical corpora have been released for the German language. While their annotations cover comparable semantic classes, the synergies of such efforts have not been explored, yet. This is due to substantial differences in the data schemas (syntax) and annotated entities (semantics), which hinder the creation of common meta-datasets. For instance, it is unclear whether named entity recognition (NER) taggers trained on one or more of such datasets are useful to detect entities in any of the other datasets. In this work, we create harmonized versions of German medical corpora using the BigBIO framework, and make them available to the community. Using these as a meta-dataset, we perform a series of cross-corpus evaluation experiments on two settings of aligned labels. These consist in fine-tuning various pre-trained Transformers on different combinations of training sets, and testing them against each dataset separately. We find that a) trained NER models generalize poorly, with F1 scores dropping approx. 20 pp. on unseen test data, and b) current pre-trained Transformer models for the German language do not systematically alleviate this issue. However, our results suggest that models benefit from additional training corpora in most cases, even if these belong to different medical fields or text genres.
@inproceedings{llorca-etal-2023-meta, abstract = {Over the last years, an increasing number of publicly available, semantically annotated medical corpora have been released for the German language. While their annotations cover comparable semantic classes, the synergies of such efforts have not been explored, yet. This is due to substantial differences in the data schemas (syntax) and annotated entities (semantics), which hinder the creation of common meta-datasets. For instance, it is unclear whether named entity recognition (NER) taggers trained on one or more of such datasets are useful to detect entities in any of the other datasets. In this work, we create harmonized versions of German medical corpora using the BigBIO framework, and make them available to the community. Using these as a meta-dataset, we perform a series of cross-corpus evaluation experiments on two settings of aligned labels. These consist in fine-tuning various pre-trained Transformers on different combinations of training sets, and testing them against each dataset separately. We find that a) trained NER models generalize poorly, with F1 scores dropping approx. 20 pp. on unseen test data, and b) current pre-trained Transformer models for the German language do not systematically alleviate this issue. However, our results suggest that models benefit from additional training corpora in most cases, even if these belong to different medical fields or text genres.}, address = {Toronto, Canada}, author = {Llorca, Ignacio and Borchert, Florian and Schapranow, Matthieu-P.}, booktitle = {Proceedings of the 5th Clinical Natural Language Processing Workshop}, keywords = {nlp myown sys:relevantfor:dhc gemtex highmed nephrocage fb-boettinger}, month = {jul}, pages = {171–181}, publisher = {Association for Computational Linguistics}, title = {A Meta-dataset of {G}erman Medical Corpora: Harmonization of Annotations and Cross-corpus {NER} Evaluation}, year = 2023 }
8.
Richter-Pechanski, P., Wiesenbach, P., Schwab, D.M., Kiriakou, C., He, M., Allers, M.M., Tiefenbacher, A.S., Kunz, N., Martynova, A., Spiller, N., Mierisch, J., Borchert, F., Schwind, C., Frey, N., Dieterich, C., Geis, N.A.: A Distributable German Clinical Corpus Containing Cardiovascular Clinical Routine Doctor’s Letters. Scientific Data. 10, 207 (2023).
[ Abstract ] [ BibTeX ] [ URL ]
 
