GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases

作者：Hellen Lesmann, Alexander Hustinx, Shahida Moosa, Hannah Klinkhammer, Elaine Marchi, Pilar Caro, Ibrahim M. Abdelrazek, Jean Tori Pantel, Merle ten Hagen, Meow‐Keong Thong, Rifhan Azwani Mazlan, Sok Kun Tae, Tom Kamphans, Wolfgang Meiswinkel, Jingmei Li, Behnam Javanmardi, Alexej Knaus, Annette Uwineza, Cordula Knopp, Tinatin Tkemaladze, Miriam Elbracht, Larissa Mattern, Rami Abou Jamra, Clara Velmans, Vincent Strehlow, Maureen Jacob, Angela Peron, Cristina Dias, Beatriz Nunes, Thainá Vilella, Isabel Furquim Pinheiro, Chong Ae Kim, Maria Isabel Melaragno, Hannah Weiland, Sophia Kaptain, Karolina Chwiałkowska, Mirosław Kwaśniewski, Ramy Saad, Sarah Wiethoff, Himanshu Goel, Clara Sze-Man Tang, Anna Hau, Tahsin Stefan Barakat, Przemysław Panek, Amira Nabil, Julia Suh, Frederik Braun, Israel Gomy, Luisa Averdunk, Ekanem N. Ekure, Gaber Bergant, Borut Peterlin, Claudio Graziano, Nagwa E. A. Gaboon, Moisés Ó. Fiesco-Roa, Alessandro Spinelli, Nina‐Maria Wilpert, Prasit Phowthongkum, Nergis Güzel, Tobias B. Haack, Rana Bitar, Andreas Tzschach, Agustí Rodríguez‐Palmero, Theresa Brunet, Sabine Rudnik‐Schöneborn, Silvina Contreras‐Capetillo, Ava Oberlack, Carole Samango‐Sprouse, Teresa Sadeghin, Margaret Olaya, Konrad Platzer, Artem Borovikov, Franziska Schnabel, Lara Heuft, Vera Herrmann, Renske Oegema, Nour Elkhateeb, Sheetal Kumar, Katalin Komlósi, Khoushoua Mohamed, Silvia Kalantari, Fabio Sirchia, Antonio F. Martinez-Monseny, Matthias Höller, Louiza Toutouna, Amal Mohamed, Amaia Lasa‐Aranzasti, John A. Sayer, Nadja Ehmke, Magdalena Danyel, Henrike L. Sczakiel, Sarina Schwartzmann, Felix Boschann, Max Zhao, R. Adam, Lara Einicke, Denise Horn, Kee Seang Chew, KAM Choy Chen, Miray Karakoyun, Ben Pode‐Shakked, Aviva Eliyahu, Rachel Rock, Teresa Carrion, Odelia Chorin, Yuri A. Zárate, Marcelo Martinez Conti, Mert Karakaya, Moon Ley Tung, Bharatendu Chandra, Arjan Bouman, Aimé Lumaka, Naveed Wasif, Marwan Shinawi, Patrick R. Blackburn, Tianyun Wang, Tim Niehues, Axel Schmidt, Regina Roth, Dagmar Wieczorek, Ping Hu, Rebekah L. Waikel, Suzanna E. Ledgister Hanchard, Gehad Elmakkawy, Sylvia Safwat, Frédéric Ebstein, Elke Krüger, Sébastien Küry, Stéphane Bezieau, Annabelle Arlt, Eric Olinger, Felix Marbach, Dong Li, Lucie Dupuis, Roberto Mendoza‐Londono, Sofia Douzgou, Denisa Weis, Brian Hon‐Yin Chung, Christopher C.Y. Mak, Hülya Kayserili, Nursel Elçioğlu, Ayça Aykut, Peli Özlem Şimşek-Kiper, Nina Bögershausen, Bernd Wollnik, Heidi Beate Bentzen, Ingo Kurth, Christian Netzer, Aleksandra Jezela‐Stanek, Koenraad Devriendt, Karen W. Gripp, Martin Mücke, Alain Verloès, Christian P. Schaaf, Christoffer Nellåker, Benjamin D. Solomon, Markus M. Nöthen, Ebtesam Abdalla, Gholson J. Lyon, Peter Krawitz, Tzung‐Chien Hsieh · 发表于：medRxiv · 年份：2023 · DOI：10.1101/2023.06.06.23290887 · 被引用次数：25 · 研究领域：AI in cancer detection、Cutaneous Melanoma Detection and Management、Genomic variations and chromosomal abnormalities

The most important factor that complicates the work of dysmorphologists is the significant phenotypic variability of the human face. Next-Generation Phenotyping (NGP) tools that assist clinicians with recognizing characteristic syndromic patterns are particularly challenged when confronted with patients from populations different from their training data. To that end, we systematically analyzed the impact of genetic ancestry on facial dysmorphism. For that purpose, we established the GestaltMatcher Database (GMDB) as a reference dataset for medical images of patients with rare genetic disorders from around the world. We collected 10,980 frontal facial images - more than a quarter previously unpublished - from 8,346 patients, representing 581 rare disorders. Although the predominant ancestry is still European (67%), data from underrepresented populations have been increased considerably via global collaborations (19% Asian and 7% African). This includes previously unpublished reports for more than 40% of the African patients. The NGP analysis on this diverse dataset revealed characteristic performance differences depending on the composition of training and test sets corresponding to genetic relatedness. For clinical use of NGP, incorporating non-European patients resulted in a profound enhancement of GestaltMatcher performance. The top-5 accuracy rate increased by +11.29%. Importantly, this improvement in delineating the correct disorder from a facial portrait was achieved wi...