Asthma severity identification from pulmonary acoustic signal for computerized decision support system

Area of Science:

• Respiratory Medicine
• Biomedical Engineering
• Acoustics

Background:

• Breath sounds offer insights into respiratory pathologies.
• Wheeze sounds are indicative of underlying conditions in subjects.
• Asthma severity assessment is crucial for effective patient management.

Purpose of the Study:

• To classify asthma severity (Mild, Moderate, Severe) using frequency features of wheeze sounds.
• To analyze the behavior of wheeze sounds across different auscultation locations and breath phases.
• To evaluate the efficacy of frequency-based features in differentiating asthma acuteness.

Main Methods:

• Wheeze sounds were collected from 55 asthmatic patients during tidal breathing.
• Sounds were segmented and categorized into nine datasets based on location and breath phase.
• Frequency features (F25, F50, F75, F90, F99, Mean Frequency) were extracted and analyzed using multivariate statistics.

Main Results:

• Frequency features demonstrated statistical significance (p < 0.05) in differentiating asthma severity across most datasets.
• Higher effect sizes were observed for frequency features derived from tracheal wheeze sounds.
• The analysis confirmed significant differences in wheeze sound characteristics based on auscultation location and breath phase.

Conclusions:

• Computerized analysis of wheeze sounds effectively identifies asthma severity levels during tidal breathing.
• Auscultation location and breath phases significantly influence wheeze sound characteristics.
• Frequency-based features hold potential for objective asthma assessment.