We present CARDIO:DE, the first freely available and distributable large German clinical corpus from the cardiovascular domain. CARDIO:DE encompasses 500 clinical routine German doctor's letters from Heidelberg University Hospital, which were manually annotated. Our prospective study design complies well with current data protection regulations and allows us to keep the original structure of clinical documents consistent. In order to ease access to our corpus, we manually de-identified all letters. To enable various information extraction tasks the temporal information in the documents was preserved. We added two high-quality manual annotation layers to CARDIO:DE, (1) medication information and (2) CDA-compliant section classes. To the best of our knowledge, CARDIO:DE is the first freely available and distributable German clinical corpus in the cardiovascular domain. In summary, our corpus offers unique opportunities for collaborative and reproducible research on natural language processing models for German clinical texts.
@article{richter-pechanskiDistributableGermanClinical2023, abstract = {We present CARDIO:DE, the first freely available and distributable large German clinical corpus from the cardiovascular domain. CARDIO:DE encompasses 500 clinical routine German doctor's letters from Heidelberg University Hospital, which were manually annotated. Our prospective study design complies well with current data protection regulations and allows us to keep the original structure of clinical documents consistent. In order to ease access to our corpus, we manually de-identified all letters. To enable various information extraction tasks the temporal information in the documents was preserved. We added two high-quality manual annotation layers to CARDIO:DE, (1) medication information and (2) CDA-compliant section classes. To the best of our knowledge, CARDIO:DE is the first freely available and distributable German clinical corpus in the cardiovascular domain. In summary, our corpus offers unique opportunities for collaborative and reproducible research on natural language processing models for German clinical texts.}, author = {{Richter-Pechanski}, Phillip and Wiesenbach, Philipp and Schwab, Dominic M. and Kiriakou, Christina and He, Mingyang and Allers, Michael M. and Tiefenbacher, Anna S. and Kunz, Nicola and Martynova, Anna and Spiller, Noemie and Mierisch, Julian and Borchert, Florian and Schwind, Charlotte and Frey, Norbert and Dieterich, Christoph and Geis, Nicolas A.}, journal = {Scientific Data}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, month = {apr}, number = 1, pages = 207, title = {A Distributable {{German}} Clinical Corpus Containing Cardiovascular Clinical Routine Doctor's Letters}, volume = 10, year = 2023 }
9.
Schapranow, M.-P., Borchert, F., Bougatf, N., Hund, H., Eils, R.: Software-Tool Support for Collaborative, Virtual, Multi-Site Molecular Tumor Boards. SN Computer Science. 4, 358 (2023).
[ BibTeX ] [ URL ]
 
@article{schapranow2023software, author = {Schapranow, Matthieu-P and Borchert, Florian and Bougatf, Nina and Hund, Hauke and Eils, Roland}, journal = {SN Computer Science}, keywords = {sys:relevantfor:dhc highmed fb-boettinger}, number = 4, pages = 358, publisher = {Springer}, title = {Software-Tool Support for Collaborative, Virtual, Multi-Site Molecular Tumor Boards}, volume = 4, year = 2023 }
10.
Schmidt, L., Ibing, S., Borchert, F., Hugo, J., Marshall, A., Peraza, J., Cho, J.H., Böttinger, E.P., Ungaro, R.C.: Extraction of Crohn’s Disease Clinical Phenotypes from Clinical Text Using Natural Language Processing. medRxiv. (2023).
[ BibTeX ] [ URL ]
 
@article{Schmidt2023.10.16.23297099, author = {Schmidt, Linea and Ibing, Susanne and Borchert, Florian and Hugo, Julian and Marshall, Allison and Peraza, Jellyana and Cho, Judy H. and B{ö}ttinger, Erwin P. and Ungaro, Ryan C.}, journal = {medRxiv}, keywords = {nlp sys:relevantfor:dhc fb-boettinger}, publisher = {Cold Spring Harbor Laboratory Press}, title = {Extraction of Crohn's Disease Clinical Phenotypes from Clinical Text Using Natural Language Processing}, year = 2023 }
11.
Steinwand, S., Borchert, F., Winkler, S., Schapranow, M.-P.: GGTWEAK: Gene Tagging with Weak Supervision for German Clinical Text. In: Juarez, J.M., Marcos, M., Stiglic, G., en Tucker, A. (reds.) Artificial Intelligence in Medicine. bll. 183–192. Springer Nature Switzerland, Cham (2023).
[ BibTeX ] [ URL ] [ Download ]
 
@inproceedings{10.1007/978-3-031-34344-5_22, address = {Cham}, author = {Steinwand, Sandro and Borchert, Florian and Winkler, Silvia and Schapranow, Matthieu-P.}, booktitle = {Artificial Intelligence in Medicine}, editor = {Juarez, Jose M. and Marcos, Mar and Stiglic, Gregor and Tucker, Allan}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, pages = {183–192}, publisher = {Springer Nature Switzerland}, title = {GGTWEAK: Gene Tagging with Weak Supervision for German Clinical Text}, year = 2023 }

