CHEST Annual Meeting

CHEST Annual Meeting

Source:

Ashraf O, et al. Phenotyping and Management of COPD. Presented at: CHEST Annual Meeting 2020; Oct. 18-22, 2020 (virtual meeting).

Disclosures: Ashraf reports no relevant financial disclosures.
October 21, 2020
1 min read
Save

Voice-based screening for chronic respiratory conditions shows promise

Source:

Ashraf O, et al. Phenotyping and Management of COPD. Presented at: CHEST Annual Meeting 2020; Oct. 18-22, 2020 (virtual meeting).

Disclosures: Ashraf reports no relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

In a new study, automated voice and breath analysis was associated with good predictive accuracy of FVC and FEV1 and classification of obstructed vs. nonobstructed lung disease.

Using voice as a biomarker enables noninvasive detection of lung function and addresses some of the drawbacks of spirometry, such as need for specialized equipment, need for skilled technicians and cross-contamination, Obaid Ashraf, MD, pulmonologist in the Allegheny Health Network at Allegheny General Hospital, Pittsburgh, said during a presentation at the virtual CHEST Annual Meeting.

Source: Adobe Stock.

The prospective cohort study included 128 participants (58% women) aged 18 to 85 years who provided pre- and post-pulmonary function test voice and breath sound samples, which corresponded with pre-and post-bronchodilator samples, during routine visits at Allegheny General Hospital in Pennsylvania. The researchers obtained pre- and post-pulmonary function test FEV1 and FVC results from each patient. Participants’ voice and breath audio samples were recorded on a smart tablet, using a proprietary software app, and then the researchers analyzed the recorded data using cloud-based software (Telling.ai).

According to the results, automated voice and breath analysis delivered good diagnostic accuracy for the prediction of FEV1 (R squared, 0.74; mean squared error, 0.17; mean absolute error, 0.32; binarized accuracy, 71.5%) and FVC (R squared, 0.79; mean squared error, 0.19; mean absolute error, 0.35; binarized accuracy, 71.4%). The obstruction classifier an accuracy of 98%, sensitivity of 96% and F1 score of 0.9, which Ashraf said is consistent with precision and power of the test.

The next steps in this research, Ashraf said, will focus on patients with chronic obstructive pulmonary disease.

“This is a study that says we can use an AI machine-learning system as an alternative to spirometry to provide a personalized, convenient, very much cost-effective and easily approachable monitoring to more people,” Ashraf said.

Reference:

Ashraf O, et al. Chest. 2020;doi:10.1016/j.chest.2020.08.1509.