Validity and reliability of acoustic analysis of respiratory sounds in infants

Validation of automatic wheeze detection in patients with obstructed airways and in healthy subjects

Computerized lung-sound analysis is a sensitive and quantitative method to identify wheezing by its typical pattern on spectral analysis. We evaluated the accuracy of the VRI, a multi-sensor, computer-based device with an automated technique of wheeze detection. The method was validated in 100 sound files from seven subjects with asthma or chronic obstructive pulmonary disease and seven healthy subjects by comparison of auscultation findings, examination of audio files, and computer detection of wheezes. Three blinded physicians identified 40 sound files with wheezes and 60 sound files without wheezes. Sensitivity and specificity were 83% and 85%, respectively. Negative predictive value and positive predictive value were 89% and 79%, respectively. Overall inter-rater agreement was 84%. False positive cases were found to contain sounds that simulate wheezes, such as background noises with high frequencies or strong noises from the throat that could be heard and identified without a stethoscope. The present findings demonstrate that the wheeze detection algorithm has good accuracy, sensitivity, specificity, negative predictive value and positive predictive value for wheeze detection in regional analyses with a single sensor and multiple sensors. Results are similar to those reported in the literature. The device is user-friendly, requires minimal patient effort, and, distinct from other devices, it provides a dynamic image of breath sound distribution with wheeze detection output in less than 1 minute.

Nasal continuous positive airway pressure decreases respiratory muscles overload in young infants with severe acute viral bronchiolitis

OBJECTIVE

To determine the efficacy of nasal continuous positive airway pressure (nCPAP) on respiratory distress symptoms and respiratory effort in young infants with acute respiratory syncytial virus bronchiolitis.

DESIGN

Prospective study.

SETTING

The paediatric intensive care unit of a university hospital.

PATIENTS

Twelve infants less than 3 months of age, with severe respiratory distress.

INTERVENTIONS

Respiratory distress was quantified with a specific scoring system. Oesophageal pressure (Pes) was measured during spontaneous ventilation before and after nCPAP, delivered through an infant-adapted ventilator. Simultaneous recording of gastric pressure (Pgas) was performed in the five oldest patients.

MEASUREMENTS AND RESULTS

The respiratory distress score decreased after nCPAP, particularly accessory muscles' use and expiratory wheezing. The breathing pattern was modified, with shorter inspiratory and longer expiratory time. Pes swings and PTPes(insp), two indices of inspiratory effort, were reduced by 54 (+/-4)% and 59 (+/-5)%. PTPgas(exp), an indicator of expiratory muscles activity, was completely abolished. A significant correlation was observed between the respiratory distress score and Pes swings at baseline and after nCPAP.

CONCLUSIONS

In young infants with severe acute respiratory syncytial virus bronchiolitis, nCPAP rapidly unloads respiratory muscles and improves respiratory distress symptoms.

Analysis of respiratory sounds: state of the art

OBJECTIVE

This paper describes state of the art, scientific publications and ongoing research related to the methods of analysis of respiratory sounds.

METHODS AND MATERIAL

Review of the current medical and technological literature using Pubmed and personal experience.

RESULTS

The study includes a description of the various techniques that are being used to collect auscultation sounds, a physical description of known pathologic sounds for which automatic detection tools were developed. Modern tools are based on artificial intelligence and on technics such as artificial neural networks, fuzzy systems, and genetic algorithms…

CONCLUSION

The next step will consist in finding new markers so as to increase the efficiency of decision aid algorithms and tools.

Reproducibility of dynamically represented acoustic lung images from healthy individuals

Background and aim:

Acoustic lung imaging offers a unique method for visualising the lung. This study was designed to demonstrate reproducibility of acoustic lung images recorded from healthy individuals at different time points and to assess intra- and inter-rater agreement in the assessment of dynamically represented acoustic lung images.

Methods:

Recordings from 29 healthy volunteers were made on three separate occasions using vibration response imaging. Reproducibility was measured using quantitative, computerised assessment of vibration energy. Dynamically represented acoustic lung images were scored by six blinded raters.

Results:

Quantitative measurement of acoustic recordings was highly reproducible with an intraclass correlation score of 0.86 (very good agreement). Intraclass correlations for inter-rater agreement and reproducibility were 0.61 (good agreement) and 0.86 (very good agreement), respectively. There was no significant difference found between the six raters at any time point. Raters ranged from 88% to 95% in their ability to identically evaluate the different features of the same image presented to them blinded on two separate occasions.