2022

1.
Borchert, F., Lohr, C., Modersohn, L., Witt, J., Langer, T., Follmann, M., Gietzelt, M., Arnrich, B., Hahn, U., Schapranow, M.-P.: GGPONC 2.0 - The German Clinical Guideline Corpus for Oncology: Curation Workflow, Annotation Policy, Baseline NER Taggers. Proceedings of the Language Resources and Evaluation Conference. bll. 3650–3660. European Language Resources Association, Marseille, France (2022).
[ Abstract ] [ BibTeX ] [ URL ]
 
Despite remarkable advances in the development of language resources over the recent years, there is still a shortage of annotated, publicly available corpora covering (German) medical language. With the initial release of the German Guideline Program in Oncology NLP Corpus (GGPONC), we have demonstrated how such corpora can be built upon clinical guidelines, a widely available resource in many natural languages with a reasonable coverage of medical terminology. In this work, we describe a major new release for GGPONC. The corpus has been substantially extended in size and re-annotated with a new annotation scheme based on SNOMED CT top level hierarchies, reaching high inter-annotator agreement (γ=.94). Moreover, we annotated elliptical coordinated noun phrases and their resolutions, a common language phenomenon in (not only German) scientific documents. We also trained BERT-based named entity recognition models on this new data set, which achieve high performance on short, coarse-grained entity spans (F1=.89), while the rate of boundary errors increases for long entity spans. GGPONC is freely available through a data use agreement. The trained named entity recognition models, as well as the detailed annotation guide, are also made publicly available.
@inproceedings{borchert-EtAl:2022:LREC, abstract = {Despite remarkable advances in the development of language resources over the recent years, there is still a shortage of annotated, publicly available corpora covering (German) medical language. With the initial release of the German Guideline Program in Oncology NLP Corpus (GGPONC), we have demonstrated how such corpora can be built upon clinical guidelines, a widely available resource in many natural languages with a reasonable coverage of medical terminology. In this work, we describe a major new release for GGPONC. The corpus has been substantially extended in size and re-annotated with a new annotation scheme based on SNOMED CT top level hierarchies, reaching high inter-annotator agreement (γ=.94). Moreover, we annotated elliptical coordinated noun phrases and their resolutions, a common language phenomenon in (not only German) scientific documents. We also trained BERT-based named entity recognition models on this new data set, which achieve high performance on short, coarse-grained entity spans (F1=.89), while the rate of boundary errors increases for long entity spans. GGPONC is freely available through a data use agreement. The trained named entity recognition models, as well as the detailed annotation guide, are also made publicly available.}, address = {Marseille, France}, author = {Borchert, Florian and Lohr, Christina and Modersohn, Luise and Witt, Jonas and Langer, Thomas and Follmann, Markus and Gietzelt, Matthias and Arnrich, Bert and Hahn, Udo and Schapranow, Matthieu-P.}, booktitle = {Proceedings of the Language Resources and Evaluation Conference}, keywords = {nlp sys:relevantfor:dhc highmed fb-boettinger}, month = {June}, pages = {3650–3660}, publisher = {European Language Resources Association}, title = {GGPONC 2.0 - The German Clinical Guideline Corpus for Oncology: Curation Workflow, Annotation Policy, Baseline NER Taggers}, year = 2022 }
2.
Borchert, F., Schapranow, M.-P.: HPI-DHC @ BioASQ DisTEMIST: Spanish Biomedical Entity Linking with Pre-trained Transformers and Cross-lingual Candidate Retrieval. Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum. bll. 244–258. , Bologna, Italy (2022).
[ BibTeX ] [ URL ]
 
