Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Leveraging cough sounds to optimize chest x-ray usage in low-resource settings

Alexander Philip, Sanya Chawla, Lola Jover, George P. Kafentzis, Joe Brew, Vishakh Saraf, Shibu Vijayan, Peter Small, Carlos Chaccour

TL;DR

This study investigates whether acoustic features from solicited coughs can predict abnormal chest radiographs to optimize chest X-ray use in low-resource settings. Using prospectively collected data from 137 patients (967 coughs) and ground-truth labels from a validated AI chest X-ray system, the authors extract frame-level and cough-level acoustic features and evaluate three classifiers via stratified grouped 4-fold cross-validation. Results show ROC-AUC values ranging from 0.70 to 0.78, with logistic regression performing best at about 0.73, indicating potential for cough-based triage to reduce unnecessary imaging. While promising, the approach requires larger, multi-site validation and potential integration with additional clinical signals to achieve robust deployment in LMIC healthcare workflows.

Abstract

Chest X-ray is a commonly used tool during triage, diagnosis and management of respiratory diseases. In resource-constricted settings, optimizing this resource can lead to valuable cost savings for the health care system and the patients as well as to and improvement in consult time. We used prospectively-collected data from 137 patients referred for chest X-ray at the Christian Medical Center and Hospital (CMCH) in Purnia, Bihar, India. Each patient provided at least five coughs while awaiting radiography. Collected cough sounds were analyzed using acoustic AI methods. Cross-validation was done on temporal and spectral features on the cough sounds of each patient. Features were summarized using standard statistical approaches. Three models were developed, tested and compared in their capacity to predict an abnormal result in the chest X-ray. All three methods yielded models that could discriminate to some extent between normal and abnormal with the logistic regression performing best with an area under the receiver operating characteristic curves ranging from 0.7 to 0.78. Despite limitations and its relatively small sample size, this study shows that AI-enabled algorithms can use cough sounds to predict which individuals presenting for chest radiographic examination will have a normal or abnormal results. These results call for expanding this research given the potential optimization of limited health care resources in low- and middle-income countries.

Leveraging cough sounds to optimize chest x-ray usage in low-resource settings

TL;DR

This study investigates whether acoustic features from solicited coughs can predict abnormal chest radiographs to optimize chest X-ray use in low-resource settings. Using prospectively collected data from 137 patients (967 coughs) and ground-truth labels from a validated AI chest X-ray system, the authors extract frame-level and cough-level acoustic features and evaluate three classifiers via stratified grouped 4-fold cross-validation. Results show ROC-AUC values ranging from 0.70 to 0.78, with logistic regression performing best at about 0.73, indicating potential for cough-based triage to reduce unnecessary imaging. While promising, the approach requires larger, multi-site validation and potential integration with additional clinical signals to achieve robust deployment in LMIC healthcare workflows.

Abstract

Chest X-ray is a commonly used tool during triage, diagnosis and management of respiratory diseases. In resource-constricted settings, optimizing this resource can lead to valuable cost savings for the health care system and the patients as well as to and improvement in consult time. We used prospectively-collected data from 137 patients referred for chest X-ray at the Christian Medical Center and Hospital (CMCH) in Purnia, Bihar, India. Each patient provided at least five coughs while awaiting radiography. Collected cough sounds were analyzed using acoustic AI methods. Cross-validation was done on temporal and spectral features on the cough sounds of each patient. Features were summarized using standard statistical approaches. Three models were developed, tested and compared in their capacity to predict an abnormal result in the chest X-ray. All three methods yielded models that could discriminate to some extent between normal and abnormal with the logistic regression performing best with an area under the receiver operating characteristic curves ranging from 0.7 to 0.78. Despite limitations and its relatively small sample size, this study shows that AI-enabled algorithms can use cough sounds to predict which individuals presenting for chest radiographic examination will have a normal or abnormal results. These results call for expanding this research given the potential optimization of limited health care resources in low- and middle-income countries.
Paper Structure (14 sections, 7 equations, 3 figures, 1 table)

This paper contains 14 sections, 7 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methods
  3. Patient Enrollment and Data Capture
  4. Digital Chest X-ray Interpretation using a Validated AI System
  5. Cough Analysis
  6. Machine Learning Models
  7. Logistic Regression
  8. Support Vector Machine
  9. Multi-Layer Perceptron
  10. Results
  11. Enrollment
  12. Performance Metrics
  13. Discussion
  14. Conclusion

Figures (3)

  • Figure 1: Mel Frequency Cepstral Coefficients of coughs from patients with (left) abnormal CXR and (right) normal CXR.
  • Figure 2: Two-dimensional tSNE plot of feature vectors. Yellow dots correspond to individuals with abnormal Xray while purple dots correspond to the ones with normal Xray.
  • Figure 3: Graphical depiction of the stratified grouped four-fold cross validation. Groups are created as per the top most bar and AI models validated with each sub-sample (fold), avoiding any overlap between the train and test groups.