Elamur Research⌕
☰
home / papers / 2026 / Extracting Medical Image-Related Entities from Spanish Electronic Health Records Using NER Methods

Extracting Medical Image-Related Entities from Spanish Electronic Health Records Using NER Methods

Summary

Radiology reports in Spanish mention many clinical concepts, but only some of them can be seen in the matching medical images. This paper builds a Spanish NER resource restricted to such visually observable entities, as a link between clinical NLP and computer vision. The authors started from Barcelona Supercomputing Center corpora that share one set of 1,000 clinical documents. From MEDDOCAN they kept only Age and Sex. For MedProcNER and DisTEMIST, they kept procedures and diseases visible in imaging using a filtering pipeline: MediPhi labelled each candidate, Gemini 2.5 Pro resolved ambiguous cases, and humans reviewed the output. The resulting four-class corpus is released in BRAT and BIO formats. The authors then benchmark three model families on it: encoder-only, encoder-decoder and decoder-only models, using fine-tuning, zero-shot and few-shot ICL. The supplied text omits the results section, so no scores or best-model claims are reported here. The resource matters because it can support multimodal tasks such as image-text retrieval and cross-modal alignment in Spanish.

Key points

  • Corpus size and split: 1,000 Spanish reports (701 train / 150 dev / 149 test), stratified by class. It is text only, with no images.
  • Entity counts: 12,470 entities in total.
    • Age (Edad): 1,100
    • Sex (Sexo): 863
    • Disease (Enfermedad): 6,862
    • Procedure (Procedimiento): 3,645
  • Filtering and curation:
    • MedProcNER went from 14,684 to 3,645 entities.
    • DisTEMIST went from 10,663 to 6,862 entities.
    • MEDDOCAN went from 22,795 to 1,963 entities. Its errors were corrected by hand, and document headers were removed.
    • PharmaCoNER, CodiEsp and SympTEMIST were excluded.
  • Encoder-only models:
    • Multilingual BERT, Biomedical RoBERTa and EriBERTa were fine-tuned for 20 epochs. Inputs longer than 512 tokens were split with a sliding window that overlaps by half.
    • GLiNER-X, GLiNER-BioMed large and NuNER-span were tested zero-shot.
  • Encoder-decoder model: Medical mT5-large was fine-tuned on sentence-level input-output pairs (15,702 sentences). It was tested with two output formats: an entity list and a tagged sentence.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Named entity recognition

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/4t6agzu5ygqr

@inproceedings{platas2026extracting,
  title = {{Extracting Medical Image-Related Entities from Spanish Electronic Health Records Using NER Methods}},
  author = {Platas, Alexander and Merino, Marcos and Zotova, Elena and Cuadros, Montse and López-Linares, Karen and Mendiola, Mikel Pérez de and Gálvez, María and Barba, Cristina and Asla, Antón},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/4t6agzu5ygqr},
  url = {https://doi.org/10.63317/4t6agzu5ygqr}
}