@inproceedings{borchert2022clef, address = {Bologna, Italy}, author = {Borchert, Florian and Schapranow, Matthieu-P.}, booktitle = {Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, month = {September}, pages = {244-258}, title = {HPI-DHC @ BioASQ DisTEMIST: Spanish Biomedical Entity Linking with Pre-trained Transformers and Cross-lingual Candidate Retrieval}, year = 2022 }
3.
Henkenjohann, R., Bergner, B., Borchert, F., Bougatf, N., Hund, H., Eils, R., Schapranow, M.-P.: An Engineering Approach towards Multi-Site Virtual Molecular Tumor Board Software Support. In: Pissaloux, E., Papadopoulos, G., Achilleos, A., en Velázquez, R. (reds.) ICT for Health, Accessibility and Wellbeing. IHAW 2021. bll. 156–170. Springer, Cham (2022).
[ BibTeX ] [ URL ]
 
@inproceedings{henkenjohann2021mtb, author = {Henkenjohann, Richard and Bergner, Benjamin and Borchert, Florian and Bougatf, Nina and Hund, Hauke and Eils, Roland and Schapranow, Matthieu-P.}, booktitle = {ICT for Health, Accessibility and Wellbeing. IHAW 2021}, editor = {Pissaloux, E. and Papadopoulos, G.A. and Achilleos, A. and Velázquez, R.}, keywords = {myown highmed mtb}, month = {mar}, pages = {156-170}, publisher = {Springer, Cham.}, series = {Communications in Computer and Information Science}, title = {An Engineering Approach towards Multi-Site Virtual Molecular Tumor Board Software Support}, volume = 1538, year = 2022 }

2021

1.
Borchert, F., Meister, L., Langer, T., Follmann, M., Arnrich, B., Schapranow, M.-P.: Controversial Trials First: Identifying Disagreement Between Clinical Guidelines and New Evidence. AMIA Annual Symposium Proceedings. bll. 237–246. American Medical Informatics Association (2021).
[ BibTeX ] [ URL ]
 
@inproceedings{borchert_amia2021, author = {Borchert, Florian and Meister, Laura and Langer, Thomas and Follmann, Markus and Arnrich, Bert and Schapranow, Matthieu-P.}, booktitle = {AMIA Annual Symposium Proceedings}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, organization = {American Medical Informatics Association}, pages = {237-246}, title = {Controversial Trials First: Identifying Disagreement Between Clinical Guidelines and New Evidence}, volume = 2021, year = 2021 }
2.
Borchert, F., Mock, A., Tomczak, A., Hügel, J., Alkarkoukly, S., Knurr, A., Volckmar, A.-L., Stenzinger, A., Schirmacher, P., Debus, J., Jäger, D., Longerich, T., Fröhling, S., Eils, R., Bougatf, N., Sax, U., Schapranow, M.-P.: Knowledge bases and software support for variant interpretation in precision oncology. Briefings in Bioinformatics. 22, (2021).
[ Abstract ] [ BibTeX ] [ URL ]
 
