GestaltMatcher is a continuously updated collection of medical images of individuals with rare diseases and open-source AIs for the interpretation of such data.[1][2] As of March 2023, GestaltMatcher DataBase (GMDB) contained approximately 10,000 case reports with a molecular diagnosis and clinical features annotated with HPO terminology.[3] Medical images include, for example, facial photographs of patients with genetic syndromes manifesting with facial dysmorphic features, as well as radiographs from those with skeletal dysplasias.
GestaltMatcher allows users to find and publish case reports, including medical images, if that option is chosen in the dynamic consent module. By that means, GMDB complements medRxiv and can also be used as a repository for re-identifiable images in preprints.
In a prospective three year multi center study, GestaltMatcher showed clinical utility as an artificial expert opinion in a multidisciplinary team.[4]
History
editThe GestaltMatcher project started in April 2021 during the revision of the manuscript from Hsieh, et al.[5] with funding from University of Bonn and the German Research Foundation (DFG). The reviewers and editors of Nature Genetics asked for FAIR data in order to reproduce the algorithmic results described in that work. Since then, the database (GMDB) has grown by contributions from its community. Since January 2022, GMDB can be used as repository for medical imaging data for preprints submitted to medRxiv. In February 2023, at the 14th ICHG meeting in Cape Town, Prof. Shahida Moosa (Stellenbosch University) reported the 10,000 case, which is a patient from South Africa with Mabry syndrome. Prof. Peter Krawitz also announced at the conference that AGD e.V., a German non-profit organization, will oversee the GMDB from this point forward. In January 2024 the GestaltMatcher project received a donation from the Eva Luise und Horst Köhler Stiftung, which is a charity of the former German president Horst Köhler and his wife, Eva Köhler, to improve the medica care for people with rare diseases.
References
edit- ^ Hustinx, Alexander; Hellmann, Fabio; Sümer, Ömer; Javanmardi, Behnam; André, Elisabeth; Krawitz, Peter; Hsieh, Tzung-Chien (2023). "Improving Deep Facial Phenotyping for Ultra-rare Disorder Verification Using Model Ensembles". 2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV). pp. 5018–5028. arXiv:2211.06764. doi:10.1109/WACV56688.2023.00499. ISBN 978-1-6654-9346-8. S2CID 253510768.
- ^ Sümer, Ömer; Hellmann, Fabio; Hustinx, Alexander; Hsieh, Tzung-Chien; André, Elisabeth; Krawitz, Peter (2022-10-23). "Few-Shot Meta Learning for Recognizing Facial Phenotypes of Genetic Disorders". arXiv:2210.12705 [cs.CV].
- ^ Lesmann, Hellen; Hustinx, Alexander; Moosa, Shahida; Klinkhammer, Hannah; Marchi, Elaine; Caro, Pilar; Abdelrazek, Ibrahim M.; Pantel, Jean Tori; Hagen, Merle ten (2024-05-21), GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases, doi:10.1101/2023.06.06.23290887, PMC 10371103, retrieved 2024-07-30
- ^ Schmidt, Axel; Danyel, Magdalena; Grundmann, Kathrin; Brunet, Theresa; Klinkhammer, Hannah; Hsieh, Tzung-Chien; Engels, Hartmut; Peters, Sophia; Knaus, Alexej; Moosa, Shahida; Averdunk, Luisa; Boschann, Felix; Sczakiel, Henrike Lisa; Schwartzmann, Sarina; Mensah, Martin Atta (2024-07-22). "Next-generation phenotyping integrated in a national framework for patients with ultrarare disorders improves genetic diagnostics and yields new molecular findings". Nature Genetics: 1–10. doi:10.1038/s41588-024-01836-1. ISSN 1546-1718.
- ^ Hsieh, Tzung-Chien; Bar-Haim, Aviram; Moosa, Shahida; Ehmke, Nadja; Gripp, Karen W.; Pantel, Jean Tori; Danyel, Magdalena; Mensah, Martin Atta; Horn, Denise; Rosnev, Stanislav; Fleischer, Nicole; Bonini, Guilherme; Hustinx, Alexander; Schmid, Alexander; Knaus, Alexej (March 2022). "GestaltMatcher facilitates rare disease matching using facial phenotype descriptors". Nature Genetics. 54 (3): 349–357. doi:10.1038/s41588-021-01010-x. ISSN 1061-4036. PMC 9272356. PMID 35145301.