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

Tzung-Chien Hsieh; Hellen Lesmann; Alexander Hustinx; Shahida Moosa; Elaine Marchi; Maria Del Pilar Caro Martin; Ibrahim Abdelrazek; Jean Tori Pantel; Hannah Klinkhammer; Merle Ten Hagen; Meow-Keong Thong; Rifhan Azwani Binti Mazlan; Sok Kun Tae; Tom Kamphans; Wolfgang Meiswinkel; 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 Pinheiro; Chong Kim; Maria Melaragno; Hannah Weiland; Sophia Kaptain; Karolina Chwiałkowska; Miroslaw Kwasniewski; Ramy Saad; Sarah Wiethoff; Himanshu Goel; Clara Tang; Anna Hau; Tahsin Stefan Barakat; Przemysław Panek; Amira Nabil; Julia Suh; Frederik Braun; Israel Gomy; Luisa Averdunk; Ekanem Ekure; Gaber Bergant; Borut Peterlin; Claudio Graziano; Nagwa Gaboon; Moisés Fiesco-Roa; Alessandro Spinelli; Nina-Maria Wilpert; Prasit Phowthongkum; Nergis Güzel; Tobias Haack; Rana Bitar; Andreas Tzschach; Agusti Rodriguez-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; Nour Elkhateeb; Sheetal Kumar; Katalin Komlosi; Khoushoua Mohamed; Silvia Kalantari; Fabio Sirchia; Antonio Martinez-Monseny; Matthias Höller; Amal Mohamed; Amaia Lasa-Aranzasti; John Sayer; Nadja Ehmke; Magdalena Danyel; Henrike Sczakiel; Sarina Schwartzmann; Felix Boschann; Max Zhao; Ronja Adams; Lara Einicke; Kee Seang Chew; Choy Chen Kam; Miray Karakoyun; Ben Pode-Shakked; Aviva Eliyahu; Rachel Rock; Teresa Carrion; Odelia Chorin; Yuri Zarate; Marcello Martinez; Mert Karakaya; Moon Ley Tung; Bharatendu Chandra; Aimé Lumaka; Marwan Shinawi; Patrick Blackburn; Tianyun Wang; Tim Niehues; Ping Hu; Rebekah Waikel; Suzanna Ledgister Hanchard; Gehad Elmakkawy; Sylvia Safwat; Frédéric Ebstein; Elke Krüger; Sébastien Küry; Stephane Bezieau; Annabelle Arlt; Felix Marbach; Dong Li; Lucie Dupuis; Roberto Mendoza-Londono; Sofia Douzgou Houge; Denisa Weis; Christopher Mak; Nursel Elcioglu; Ayca Aykut; Peli Şimşek-Kiper; Nina Bögershausen; Bernd Wollnik; Heidi Beate Bentzen; Ingo Kurth; Christian Netzer; Aleksandra Jezela-Stanek; Koen Devriendt; Karen Gripp; Martin Mücke; Alain Verloes; Christian Schaaf; Christoffer Nellåker; Benjamin Solomon; Markus Nöthen; Ebtesam Abdalla; Gholson Lyon; Peter Krawitz; Hulya Kayserili; Louiza Toutouna; Axel Schmidt; Regina Roth; Dagmar Wieczorek; Eric Olinger