Precision oncology is a rapidly evolving interdisciplinary medical specialty. Comprehensive cancer panels are becoming increasingly available at pathology departments worldwide, creating the urgent need for scalable cancer variant annotation and molecularly informed treatment recommendations. A wealth of mainly academia-driven knowledge bases calls for software tools supporting the multi-step diagnostic process. We derive a comprehensive list of knowledge bases relevant for variant interpretation by a review of existing literature followed by a survey among medical experts from university hospitals in Germany. In addition, we review cancer variant interpretation tools, which integrate multiple knowledge bases. We categorize the knowledge bases along the diagnostic process in precision oncology and analyze programmatic access options as well as the integration of knowledge bases into software tools. The most commonly used knowledge bases provide good programmatic access options and have been integrated into a range of software tools. For the wider set of knowledge bases, access options vary across different parts of the diagnostic process. Programmatic access is limited for information regarding clinical classifications of variants and for therapy recommendations. The main issue for databases used for biological classification of pathogenic variants and pathway context information is the lack of standardized interfaces. There is no single cancer variant interpretation tool that integrates all identified knowledge bases. Specialized tools are available and need to be further developed for different steps in the diagnostic process.
@article{10.1093/bib/bbab134, abstract = {Precision oncology is a rapidly evolving interdisciplinary medical specialty. Comprehensive cancer panels are becoming increasingly available at pathology departments worldwide, creating the urgent need for scalable cancer variant annotation and molecularly informed treatment recommendations. A wealth of mainly academia-driven knowledge bases calls for software tools supporting the multi-step diagnostic process. We derive a comprehensive list of knowledge bases relevant for variant interpretation by a review of existing literature followed by a survey among medical experts from university hospitals in Germany. In addition, we review cancer variant interpretation tools, which integrate multiple knowledge bases. We categorize the knowledge bases along the diagnostic process in precision oncology and analyze programmatic access options as well as the integration of knowledge bases into software tools. The most commonly used knowledge bases provide good programmatic access options and have been integrated into a range of software tools. For the wider set of knowledge bases, access options vary across different parts of the diagnostic process. Programmatic access is limited for information regarding clinical classifications of variants and for therapy recommendations. The main issue for databases used for biological classification of pathogenic variants and pathway context information is the lack of standardized interfaces. There is no single cancer variant interpretation tool that integrates all identified knowledge bases. Specialized tools are available and need to be further developed for different steps in the diagnostic process.}, author = {Borchert, Florian and Mock, Andreas and Tomczak, Aurelie and Hügel, Jonas and Alkarkoukly, Samer and Knurr, Alexander and Volckmar, Anna-Lena and Stenzinger, Albrecht and Schirmacher, Peter and Debus, Jürgen and Jäger, Dirk and Longerich, Thomas and Fröhling, Stefan and Eils, Roland and Bougatf, Nina and Sax, Ulrich and Schapranow, Matthieu-P}, journal = {Briefings in Bioinformatics}, keywords = {HiGHmed myown sys:relevantfor:dhc fb-boettinger}, month = {05}, note = {bbab134}, number = 6, title = {Knowledge bases and software support for variant interpretation in precision oncology}, volume = 22, year = 2021 }
3.
Rasheed, A., Borchert, F., Kohlmeyer, L., Henkenjohann, R., Schapranow, M.-P.: A Comparison of Concept Embeddings for German Clinical Corpora. 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). bll. 2314–2321 (2021).
[ Abstract ] [ BibTeX ] [ URL ]
 
Clinical concept embeddings enable unsupervised learning of relationships among medical concepts. A range of benchmarks quantifies the degree to which learned representations capture medical semantics. However, training and evaluation of embeddings require a large amount of data. In addition, embeddings’ benchmark score varies in different languages because it differs with the size of the available corpora. Multi-modal data increases the corpus size, but data protection regulations limit access to clinical multi-modal data. We present an extendable pipeline for training clinical concept embeddings on various text corpora and evaluating the quality of trained embeddings on selected benchmark tasks. Our work provides different ways to identify clinical concepts in textual corpora. We train embeddings on selected German clinical text corpora and evaluate them on various benchmark scores. Our work can be extended to train embeddings in other languages in which a large multi-modal dataset is not available.
@inproceedings{rasheed_concept_embeddings, abstract = {Clinical concept embeddings enable unsupervised learning of relationships among medical concepts. A range of benchmarks quantifies the degree to which learned representations capture medical semantics. However, training and evaluation of embeddings require a large amount of data. In addition, embeddings’ benchmark score varies in different languages because it differs with the size of the available corpora. Multi-modal data increases the corpus size, but data protection regulations limit access to clinical multi-modal data. We present an extendable pipeline for training clinical concept embeddings on various text corpora and evaluating the quality of trained embeddings on selected benchmark tasks. Our work provides different ways to identify clinical concepts in textual corpora. We train embeddings on selected German clinical text corpora and evaluate them on various benchmark scores. Our work can be extended to train embeddings in other languages in which a large multi-modal dataset is not available.}, author = {Rasheed, Aadil and Borchert, Florian and Kohlmeyer, Lasse and Henkenjohann, Richard and Schapranow, Matthieu-P.}, booktitle = {2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)}, keywords = {nlp myown sys:relevantfor:dhc highmed fb-boettinger}, pages = {2314-2321}, title = {A Comparison of Concept Embeddings for German Clinical Corpora}, year = 2021 }

