Scientific article
Open access

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

Published innpj digital medicine, vol. 6, no. 1, 104
Publication date2023-06-02
First online date2023-06-02

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.

Citation (ISO format)
HEITMANN, Julien et al. DeepBreath-automated detection of respiratory pathology from lung auscultation in 572 pediatric outpatients across 5 countries. In: npj digital medicine, 2023, vol. 6, n° 1, p. 104. doi: 10.1038/s41746-023-00838-3
Main files (1)
Article (Published version)
Secondary files (2)
ISSN of the journal2398-6352

Technical informations

Creation08/17/2023 3:15:30 PM
First validation10/02/2023 7:56:33 AM
Update time10/02/2023 7:56:33 AM
Status update10/02/2023 7:56:33 AM
Last indexation02/01/2024 10:44:12 AM
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack