Nicholas L Rider1,2,3, Gina Cahill1,2, Tina Motazedi4, Lei Wei3, Ashok Kurian3, Lenora M Noroski1,2, Filiz O Seeborg1,2, Ivan K Chinn1,2, Kirk Roberts5. 1. Department of Pediatrics, Baylor College of Medicine, Houston, Texas, United States of America. 2. Section of Immunology, Allergy and Retrovirology, Texas Children's Hospital, Houston, Texas, United States of America. 3. Department of Information Services, Texas Children's Hospital, Houston, Texas, United States of America. 4. Division of Allergy and Immunology, Massachusetts General Hospital, Boston, Massachusetts, United States of America. 5. The University of Texas School of Biomedical Informatics, Houston, Texas, United States of America.
Abstract
BACKGROUND: Primary immunodeficiency diseases represent an expanding set of heterogeneous conditions which are difficult to recognize clinically. Diagnostic rates outside of the newborn period have not changed appreciably. This concern underscores a need for novel methods of disease detection. OBJECTIVE: We built a Bayesian network to provide real-time risk assessment about primary immunodeficiency and to facilitate prescriptive analytics for initiating the most appropriate diagnostic work up. Our goal is to improve diagnostic rates for primary immunodeficiency and shorten time to diagnosis. We aimed to use readily available health record data and a small training dataset to prove utility in diagnosing patients with relatively rare features. METHODS: We extracted data from the Texas Children's Hospital electronic health record on a large population of primary immunodeficiency patients (n = 1762) and appropriately-matched set of controls (n = 1698). From the cohorts, clinically relevant prior probabilities were calculated enabling construction of a Bayesian network probabilistic model(PI Prob). Our model was constructed with clinical-immunology domain expertise, trained on a balanced cohort of 100 cases-controls and validated on an unseen balanced cohort of 150 cases-controls. Performance was measured by area under the receiver operator characteristic curve (AUROC). We also compared our network performance to classic machine learning model performance on the same dataset. RESULTS: PI Prob was accurate in classifying immunodeficiency patients from controls (AUROC = 0.945; p<0.0001) at a risk threshold of ≥6%. Additionally, the model was 89% accurate for categorizing validation cohort members into appropriate International Union of Immunological Societies diagnostic categories. Our network outperformed 3 other machine learning models and provides superior transparency with a prescriptive output element. CONCLUSION: Artificial intelligence methods can classify risk for primary immunodeficiency and guide management. PI Prob enables accurate, objective decision making about risk and guides the user towards the appropriate diagnostic evaluation for patients with recurrent infections. Probabilistic models can be trained with small datasets underscoring their utility for rare disease detection given appropriate domain expertise for feature selection and network construction.
BACKGROUND:Primary immunodeficiency diseases represent an expanding set of heterogeneous conditions which are difficult to recognize clinically. Diagnostic rates outside of the newborn period have not changed appreciably. This concern underscores a need for novel methods of disease detection. OBJECTIVE: We built a Bayesian network to provide real-time risk assessment about primary immunodeficiency and to facilitate prescriptive analytics for initiating the most appropriate diagnostic work up. Our goal is to improve diagnostic rates for primary immunodeficiency and shorten time to diagnosis. We aimed to use readily available health record data and a small training dataset to prove utility in diagnosing patients with relatively rare features. METHODS: We extracted data from the Texas Children's Hospital electronic health record on a large population of primary immunodeficiencypatients (n = 1762) and appropriately-matched set of controls (n = 1698). From the cohorts, clinically relevant prior probabilities were calculated enabling construction of a Bayesian network probabilistic model(PI Prob). Our model was constructed with clinical-immunology domain expertise, trained on a balanced cohort of 100 cases-controls and validated on an unseen balanced cohort of 150 cases-controls. Performance was measured by area under the receiver operator characteristic curve (AUROC). We also compared our network performance to classic machine learning model performance on the same dataset. RESULTS: PI Prob was accurate in classifying immunodeficiencypatients from controls (AUROC = 0.945; p<0.0001) at a risk threshold of ≥6%. Additionally, the model was 89% accurate for categorizing validation cohort members into appropriate International Union of Immunological Societies diagnostic categories. Our network outperformed 3 other machine learning models and provides superior transparency with a prescriptive output element. CONCLUSION: Artificial intelligence methods can classify risk for primary immunodeficiency and guide management. PI Prob enables accurate, objective decision making about risk and guides the user towards the appropriate diagnostic evaluation for patients with recurrent infections. Probabilistic models can be trained with small datasets underscoring their utility for rare disease detection given appropriate domain expertise for feature selection and network construction.
