DeepBreath-automated detection of respiratory pathology from lung auscultation in 572 pediatric outpatients across 5 countries

Article Article scientifique Article DeepBreath-automated detection of respiratory pathology from lung auscultation in 572 pediatric outpatients across 5 countries eng 2023-06-02 2023-06-02 author Heitmann Julien Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Glangetas Alban Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 779424 author Doenz Jonathan 0000-0001-6047-9488 Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Dervaux Juliane Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Shama Deeksha M Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Garcia Daniel Hinjos Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Benissa Mohamed-Rida Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 836994 author Cantais Aymeric 0000-0002-8650-2129 Pediatric Emergency Department, Hospital University of Saint Etienne, Saint Etienne, France. author Perez Alexandre 0000-0003-1454-4327 Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 658747 author Müller Daniel 0000-0003-2019-233X Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Chavdarova Tatjana Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Ruchonnet Isabelle 0000-0003-4771-875X Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 106509 Département de Pédiatrie, Gynécologie et Obstétrique (PETRI) author Siebert Johan 0000-0002-2871-3382 Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 507348 Département de Pédiatrie, Gynécologie et Obstétrique (PETRI) author Lacroix Laurence Elisabeth Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 84201 Département de Pédiatrie, Gynécologie et Obstétrique (PETRI) author Jaggi Martin Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. author Gervaix Alain Division of Pediatric Emergency Medicine, Department of Women, Child and Adolescent, Geneva University Hospitals (HUG), University of Geneva, Switzerland, Geneva, Switzerland. 31715 Département de Pédiatrie, Gynécologie et Obstétrique (PETRI) author Hartley Mary-Anne 0000-0002-8826-3870 Intelligent Global Health Research Group, Machine Learning and Optimization Laboratory, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. mary-anne.hartley@epfl.ch.,Center for Intelligent Systems (CIS), Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. mary-anne.hartley@epfl.ch.,Division of Pediatric Emergency Medicine, Department of Pediatrics, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland. mary-anne.hartley@epfl.ch. Pneumoscope Study Group collaborator Hugon Florence Anne-Marie 74225 104 10.1038/s41746-023-00838-3 37268730 2398-6352 PMC10238513 The interpretation of lung auscultation is highly subjective and relies on non-specific nomenclature. Computer-aided analysis has the potential to better standardize and automate evaluation. We used 35.9 hours of auscultation audio from 572 pediatric outpatients to develop DeepBreath : a deep learning model identifying the audible signatures of acute respiratory illness in children. It comprises a convolutional neural network followed by a logistic regression classifier, aggregating estimates on recordings from eight thoracic sites into a single prediction at the patient-level. Patients were either healthy controls (29%) or had one of three acute respiratory illnesses (71%) including pneumonia, wheezing disorders (bronchitis/asthma), and bronchiolitis). To ensure objective estimates on model generalisability, DeepBreath is trained on patients from two countries (Switzerland, Brazil), and results are reported on an internal 5-fold cross-validation as well as externally validated (extval) on three other countries (Senegal, Cameroon, Morocco). DeepBreath differentiated healthy and pathological breathing with an Area Under the Receiver-Operator Characteristic (AUROC) of 0.93 (standard deviation [SD] ± 0.01 on internal validation). Similarly promising results were obtained for pneumonia (AUROC 0.75 ± 0.10), wheezing disorders (AUROC 0.91 ± 0.03), and bronchiolitis (AUROC 0.94 ± 0.02). Extval AUROCs were 0.89, 0.74, 0.74 and 0.87 respectively. All either matched or were significant improvements on a clinical baseline model using age and respiratory rate. Temporal attention showed clear alignment between model prediction and independently annotated respiratory cycles, providing evidence that DeepBreath extracts physiologically meaningful representations. DeepBreath provides a framework for interpretable deep learning to identify the objective audio signatures of respiratory pathology. Dewey 618 58 Département de pédiatrie, gynécologie et obstétrique 223 false 380 Bochaton-Piallat Marie-Luce (pathologie et immunologie) 646 149 Etiologie des pneumonies et marqueurs inflammatoires chez l'enfant fébrile 183 npj digital medicine 6 1 64f718be-b669-4bed-b208-b7633279b82c Article (Published version) Public Article-scientifique-from-unpaywall.pdf 2530235 application/pdf CC BY-4.0 d70a864cdcd171fcfcbea54938ad9baa841bc925f765975c112efb51bdf3b817 4c63eba9-440c-48d5-bc36-6a4c31f8fabd Appendix Public 41746_2023_838_MOESM1_ESM.pdf 273319 application/pdf fe1d8f138a5f2b32c625cae2e5054d1eb5bf21305e64d9fb0c930447af793b86 a46b3a7c-aad3-4119-843d-4bbf66c18d9d Appendix Public 41746_2023_838_MOESM2_ESM.pdf 2145306 application/pdf 29d82dff3cd01e05104fec83ac83fc863ff360b8adff802d749f6cf8e6fb5188 openAccess Faculté de médecine / Section de médecine clinique / Département de pédiatrie, gynécologie et obstétrique

Title	Published in	Year	Views	Downloads

DeepBreath-automated detection of respiratory pathology from lung auscultation in 572 pediatric outpatients across 5 countries	npj digital medicine	2023	161	64
Immediate Implants in the esthetic area: Our perspective and guidelines	Journal of Oral Science & Rehabilitation	2018	542	12

Archive ouverte UNIGE

Perez, Alexandre