Poster Presentation
IEEE Signal Processing in Medicine and Biology Symposium 2019
Strados Labs: An Efficient Process to Acquire and Characterize Clinically Validated Respiratory System Information Using a Non-Invasive Biosensor
Y. K. Au, N. Capp, A. Cardenas, V. Fauveau, P. Glasser, G. Hassen, T. Muqueem
What’s Inside?
This study describes how the RESP® Biosensor system captures lung sound and chest wall motion data continuously and non-invasively, to provide objective respiratory biomarkers for both outpatient and inpatient care. It reports on a clinical feasibility trial involving over 51 patients with respiratory disease at NYC Health + Hospitals, where the device recorded nearly 24 hours of lung sound data and collected more than 4,500 annotated respiratory events.
Using professional annotation plus validation by medically-trained clinicians, the study demonstrates that the system can automatically label respiratory events, such as cough, wheeze, crackles, and other adventitious breath sounds, with high sensitivity (≈ 93%) and specificity (≈ 97%). The device’s continuous monitoring, compared to episodic auscultation, shows promise for better tracking of disease progression and more timely detection of changes in respiratory health.