doi:10.21203/rs.3.rs-4438861/v1

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

Tzung-Chien Hsieh, Hellen Lesmann, Alexander Hustinx, Shahida Moosa, Elaine Marchi, Maria Del Pilar Caro Martin, Ibrahim Abdelrazek, Jean Tori Pantel, Hannah Klinkhammer, Merle Ten Hagen, Meow-Keong Thong, Rifhan Azwani Binti Mazlan, Sok Kun Tae, Tom Kamphans, Wolfgang Meiswinkel, Behnam Javanmardi, Alexej Knaus, Annette Uwineza, Cordula Knopp, Tinatin TkemaladzeMiriam Elbracht, Larissa Mattern, Rami Abou Jamra, Clara Velmans, Vincent Strehlow, Maureen Jacob, Angela Peron, Cristina Dias, Beatriz Nunes, Thainá Vilella, Isabel Pinheiro, Chong Kim, Maria Melaragno, Hannah Weiland, Sophia Kaptain, Karolina Chwiałkowska, Miroslaw Kwasniewski, Ramy Saad, Sarah Wiethoff, Himanshu Goel, Clara Tang, Anna Hau, Tahsin Stefan Barakat, Przemysław Panek, Amira Nabil, Julia Suh, Frederik Braun, Israel Gomy, Luisa Averdunk, Ekanem Ekure, Gaber Bergant, Borut Peterlin, Claudio Graziano, Nagwa Gaboon, Moisés Fiesco-Roa, Alessandro Spinelli, Nina-Maria Wilpert, Prasit Phowthongkum, Nergis Güzel, Tobias Haack, Rana Bitar, Andreas Tzschach, Agusti Rodriguez-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, Nour Elkhateeb, Sheetal Kumar, Katalin Komlosi, Khoushoua Mohamed, Silvia Kalantari, Fabio Sirchia, Antonio Martinez-Monseny, Matthias Höller, Amal Mohamed, Amaia Lasa-Aranzasti, John Sayer, Nadja Ehmke, Magdalena Danyel, Henrike Sczakiel, Sarina Schwartzmann, Felix Boschann, Max Zhao, Ronja Adams, Lara Einicke, Kee Seang Chew, Choy Chen Kam, Miray Karakoyun, Ben Pode-Shakked, Aviva Eliyahu, Rachel Rock, Teresa Carrion, Odelia Chorin, Yuri Zarate, Marcello Martinez, Mert Karakaya, Moon Ley Tung, Bharatendu Chandra, Aimé Lumaka, Marwan Shinawi, Patrick Blackburn, Tianyun Wang, Tim Niehues, Ping Hu, Rebekah Waikel, Suzanna Ledgister Hanchard, Gehad Elmakkawy, Sylvia Safwat, Frédéric Ebstein, Elke Krüger, Sébastien Küry, Stephane Bezieau, Annabelle Arlt, Felix Marbach, Dong Li, Lucie Dupuis, Roberto Mendoza-Londono, Sofia Douzgou Houge, Denisa Weis, Christopher Mak, Nursel Elcioglu, Ayca Aykut, Peli Şimşek-Kiper, Nina Bögershausen, Bernd Wollnik, Heidi Beate Bentzen, Ingo Kurth, Christian Netzer, Aleksandra Jezela-Stanek, Koen Devriendt, Karen Gripp, Martin Mücke, Alain Verloes, Christian Schaaf, Christoffer Nellåker, Benjamin Solomon, Markus Nöthen, Ebtesam Abdalla, Gholson Lyon^*, Peter Krawitz, Hulya Kayserili, Louiza Toutouna, Axel Schmidt, Regina Roth, Dagmar Wieczorek, Eric Olinger

^*Corresponding author for this work

Clinical Genetics

Research output: Contribution to journal › Article › Academic

22 Downloads (Pure)

Abstract

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 without decreasing the performance on European patients. By design, GMDB complies with the FAIR principles by rendering the curated medical data findable, accessible, interoperable, and reusable. This means GMDB can also serve as data for training and benchmarking. In summary, our study on facial dysmorphism on a global sample revealed a considerable cross ancestral phenotypic variability confounding NGP that should be counteracted by international efforts for increasing data diversity. GMDB will serve as a vital reference database for clinicians and a transparent training set for advancing NGP technology.