2020

1.
Borchert, F., Lohr, C., Modersohn, L., Hahn, U., Langer, T., Wenzel, G., Follmann, M., Schapranow, M.-P.: "Herr Doktor, verstehen Sie mich?“: Wie lernende Systeme helfen medizinische Fachsprache zu verstehen und welche Rolle klinische Leitlinien dabei spielen. gesundhyte.de: Das Magazin für Digitale Gesundheit in Deutschland. 13, 19–22 (2020).
[ BibTeX ]
 
@article{borchert2020doktor, author = {Borchert, Florian and Lohr, Christina and Modersohn, Luise and Hahn, Udo and Langer, Thomas and Wenzel, Gregor and Follmann, Markus and Schapranow, Matthieu-P.}, journal = {gesundhyte.de: Das Magazin für Digitale Gesundheit in Deutschland}, keywords = {HiGHmed myown}, month = {dec}, pages = {19–22}, title = {"Herr Doktor, verstehen Sie mich?“: Wie lernende Systeme helfen medizinische Fachsprache zu verstehen und welche Rolle klinische Leitlinien dabei spielen}, volume = 13, year = 2020 }
2.
Borchert, F., Lohr, C., Modersohn, L., Langer, T., Follmann, M., Sachs, J.P., Hahn, U., Schapranow, M.-P.: GGPONC: A Corpus of German Medical Text with Rich Metadata Based on Clinical Practice Guidelines. Proceedings of the 11th International Workshop on Health Text Mining and Information Analysis. bll. 38–48. Association for Computational Linguistics, Online (2020).
[ Abstract ] [ BibTeX ] [ URL ]
 
The lack of publicly accessible text corpora is a major obstacle for progress in natural language processing. For medical applications, unfortunately, all language communities other than English are low-resourced. In this work, we present GGPONC (German Guideline Program in Oncology NLP Corpus), a freely distributable German language corpus based on clinical practice guidelines for oncology. This corpus is one of the largest ever built from German medical documents. Unlike clinical documents, clinical guidelines do not contain any patient-related information and can therefore be used without data protection restrictions. Moreover, GGPONC is the first corpus for the German language covering diverse conditions in a large medical subfield and provides a variety of metadata, such as literature references and evidence levels. By applying and evaluating existing medical information extraction pipelines for German text, we are able to draw comparisons for the use of medical language to other corpora, medical and non-medical ones.
@inproceedings{borchert-etal-2020-ggponc, abstract = {The lack of publicly accessible text corpora is a major obstacle for progress in natural language processing. For medical applications, unfortunately, all language communities other than English are low-resourced. In this work, we present GGPONC (German Guideline Program in Oncology NLP Corpus), a freely distributable German language corpus based on clinical practice guidelines for oncology. This corpus is one of the largest ever built from German medical documents. Unlike clinical documents, clinical guidelines do not contain any patient-related information and can therefore be used without data protection restrictions. Moreover, GGPONC is the first corpus for the German language covering diverse conditions in a large medical subfield and provides a variety of metadata, such as literature references and evidence levels. By applying and evaluating existing medical information extraction pipelines for German text, we are able to draw comparisons for the use of medical language to other corpora, medical and non-medical ones.}, address = {Online}, author = {Borchert, Florian and Lohr, Christina and Modersohn, Luise and Langer, Thomas and Follmann, Markus and Sachs, Jan Philipp and Hahn, Udo and Schapranow, Matthieu-P.}, booktitle = {Proceedings of the 11th International Workshop on Health Text Mining and Information Analysis}, keywords = {nlp HiGHmed myown sys:relevantfor:dhc fb-boettinger}, month = {nov}, pages = {38–48}, publisher = {Association for Computational Linguistics}, title = {{GGPONC}: A Corpus of {G}erman Medical Text with Rich Metadata Based on Clinical Practice Guidelines}, year = 2020 }