Conclusion:

Acoustic lung imaging is reproducible in healthy individuals. Graphic representation of lung images can be interpreted with a high degree of accuracy by the same and by different reviewers.

Dynamic visualization of lung sounds with a vibration response device: a case series

BACKGROUND

The field of computer-assisted mapping of lung sounds is constantly evolving and several devices have been developed in this field.

OBJECTIVES

Our objective was to evaluate a new computer-assisted lung sound imaging system, 'vibration response imaging' (VRI), that records and creates a dynamic image of breath sounds. We postulated that the VRI display format would qualitatively and quantitatively reveal breath sound distribution throughout the breathing cycle.

METHODS

Lung sounds were recorded from 5 healthy adults and 14 patients with various respiratory illnesses using VRI. The lung sounds were processed by the VRI software, which incorporates an algorithm to convert breath sounds in the frequency range of 150-250 Hz to a dynamic image and quantitative assessment of breath sound distribution.

RESULTS

Images and quantifications from recordings of the healthy adults showed distinct patterns for inspiration and expiration. Images and quantifications from the subjects with respiratory illness differed substantially from the images of the healthy subjects. Both healthy and pathological subjects presented some expected characteristics of breath sound distribution.

CONCLUSIONS

The VRI device may provide a new perspective in acoustic imaging and quantification of breath sounds by adding aspects of time analysis and quantification of distribution to existing methods. Further studies will be required in order to establish reliability of repeated recordings and to validate the sensitivity of the system in detecting various lung pathologies.

Bronchiolitis

Bronchiolitis is a distressing, potentially life-threatening respiratory condition that affects young babies. Around 2-3% of all infants younger than 1 year are admitted to hospital with bronchiolitis, usually during the seasonal epidemic. The majority of these infants are infected with respiratory syncytial virus and all have an intense inflammatory response in their airways. Although most infants recover, they have an increased risk of recurrent wheezing. Although bronchiolitis is common, little is known about what causes infants to be susceptible. Diagnostic interventions have little effect on clinical outcome, and apart from supportive measures, there is no specific treatment. Bronchiolitis therefore presents an intriguing clinical conundrum and a major challenge to researchers. High quality clinical studies are needed to clarify assessment of disease severity and criteria for hospital admission, particularly the use of pulse oximetry and chest radiography. Careful mapping of the inflammatory pathways in the pathogenesis of bronchiolitis should lead to development of new therapies to alleviate symptoms.

Response to bronchodilator in infants with bronchiolitis can be predicted from wheeze characteristics

OBJECTIVE

Lung sounds analysis has been used for clinical care. Our objectives were to characterize the spectral pattern of lung sounds and their relation to bronchodilator effects in acute bronchiolitis (AB). We hypothesized that patients with sinusoidal wheezes (SW) would show a more significant bronchodilator response.

METHODOLOGY

We studied 22 asleep hospitalized infants (14 boys, eight girls), aged 5.2 +/- 1 months, 16 with a positive respiratory syncytial virus test, during their first 3 days after admission. Patients breathed spontaneously through a face mask connected to a pneumotachograph during normal breathing, and only target flows of 0.1 +/- 0.02 L/s were analyzed. Sounds were obtained using two contact sensors attached over both posterior lower lobes. For inspiratory and expiratory sounds, we determined the frequencies below which 25% (F25), 50% (F50), 75% (F75) and 99% (SEF99) of the spectral power between 100 and 1000 Hz was contained. We repeated the measurements 20 min after bronchodilator therapy in all patients.

RESULTS

We found classic SW in 11 patients, while the other 11 had complex wheezes (CW). There were positive bronchodilator responses in 9/11 with SW and 3/11 with CW (P < 0.01). Patients who responded to salbutamol showed an increase in power at low frequencies after medication (P < 0.01), and a positive correlation between wheezing and the increase in the power spectra measured by F50 and SEF99 (P < 0.001).

CONCLUSIONS

We conclude that sinusoidal and complex wheezes occur in patients with AB, that a positive response to bronchodilator is significantly more common in those with classic SW and that lung sounds analysis is a reproducible, safe and non-invasive method for assessing wheeze in infants.