Original language	English
Journal	Research Square
DOIs	https://doi.org/10.21203/rs.3.rs-4438861/v1
Publication status	Published - 10 Jun 2024

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.21203/rs.3.rs-4438861/v1Licence: CC BY

GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseasesFinal published version, 4.58 MBLicence: CC BY

Cite this

Hsieh, T.-C., Lesmann, H., Hustinx, A., Moosa, S., Marchi, E., Martin, M. D. P. C., Abdelrazek, I., Pantel, J. T., Klinkhammer, H., Hagen, M. T., Thong, M.-K., Mazlan, R. A. B., Tae, S. K., Kamphans, T., Meiswinkel, W., Javanmardi, B., Knaus, A., Uwineza, A., Knopp, C., ... Olinger, E. (2024). GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases. Research Square. https://doi.org/10.21203/rs.3.rs-4438861/v1

@article{4292a01945784cc0aeab58f7f92e1eab,

title = "GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases",

abstract = "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 without decreasing the performance on European patients. By design, GMDB complies with the FAIR principles by rendering the curated medical data findable, accessible, interoperable, and reusable. This means GMDB can also serve as data for training and benchmarking. In summary, our study on facial dysmorphism on a global sample revealed a considerable cross ancestral phenotypic variability confounding NGP that should be counteracted by international efforts for increasing data diversity. GMDB will serve as a vital reference database for clinicians and a transparent training set for advancing NGP technology.",

author = "Tzung-Chien Hsieh and Hellen Lesmann and Alexander Hustinx and Shahida Moosa and Elaine Marchi and Martin, {Maria Del Pilar Caro} and Ibrahim Abdelrazek and Pantel, {Jean Tori} and Hannah Klinkhammer and Hagen, {Merle Ten} and Meow-Keong Thong and Mazlan, {Rifhan Azwani Binti} and Tae, {Sok Kun} and Tom Kamphans and Wolfgang Meiswinkel and Behnam Javanmardi and Alexej Knaus and Annette Uwineza and Cordula Knopp and Tinatin Tkemaladze and Miriam Elbracht and Larissa Mattern and Jamra, {Rami Abou} and Clara Velmans and Vincent Strehlow and Maureen Jacob and Angela Peron and Cristina Dias and Beatriz Nunes and Thain{\'a} Vilella and Isabel Pinheiro and Chong Kim and Maria Melaragno and Hannah Weiland and Sophia Kaptain and Karolina Chwia{\l}kowska and Miroslaw Kwasniewski and Ramy Saad and Sarah Wiethoff and Himanshu Goel and Clara Tang and Anna Hau and Barakat, {Tahsin Stefan} and Przemys{\l}aw Panek and Amira Nabil and Julia Suh and Frederik Braun and Israel Gomy and Luisa Averdunk and Ekanem Ekure and Gaber Bergant and Borut Peterlin and Claudio Graziano and Nagwa Gaboon and Mois{\'e}s Fiesco-Roa and Alessandro Spinelli and Nina-Maria Wilpert and Prasit Phowthongkum and Nergis G{\"u}zel and Tobias Haack and Rana Bitar and Andreas Tzschach and Agusti Rodriguez-Palmero and Theresa Brunet and Sabine Rudnik-Sch{\"o}neborn and Silvina Contreras-Capetillo and Ava Oberlack and Carole Samango-Sprouse and Teresa Sadeghin and Margaret Olaya and Konrad Platzer and Artem Borovikov and Franziska Schnabel and Lara Heuft and Vera Herrmann and Nour Elkhateeb and Sheetal Kumar and Katalin Komlosi and Khoushoua Mohamed and Silvia Kalantari and Fabio Sirchia and Antonio Martinez-Monseny and Matthias H{\"o}ller and Amal Mohamed and Amaia Lasa-Aranzasti and John Sayer and Nadja Ehmke and Magdalena Danyel and Henrike Sczakiel and Sarina Schwartzmann and Felix Boschann and Max Zhao and Ronja Adams and Lara Einicke and Chew, {Kee Seang} and Kam, {Choy Chen} and Miray Karakoyun and Ben Pode-Shakked and Aviva Eliyahu and Rachel Rock and Teresa Carrion and Odelia Chorin and Yuri Zarate and Marcello Martinez and Mert Karakaya and Tung, {Moon Ley} and Bharatendu Chandra and Aim{\'e} Lumaka and Marwan Shinawi and Patrick Blackburn and Tianyun Wang and Tim Niehues and Ping Hu and Rebekah Waikel and Hanchard, {Suzanna Ledgister} and Gehad Elmakkawy and Sylvia Safwat and Fr{\'e}d{\'e}ric Ebstein and Elke Kr{\"u}ger and S{\'e}bastien K{\"u}ry and Stephane Bezieau and Annabelle Arlt and Felix Marbach and Dong Li and Lucie Dupuis and Roberto Mendoza-Londono and Houge, {Sofia Douzgou} and Denisa Weis and Christopher Mak and Nursel Elcioglu and Ayca Aykut and Peli {\c S}im{\c s}ek-Kiper and Nina B{\"o}gershausen and Bernd Wollnik and Bentzen, {Heidi Beate} and Ingo Kurth and Christian Netzer and Aleksandra Jezela-Stanek and Koen Devriendt and Karen Gripp and Martin M{\"u}cke and Alain Verloes and Christian Schaaf and Christoffer Nell{\aa}ker and Benjamin Solomon and Markus N{\"o}then and Ebtesam Abdalla and Gholson Lyon and Peter Krawitz and Hulya Kayserili and Louiza Toutouna and Axel Schmidt and Regina Roth and Dagmar Wieczorek and Eric Olinger",

