Final CFP: SYMPTEMIST Shared Task (BioCreative VIII run with AMIA 2023)

NER and entity linking of symptoms

https://temu.bsc.es/SYMPTEMIST/



The SYMPTEMIST focuses on the detection of mentions of symptoms (NER) and mapping to concept identifiers in clinical case reports in Spanish (entity linking). Also a multilingual version of the dataset will be released including versions in English, French, Italian, Dutch, Portuguese, Romanian and Swedish.

Key information:

Web: https://temu.bsc.es/symptemist

Data: https://zenodo.org/record/8223654

Annotation guidelines: https://zenodo.org/record/8246440

BioCreative web: https://biocreative.bioinformatics.udel.edu

Registration form (Track 2- SYMPTEMIST): https://docs.google.com/forms/d/e/1FAIpQLScoSNulOoxRju3c8v9Q-CSv-w5jJcXu93G7A7v343AWfonpPw/viewform


Motivation
Systems able to detect and normalize clinical symptom mentions from medical texts are crucial for almost any healthcare data mining, AI, medical analytics or predictive application. As opposed to other clinical information types, such as diagnoses (diseases/procedures), lab test results or even medications, clinical symptoms can only be recovered directly from written clinical narratives. Due to the high complexity, variability and difficulty in generating annotated corpora for clinical symptoms, only few large manually annotated data collections have been constructed so far, with certain underlying limitations in terms of a) entity linking / normalization of the symptom mentions to controlled vocabularies and b) a lack of attempts to promote the development of multilingual solutions and b) provide detailed annotation criteria and guidelines. To address these issues, we have posed the SYMPTEMIST track at the upcoming BioCreative VIII initiative, which will be run in the context of the prestigious AMIA 2023 conference, which received over 1400 submissions this year.

Automatic detection of symptoms mentions are key for a range of clinical use cases and real world applications like:

Predictive modeling of diseases

Differential diagnosis of complex diseases

Rare disease characterization & analysis

Selection of appropriate treatment & therapy

Study of disease-symptom associations

Early detection of disease outbreaks & epidemiological surveillance

Extraction of phenotypes

Drug repurposing & off label indications


The SYMPTEMIST organizers will also release multilingual resources to foster the development of multilingual tools and generate systems not only for Spanish but also for content in English and Romance languages (French, Portuguese, Italian, Romanian and Catalan) as well as versions in Dutch, Swedish and Czech.

Inspired by previous initiatives (e.g. n2c2, CLEF or TREC) and shared tasks (CANTEMIST, PharmaCoNER, or CodiEsp), we are launching the SYMPTEMIST shared task as part of the BioCreative 2023 evaluation initiative, with the following three sub-tracks:

SYMPTEMIST-entities: automatic detection of mentions of symptoms.

SYMPTEMIST-linking: finding mentions of symptoms and normalizing them to their Snomed-CT concept identifiers.

SYMPTEMIST-multilingual: automatic detection of mentions of symptoms in versions of the corpus generated in English, French, Italian, Portuguese, Romanian, Catalan, Dutch, Swedish and Czech.




BioCreative proceedings and AMIA workshop


Teams participating in SYMPTEMIST will be invited to contribute a systems description paper for the BioCreative 2023 Working Notes proceedings and a flash presentation of their approach at the BioCreative 2023 session. The BioCreative VIII workshop will run with AMIA 2023, November 11-15, 2023, In New Orleans, LA. See: https://amia.org/education-events/amia-2023-annual-symposium



Workshop Proceedings and Special Issue:


The BioCreative VIII Proceedings will host all the submissions from participating teams, and it will be freely available by the time of the workshop. In addition, we are happy to announce that the journal Database will host the BioCreative VIII special issue for work that has passed their peer-review process. Invitation to submit will be sent after the workshop.




All BioCreative VIII tracks

Track 1: BioRED (Biomedical Relation Extraction Dataset)

*Track 2: SYMPTEMIST (Symptom TExt Mining Shared Task)

Track 3: Genetic Phenotype Extraction and Normalization from Dysmorphology Physical Examination Entries

Track 4: Clinical Annotation Tool Track



Main Organizers

Martin Krallinger, Barcelona Supercomputing Center, Spain

Eulàlia Farré-Maduell, Barcelona Supercomputing Center, Spain

Luis Gascó, Barcelona Supercomputing Center, Spain

Salvador Lima, Barcelona Supercomputing Center, Spain

Jan Rodriguez, Barcelona Supercomputing Center, Spain