Mapping biomedical terminologies using natural language processing tools and UMLS: mapping the Orphanet thesaurus to the MeSH

Background: Orphanet aims to provide rare disease information to healthcare professionals, patients, and their relatives. Objective: The objective of this work is to evaluate two methodologies (UMLS and manual Orphanet-ICD-10 link-based mapping & String Based matching) used to map Orphanet thesaurus to the MeSH thesaurus. Results: On a corpus of 375 mappings, the string based matching provides significantly better results than the UMLS and manual Orphanet-ICD-10 link-based mapping. Conclusion: String based matching could be applied to any biomedical terminology in French not yet included into UMLS. Mapping biomedical terminologies using natural language processing tools and UMLS: mapping the Orphanet thesaurus to the MeSH Tayeb Merabti, MS 1, 2 , Michel Joubert, PhD 2 , Thierry Lecroq 1 , PhD, A. Rath, MD 3 , Stefan J. Darmoni, MD, PhD 1 1 CISMeF, University Hospital, Rouen, France & TIBS, LITIS EA 4108, Institute of Biomedical Research, University of Rouen, France; 2 LERTIM EA 3283, Faculty of Medicine, Marseille, France; 3 Orphanet INSERM SC11, Paris, France Abstract Background: Orphanet aims to provide rare disease information to healthcare professionals, patients, and their relatives. Objective: The objective of this work is to evaluate two methodologies (UMLS and manual OrphanetICD-10 link-based mapping & String Based matching) used to map Orphanet thesaurus to the MeSH thesaurus. Results: On a corpus of 375 mappings, the string based matching provides significantly better results than the UMLS and manual Orphanet-ICD-10 link-based mapping. Conclusion: String based matching could be applied to any biomedical terminology in French not yet included into UMLS.Background: Orphanet aims to provide rare disease information to healthcare professionals, patients, and their relatives. Objective: The objective of this work is to evaluate two methodologies (UMLS and manual OrphanetICD-10 link-based mapping & String Based matching) used to map Orphanet thesaurus to the MeSH thesaurus. Results: On a corpus of 375 mappings, the string based matching provides significantly better results than the UMLS and manual Orphanet-ICD-10 link-based mapping. Conclusion: String based matching could be applied to any biomedical terminology in French not yet included into UMLS.