year = "2024",

month = jun,

day = "10",

doi = "10.21203/rs.3.rs-4438861/v1",

language = "English",

}

Hsieh, T-C, Lesmann, H, Hustinx, A, Moosa, S, Marchi, E, Martin, MDPC, Abdelrazek, I, Pantel, JT, Klinkhammer, H, Hagen, MT, Thong, M-K, Mazlan, RAB, Tae, SK, Kamphans, T, Meiswinkel, W, Javanmardi, B, Knaus, A, Uwineza, A, Knopp, C, Tkemaladze, T, Elbracht, M, Mattern, L, Jamra, RA, Velmans, C, Strehlow, V, Jacob, M, Peron, A, Dias, C, Nunes, B, Vilella, T, Pinheiro, I, Kim, C, Melaragno, M, Weiland, H, Kaptain, S, Chwiałkowska, K, Kwasniewski, M, Saad, R, Wiethoff, S, Goel, H, Tang, C, Hau, A, Barakat, TS, Panek, P, Nabil, A, Suh, J, Braun, F, Gomy, I, Averdunk, L, Ekure, E, Bergant, G, Peterlin, B, Graziano, C, Gaboon, N, Fiesco-Roa, M, Spinelli, A, Wilpert, N-M, Phowthongkum, P, Güzel, N, Haack, T, Bitar, R, Tzschach, A, Rodriguez-Palmero, A, Brunet, T, Rudnik-Schöneborn, S, Contreras-Capetillo, S, Oberlack, A, Samango-Sprouse, C, Sadeghin, T, Olaya, M, Platzer, K, Borovikov, A, Schnabel, F, Heuft, L, Herrmann, V, Elkhateeb, N, Kumar, S, Komlosi, K, Mohamed, K, Kalantari, S, Sirchia, F, Martinez-Monseny, A, Höller, M, Mohamed, A, Lasa-Aranzasti, A, Sayer, J, Ehmke, N, Danyel, M, Sczakiel, H, Schwartzmann, S, Boschann, F, Zhao, M, Adams, R, Einicke, L, Chew, KS, Kam, CC, Karakoyun, M, Pode-Shakked, B, Eliyahu, A, Rock, R, Carrion, T, Chorin, O, Zarate, Y, Martinez, M, Karakaya, M, Tung, ML, Chandra, B, Lumaka, A, Shinawi, M, Blackburn, P, Wang, T, Niehues, T, Hu, P, Waikel, R, Hanchard, SL, Elmakkawy, G, Safwat, S, Ebstein, F, Krüger, E, Küry, S, Bezieau, S, Arlt, A, Marbach, F, Li, D, Dupuis, L, Mendoza-Londono, R, Houge, SD, Weis, D, Mak, C, Elcioglu, N, Aykut, A, Şimşek-Kiper, P, Bögershausen, N, Wollnik, B, Bentzen, HB, Kurth, I, Netzer, C, Jezela-Stanek, A, Devriendt, K, Gripp, K, Mücke, M, Verloes, A, Schaaf, C, Nellåker, C, Solomon, B, Nöthen, M, Abdalla, E, Lyon, G, Krawitz, P, Kayserili, H, Toutouna, L, Schmidt, A, Roth, R, Wieczorek, D & Olinger, E 2024, 'GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases', Research Square. https://doi.org/10.21203/rs.3.rs-4438861/v1