References

35 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ricardo Ahumada, Jocelyn Dunstan, Matías Rojas, Sergio Peñafiel, Inti Paredes, and Pablo Báez. 2024. Automatic detection of distant metastasis mentions in radiology reports in spanish. JCO Clinical Cancer Informatics, 8:e2300130. find
  2. Liliya Akhtyamova. 2020. Named Entity Recognition in Spanish Biomedical Literature: Short Review and BERT Model. In 2020 26th Conference of Open Innovations Association (FRUCT), pages 1–7. find
  3. Anthropic. 2025. System card: Claude sonnet 4.5. find
  4. Pablo Báez, Felipe Bravo-Marquez, Jocelyn Dunstan, Matías Rojas, and Fabián Villena. 2022. Automatic extraction of nested entities in clinical referrals in spanish. ACM Transactions on Computing for Healthcare (HEALTH), 3(3):1–22. find
  5. Sergei Bogdanov, Alexandre Constantin, Timothée Bernard, Benoit Crabbé, and Etienne Bernard. 2024. Nuner: Entity recognition encoder pre-training via llm-annotated data. find
  6. Casimiro Pio Carrino, Jordi Armengol-Estapé, Asier Gutiérrez-Fandiño, Joan Llop-Palao, Marc Pàmies, Aitor Gonzalez-Agirre, and Marta Villegas. 2021. Biomedical and clinical language models for spanish: On the benefits of domain-specific pretraining in a mid-resource scenario. arXiv preprint arXiv:21… — [arXiv:2109.03570](https://arxiv.org/abs/2109.03570) find
  7. Procedure 4.20 20.93 57.53 Procedure 10.28 41.31 Procedure 63.54 69.75 Procedure 75.18 Procedure 28.88 Procedure 77.03 find
  8. Jean-Philippe Corbeil, Amin Dada, Jean-Michel Attendu, Asma Ben Abacha, Alessandro Sordoni, Lucas Caccia, Francois Beaulieu, Thomas Lin, Jens Kleesiek, and Paul Vozila. 2025. A modular approach for clinical slms driven by synthetic data with pre-instruction tuning, model merging, and clinical-tasks… find
  9. Viviana Cotik, Laura Alonso Alemany, Darío Filippo, Franco Luque, Roland Roller, Jorge Vivaldi, Ammer Ayach, Fernando Carranza, Lucas Defrancesca, Antonella Dellanzo, and Macarena Fernández Urquiza. 2021. Overview of CLEF eHealth Task 1 - SpRadIE: A Challenge on Information Extraction from Spanish R… find
  10. Iker de la Iglesia, Aitziber Atutxa, Koldo Gojenola, and Ander Barrena. 2023. Eriberta: A bilingual pre-trained language model for clinical natural language processing. arXiv preprint arXiv:2306.07373. — [arXiv:2306.07373](https://arxiv.org/abs/2306.07373) find
  11. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. BERT: pre-training of deep bidirectional transformers for language understanding. CoRR, abs/1810.04805. find
  12. Álvaro García-Barragán, Alberto González Calatayud, Oswaldo Solarte-Pabón, Mariano Provencio, Ernestina Menasalvas, and Víctor Robles. 2024. Gpt for medical entity recognition in spanish. Multimedia Tools and Applications, pages 1–20. find
  13. Iker García-Ferrero, Rodrigo Agerri, Aitziber Atutxa Salazar, Elena Cabrio, Iker de la Iglesia, Alberto Lavelli, Bernardo Magnini, Benjamin Molinet, Johana Ramirez-Romero, German Rigau, Jose Maria Villa-Gonzalez, Serena Villata, and Andrea Zaninello. 2024. MedMT5: An open-source multilingual text-to… find
  14. Aitor García-Pablos, Naiara Pérez, and Montse Cuadros. 2020. Vicomtech at CANTEMIST 2020. In Proceedings of the Iberian Languages Evaluation Forum (IberLEF 2020) co-located with 36th Conference of the Spanish Society for Natural Language Processing (SEPLN 2020), pages 489–498. find
  15. Google Gemini Team. 2025. Gemini 2.5 pro. find
  16. Eduardo Godoy, Steren Chabert, Marvin Querales, Julio Sotelo, Denis Parra, Carlos Fernández, Diego Mellado, Alejandro Veloz, Scarlett Lever, Favian Pardo, Ayleen Bertini Rojas, Yomar Molina, Claudia Díaz, Rodrigo Ferreira, and Rodrigo Salas. 2023. A named entity recognition framework using transform… find
  17. Iakes Goenaga, Edgar Andres, Koldo Gojenola, and Aitziber Atutxa. 2023. Advances in monolingual and crosslingual automatic disability annotation in spanish. BMC Bioinformatics, 24(1):265. find