Authors: Vicki Modell; Bonnie Gee; David B Lewis; Jordan S Orange; Chaim M Roifman; John M Routes; Ricardo U Sorensen; Luigi D Notarangelo; Fred Modell Journal: Immunol Res Date: 2011-10 Impact factor: 2.829
Authors: Paul Arora; Devon Boyne; Justin J Slater; Alind Gupta; Darren R Brenner; Marek J Druzdzel Journal: Value Health Date: 2019-03-15 Impact factor: 5.725
Authors: Alexander Stojadinovic; Anton Bilchik; David Smith; John S Eberhardt; Elizabeth Ben Ward; Aviram Nissan; Eric K Johnson; Mladjan Protic; George E Peoples; Itzhak Avital; Scott R Steele Journal: Ann Surg Oncol Date: 2012-08-17 Impact factor: 5.344
Authors: Asbjørg Stray-Pedersen; Hanne Sørmo Sorte; Pubudu Samarakoon; Tomasz Gambin; Ivan K Chinn; Zeynep H Coban Akdemir; Hans Christian Erichsen; Lisa R Forbes; Shen Gu; Bo Yuan; Shalini N Jhangiani; Donna M Muzny; Olaug Kristin Rødningen; Ying Sheng; Sarah K Nicholas; Lenora M Noroski; Filiz O Seeborg; Carla M Davis; Debra L Canter; Emily M Mace; Timothy J Vece; Carl E Allen; Harshal A Abhyankar; Philip M Boone; Christine R Beck; Wojciech Wiszniewski; Børre Fevang; Pål Aukrust; Geir E Tjønnfjord; Tobias Gedde-Dahl; Henrik Hjorth-Hansen; Ingunn Dybedal; Ingvild Nordøy; Silje F Jørgensen; Tore G Abrahamsen; Torstein Øverland; Anne Grete Bechensteen; Vegard Skogen; Liv T N Osnes; Mari Ann Kulseth; Trine E Prescott; Cecilie F Rustad; Ketil R Heimdal; John W Belmont; Nicholas L Rider; Javier Chinen; Tram N Cao; Eric A Smith; Maria Soledad Caldirola; Liliana Bezrodnik; Saul Oswaldo Lugo Reyes; Francisco J Espinosa Rosales; Nina Denisse Guerrero-Cursaru; Luis Alberto Pedroza; Cecilia M Poli; Jose L Franco; Claudia M Trujillo Vargas; Juan Carlos Aldave Becerra; Nicola Wright; Thomas B Issekutz; Andrew C Issekutz; Jordan Abbott; Jason W Caldwell; Diana K Bayer; Alice Y Chan; Alessandro Aiuti; Caterina Cancrini; Eva Holmberg; Christina West; Magnus Burstedt; Ender Karaca; Gözde Yesil; Hasibe Artac; Yavuz Bayram; Mehmed Musa Atik; Mohammad K Eldomery; Mohammad S Ehlayel; Stephen Jolles; Berit Flatø; Alison A Bertuch; I Celine Hanson; Victor W Zhang; Lee-Jun Wong; Jianhong Hu; Magdalena Walkiewicz; Yaping Yang; Christine M Eng; Eric Boerwinkle; Richard A Gibbs; William T Shearer; Robert Lyle; Jordan S Orange; James R Lupski Journal: J Allergy Clin Immunol Date: 2016-07-16 Impact factor: 10.793
Authors: Ahmed Hosny; Chintan Parmar; Thibaud P Coroller; Patrick Grossmann; Roman Zeleznik; Avnish Kumar; Johan Bussink; Robert J Gillies; Raymond H Mak; Hugo J W L Aerts Journal: PLoS Med Date: 2018-11-30 Impact factor: 11.069