Discriminative ability for abnormal ventilatory function using acoustic pattern of inhale and exhale in patients...

Despite various available methods for monitoring a patient's respiratory system, conventional monitors provide limited ventilatory function information. This study explored the feasibility of intraoperative lung sound patterns to discriminate preoperatively diagnosed ventilatory dysfunction.

Forty-five patients who had undergone preoperative pulmonary function testing were enrolled for analysis, comprising 15 patients per normal, obstructive, and restrictive group. High-fidelity lung sounds were recorded intraoperatively using esophageal stethoscopes equipped with digital microphone devices. After signal processing, morphological features of the acoustic data, including the inhale/exhale peak ratio (I/Ep), were extracted. Their discriminative abilities for obstructive and restrictive types were assessed and compared with conventional monitoring parameters.

I/Ep showed strong discriminative performance, with an area under the receiver operating characteristic curve of 0.950 (95% CI: 0.887-0.991) for obstructive and 0.950 (95% CI: 0.867-0.995) for restrictive types. The median values of I/Ep were 2.9 in the restrictive, 2.2 in the normal, and 1.5 in the obstructive group (P < 0.001). Conventional ventilatory parameters (compliance, peak inspiratory pressure, and slope of end-tidal CO2) did not significantly differ among the groups.

Our study demonstrated that the I/Ep derived from intraoperative acoustic data differed according to preoperatively diagnosed ventilatory dysfunction. Hence, acoustic inhale and exhale patterns contain clinically useful information not captured by conventional ventilator parameters. Further studies are warranted to explore the clinical application of acoustic feature analysis for real-time intraoperative monitoring.