TBscreen: A passive cough classifier for tuberculosisscreening with a controlled dataset
TBscreen: A passive cough classifier for tuberculosisscreening with a controlled dataset
Abstract
Recent respiratory disease screening studies suggest promising performance of cough classifiers, but potentialbiases in model training and dataset quality preclude robust conclusions. To examine tuberculosis (TB) cough di-agnostic features, we enrolled subjects with pulmonary TB (N = 149) and controls with other respiratory illnesses(N = 46) in Nairobi. We collected a dataset with 33,000 passive coughs and 1600 forced coughs in a controlledsetting with similar demographics. We trained a ResNet18-based cough classifier using images of passive coughscalogram as input and obtained a fivefold cross-validation sensitivity of 0.70 (±0.11 SD). The smartphone-basedmodel had better performance in subjects with higher bacterial load {receiver operating characteristic–area un-der the curve (ROC-AUC): 0.87 [95% confidence interval (CI): 0.87 to 0.88], P < 0.001} or lung cavities [ROC-AUC:0.89 (95% CI: 0.88 to 0.89), P < 0.001]. Overall, our data suggest that passive cough features distinguish TB fromnon-TB subjects and are associated with bacterial burden and disease severity.