TY - JOUR

T1 - GestaltMatcher Database - A global reference for facial phenotypic variability in rare human diseases

AU - Hsieh, Tzung-Chien

AU - Lesmann, Hellen

AU - Hustinx, Alexander

AU - Moosa, Shahida

AU - Marchi, Elaine

AU - Martin, Maria Del Pilar Caro

AU - Abdelrazek, Ibrahim

AU - Pantel, Jean Tori

AU - Klinkhammer, Hannah

AU - Hagen, Merle Ten

AU - Thong, Meow-Keong

AU - Mazlan, Rifhan Azwani Binti

AU - Tae, Sok Kun

AU - Kamphans, Tom

AU - Meiswinkel, Wolfgang

AU - Javanmardi, Behnam

AU - Knaus, Alexej

AU - Uwineza, Annette

AU - Knopp, Cordula

AU - Tkemaladze, Tinatin

AU - Elbracht, Miriam

AU - Mattern, Larissa

AU - Jamra, Rami Abou

AU - Velmans, Clara

AU - Strehlow, Vincent

AU - Jacob, Maureen

AU - Peron, Angela

AU - Dias, Cristina

AU - Nunes, Beatriz

AU - Vilella, Thainá

AU - Pinheiro, Isabel

AU - Kim, Chong

AU - Melaragno, Maria

AU - Weiland, Hannah

AU - Kaptain, Sophia

AU - Chwiałkowska, Karolina

AU - Kwasniewski, Miroslaw

AU - Saad, Ramy

AU - Wiethoff, Sarah

AU - Goel, Himanshu

AU - Tang, Clara

AU - Hau, Anna

AU - Barakat, Tahsin Stefan

AU - Panek, Przemysław

AU - Nabil, Amira

AU - Suh, Julia

AU - Braun, Frederik

AU - Gomy, Israel

AU - Averdunk, Luisa

AU - Ekure, Ekanem

AU - Bergant, Gaber

AU - Peterlin, Borut

AU - Graziano, Claudio

AU - Gaboon, Nagwa

AU - Fiesco-Roa, Moisés

AU - Spinelli, Alessandro

AU - Wilpert, Nina-Maria

AU - Phowthongkum, Prasit

AU - Güzel, Nergis

AU - Haack, Tobias

AU - Bitar, Rana

AU - Tzschach, Andreas

AU - Rodriguez-Palmero, Agusti

AU - Brunet, Theresa

AU - Rudnik-Schöneborn, Sabine

AU - Contreras-Capetillo, Silvina

AU - Oberlack, Ava

AU - Samango-Sprouse, Carole

AU - Sadeghin, Teresa

AU - Olaya, Margaret

AU - Platzer, Konrad

AU - Borovikov, Artem

AU - Schnabel, Franziska

AU - Heuft, Lara

AU - Herrmann, Vera

AU - Elkhateeb, Nour

AU - Kumar, Sheetal

AU - Komlosi, Katalin

AU - Mohamed, Khoushoua

AU - Kalantari, Silvia

AU - Sirchia, Fabio

AU - Martinez-Monseny, Antonio

AU - Höller, Matthias

AU - Mohamed, Amal

AU - Lasa-Aranzasti, Amaia

AU - Sayer, John

AU - Ehmke, Nadja

AU - Danyel, Magdalena

AU - Sczakiel, Henrike

AU - Schwartzmann, Sarina

AU - Boschann, Felix

AU - Zhao, Max

AU - Adams, Ronja

AU - Einicke, Lara

AU - Chew, Kee Seang

AU - Kam, Choy Chen

AU - Karakoyun, Miray

AU - Pode-Shakked, Ben

AU - Eliyahu, Aviva

AU - Rock, Rachel

AU - Carrion, Teresa

AU - Chorin, Odelia

AU - Zarate, Yuri

AU - Martinez, Marcello

AU - Karakaya, Mert

AU - Tung, Moon Ley

AU - Chandra, Bharatendu

AU - Lumaka, Aimé

AU - Shinawi, Marwan

AU - Blackburn, Patrick

AU - Wang, Tianyun