  18. Aitor Gonzalez-Agirre, Antonio Miranda-Escalada, Obdulia Rabal, and Martin Krallinger. 2020. Pharmaconer corpus: gold standard annotations of pharmacological substances, compounds and proteins in spanish clinical case reports. Zenodo. find
  19. Richard Jonker, Tiago Almeida, Rui Antunes, João Almeida, and Sérgio Matos. 2024. Multi-head CRF classifier for biomedical multi-class named entity recognition on Spanish clinical notes. Database, 2024:baae068. find
  20. Salvador Lima López, Eulàlia Farré Maduell, Luis Gascó Sánchez, and Martin Krallinger. 2023. Medprocner corpus: Gold standard annotations for clinical procedures information extraction. Zenodo. find
  21. Salvador Lima López, Luis Gascó Sánchez, Eulalia Farré, Laura Vigil Gimenez, and Martin Krallinger. 2024. Symptemist corpus: Gold standard annotations for clinical symptoms, signs and findings information extraction. Zenodo. find
  22. Montserrat Marimon, Aitor Gonzalez-Agirre, Ander Intxaurrondo, Heidy Rodríguez, Jose Antonio Lopez Martin, Marta Villegas, and Martin Krallinger. 2020. Meddocan corpus: gold standard annotations for medical document anonymization on spanish clinical case reports. Zenodo. find
  23. Antonio Miranda-Escalada, Eulàlia Farré, Luis Gasco, Salvador Lima, and Martin Krallinger. 2022. Distemist corpus: detection and normalization of disease mentions in spanish clinical cases. Zenodo. find
  24. Antonio Miranda-Escalada, Aitor Gonzalez-Agirre, and Martin Krallinger. 2020. Codiesp corpus: gold standard spanish clinical cases coded in icd10 (cie10) - ehealth clef2020. Zenodo. find
  25. OpenAI. 2025. Gpt-5 is here. find
  26. Alibaba Qwen Team. 2025. Qwen3 technical report. find
  27. Sara Santiso, Alicia Pérez, and Arantza Casillas. 2021. Adverse drug reaction extraction: Tolerance to entity recognition errors and sub-domain variants. Computer Methods and Programs in Biomedicine, 199:105891. find
  28. Sergio Santiso, Arantza Casillas, Alicia Pérez, and Maite Oronoz. 2017. Medical entity recognition and negation extraction: Assessment of negex on health records in spanish. In Bioinformatics and Biomedical Engineering - 5th International Work-Conference, IWBBIO 2017, Proceedings, Part I, pages 177–… find
  29. Andrew Sellergren, Sahar Kazemzadeh, Tiam Jaroensri, Atilla Kiraly, Madeleine Traverse, Timo Kohlberger, Shawn Xu, Fayaz Jamil, Cían Hughes, Charles Lau, et al. 2025. Medgemma technical report. arXiv preprint arXiv:2507.05201. find
  30. Ihor Stepanov and Mykhailo Shtopko. 2024. Gliner multi-task: Generalist lightweight model for various information extraction tasks. find
  31. Víctor Suárez-Paniagua, Hang Dong, and Arlene Casey. 2021. A multi-BERT hybrid system for Named Entity Recognition in Spanish radiology reports. In 2021 Working Notes of CLEF: Conference and Labs of the Evaluation Forum, CLEF-WN 2021, pages 846–856. find
  32. Antonio Tamayo, Diego Burgos, and Alexander Gelbukh. 2022. Partner: Paragraph tuning for named entity recognition on clinical cases in spanish using mbert+ rules. In Proceedings of the Iberian Languages Evaluation Forum (Iber-LEF 2022) co-located with the Conference of the Spanish Society for Natura… find
  33. Aitana Villaplana, Raquel Martínez, and Soto Montalvo. 2023. Improving medical entity recognition in spanish by means of biomedical language models. Electronics, 12(23). find
  34. Rebecka Weegar, Alicia Pérez, Arantza Casillas, and Maite Oronoz. 2019. Recent advances in swedish and spanish medical entity recognition in clinical texts using deep neural approaches. BMC Medical Informatics and Decision Making, 19(7):274. find
  35. Anthony Yazdani, Ihor Stepanov, and Douglas Teodoro. 2025. Gliner-biomed: A suite of efficient models for open biomedical named entity recognition. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/platas-2026-extracting-medical-image-related-entities/