Dataset

Description

The merged dataset utilized in this project combines four distinct annotated datasets, all based on the Spanish Clinical Case Corpus (SPACCC), a compilation of clinical case reports from Spanish medical publications. This merged dataset encompasses a total of 16,504 sentences across 1,000 clinical cases. The dataset focuses on identifying various medical entities within clinical narratives, including symptoms, medical procedures, diseases, proteins, and chemicals.

The dataset is further provided on Zenodo, and construction is detailed in our GitHub repository.

In order to use this dataset, the documents are also provided.

Origin

The constituent datasets include:

• SympTEMIST: Designed for symptom, signs, and findings annotations, contributing 12,193 annotations.
• MedProcNER: Focused on medical procedure identification with 14,683 instances.
• DisTEMIST: Targeting disease identification, containing 10,663 annotations.
• PharmaCoNER: Primarily for identifying chemicals and proteins, with 7,624 entities.

Annotations and Classes

The annotations are normalized to SNOMED CT, ensuring consistency across datasets. Classes include symptoms, procedures, diseases, proteins, and chemicals. Intra-class overlapping entities are addressed, ensuring model training on the longest span.

Statistics

The dataset is split into training and test sets, with consistent splits across corpora. For detailed statistics, refer to Table 1 and Table 2 below.

Use Case

The dataset serves as a resource for training and evaluating models for medical entity recognition and normalization tasks in Spanish clinical text.

References

1. Miranda-Escalada, A., Gascó, L., Lima-López, S., Farré-Maduell, E., Estrada, D., Nentidis, A., Krithara, A., Katsimpras, G., Paliouras, G., & Krallinger, M. (2022). Overview of DisTEMIST at BioASQ: Automatic detection and normalization of diseases from clinical texts: results, methods, evaluation and multilingual resources. Working Notes of Conference and Labs of the Evaluation (CLEF) Forum. CEUR Workshop Proceedings
2. Lima-López S, Farré-Maduell E, Gascó L, Nentidis A, Krithara A, Katsimpras G, Paliouras G, Krallinger M. Overview of MedProcNER task on medical procedure detection and entity linking at BioASQ 2023. Working Notes of CLEF. 2023.
3. Lima-López, S., Farré-Maduell, E., Gasco-Sánchez, L., Rodríguez-Miret, J. and Krallinger, M. (2023). Overview of SympTEMIST at BioCreative VIII: corpus, guidelines and evaluation of systems for the detection and normalization of symptoms, signs and findings from text. In: Proceedings of the BioCreative VIII Challenge and Workshop: Curation and Evaluation in the era of Generative Models.
4. A. G. Agirre, M. Marimon, A. Intxaurrondo, O. Rabal, M. Villegas, M. Krallinger, Pharmaconer: Pharmacological substances, compounds and proteins named entity recognition track, in: Proceedings of The 5th Workshop on BioNLP Open Shared Tasks, 2019, pp. 1–10.
5. Ander Intxaurrondo, & Krallinger, M. (2018). SPACCC (2019-02-01) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.2560316

---

Table 1: Datasets statistics with the number of entity mentions.

Table 2: Total number of overlapping entities within the datasets.

Licensed under CC4