AU - Niehues, Tim

AU - Hu, Ping

AU - Waikel, Rebekah

AU - Hanchard, Suzanna Ledgister

AU - Elmakkawy, Gehad

AU - Safwat, Sylvia

AU - Ebstein, Frédéric

AU - Krüger, Elke

AU - Küry, Sébastien

AU - Bezieau, Stephane

AU - Arlt, Annabelle

AU - Marbach, Felix

AU - Li, Dong

AU - Dupuis, Lucie

AU - Mendoza-Londono, Roberto

AU - Houge, Sofia Douzgou

AU - Weis, Denisa

AU - Mak, Christopher

AU - Elcioglu, Nursel

AU - Aykut, Ayca

AU - Şimşek-Kiper, Peli

AU - Bögershausen, Nina

AU - Wollnik, Bernd

AU - Bentzen, Heidi Beate

AU - Kurth, Ingo

AU - Netzer, Christian

AU - Jezela-Stanek, Aleksandra

AU - Devriendt, Koen

AU - Gripp, Karen

AU - Mücke, Martin

AU - Verloes, Alain

AU - Schaaf, Christian

AU - Nellåker, Christoffer

AU - Solomon, Benjamin

AU - Nöthen, Markus

AU - Abdalla, Ebtesam

AU - Lyon, Gholson

AU - Krawitz, Peter

AU - Kayserili, Hulya

AU - Toutouna, Louiza

AU - Schmidt, Axel

AU - Roth, Regina

AU - Wieczorek, Dagmar

AU - Olinger, Eric

PY - 2024/6/10

Y1 - 2024/6/10

N2 - 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 without decreasing the performance on European patients. By design, GMDB complies with the FAIR principles by rendering the curated medical data findable, accessible, interoperable, and reusable. This means GMDB can also serve as data for training and benchmarking. In summary, our study on facial dysmorphism on a global sample revealed a considerable cross ancestral phenotypic variability confounding NGP that should be counteracted by international efforts for increasing data diversity. GMDB will serve as a vital reference database for clinicians and a transparent training set for advancing NGP technology.

AB - 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 without decreasing the performance on European patients. By design, GMDB complies with the FAIR principles by rendering the curated medical data findable, accessible, interoperable, and reusable. This means GMDB can also serve as data for training and benchmarking. In summary, our study on facial dysmorphism on a global sample revealed a considerable cross ancestral phenotypic variability confounding NGP that should be counteracted by international efforts for increasing data diversity. GMDB will serve as a vital reference database for clinicians and a transparent training set for advancing NGP technology.

U2 - 10.21203/rs.3.rs-4438861/v1

DO - 10.21203/rs.3.rs-4438861/v1

M3 - Article

C2 - 38903062

SN - 2693-5015

JO - Research Square

JF - Research Square

ER -

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

Abstract

UN SDGs

Access to Document

Fingerprint

Cite this