Contactless screening for sleep apnea with breathing vibration signals based on modified U-Net

BACKGROUND

Obstructive sleep apnea (OSA) is a chronic sleep disorder characterized by frequent cessations or reductions of breathing during sleep. Polysomnography (PSG) is a definitive diagnostic tool for OSA. The costly and obtrusive nature of PSG and poor access to sleep clinics have created a demand for accurate home-based screening devices.

METHODS

This paper proposes a novel OSA screening method based solely on breathing vibration signals with a modified U-Net, allowing patients to be tested at home. Sleep recordings over a whole night are collected in a contactless manner, and sleep apnea-hypopnea events are labeled by a deep neural network. The apnea-hypopnea index (AHI) calculated from events estimation is then used to screen for the apnea. The performance of the model is tested by event-based analysis and comparing the estimated AHI with the manually obtained values.

RESULTS

The accuracy and sensitivity of sleep apnea events detection are 97.5% and 76.4%, respectively. The mean absolute error of AHI estimation for the patients is 3.0 events/hour. The correlation between the ground truth AHI and predicted AHI shows an R² of 0.95. In addition, 88.9% of all participants are classified into correct AHI categories.

CONCLUSIONS

The proposed scheme has great potential as a simple screening tool for sleep apnea. It can accurately detect potential OSA and help the patients to be referred for differential diagnosis of home sleep apnea test (HSAT) or polysomnographic evaluation.

Clinical application of a previously validated pregnancy-specific screening tool for sleep apnea in a cohort with a high prevalence of obesity

Objective

The purpose of this project was to determine the positive predictive value of existing obstructive sleep apnea (OSA) screening tools in clinical use, in a real-world clinical population of gravidae, and to explore the development of a new questionnaire for screening for OSA during pregnancy.

Methods

Pregnant people were administered sleep screening questionnaires as part of routine clinical care. These included Facco's four variable OSA screening tool, the STOP-BANG, and the Epworth Sleepiness Scale. Those who screened positive were referred for diagnostic sleep testing, typically with a type III home monitoring device. Here we analyzed the screening responses used by those who completed diagnostic testing to determine the positive predictive value of the existing tools.

Results

159 pregnant people completed diagnostic OSA testing and were included in this analysis. The positive predictive value of Facco's four variable sleep screening tool was 74.3%, STOP-BANG was 75.3%, and the Epworth Sleepiness Scale was 69.8%. Our sample size was insufficient to create a new screening tool.

Conclusions

Here we calculated the positive predictive value of Facco's 4 variable screening tool for screening for OSA in pregnancy in a real-world pregnant population. While we were not able to generate a new screening tool for screening for OSA during pregnancy, both STOP-BANG and Facco's four variable tool had positive predictive values over 70% in our population which was characterized by high BMI and advanced maternal age. Increased clinical use of the pregnancy-specific tool may be warranted.

Feasibility of automated detection of sleep apnea using implantable pacemakers and defibrillators: a comparison with simultaneous polysomnography recording

PURPOSE

Sleep-related breathing disorders (SRBD) have been associated with adverse cardiovascular events and prognosis. Some modern pacemakers (PMK) and implantable cardioverter defibrillators (ICD) are equipped with thoracic impedance monitoring systems which allow detecting fluctuations in tidal volume occurring in SRBD. This study aimed at assessing the accuracy of a novel transthoracic impedance-based PMK/ICD sensor for automatic detection of SRBD.

METHODS

We prospectively enrolled 31 consecutive patients scheduled for implantation of a PMK or an ICD with sleep respiratory disturbances monitoring capability. The results of unattended home nocturnal recordings expressed as apnea/hypopnea index (AHI) were compared with the respiratory disturbance index (RDI) stored in the cardiac device memory at the time polygraphy was performed.

RESULTS

Among the 31 enrolled patients (20 men, mean age 70 ± 8 years), a comparison between the data from polygraphy and cardiac device was feasible in 29 patients. According to the results of home nocturnal recording, a severe SRBD was observed in 7/29 (24.1%) patients. On the basis of receiver-operating characteristic curve analysis of RDI values, the optimal RDI cutoff value to identify severe SRBD was 47 episodes/h (sensitivity 100%, specificity 100%). Bland-Altman agreement analysis of AHI-RDI revealed a bias between measurements of - 11 episodes/h, with limits of agreement - 38 to 15 episodes/h.

CONCLUSIONS

In the present study, the novel transthoracic impedance-based monitoring system ApneaScan^TM appeared effective in screening PMK and ICD patients for SRBD.

Comparison of a simple obstructive sleep apnea screening device with standard in-laboratory polysomnography

BACKGROUND

Obstructive sleep apnea (OSA) is a common underdiagnosed sleep disorder. Various strategies have been employed to easily screen for OSA. The ApneaStrip® (AS - S.L.P. Ltd, Tel Aviv, Israel) is an FDA approved OSA screening device applied to the upper lip at home. We evaluated the performance of this device against simultaneous in-laboratory polysomnography (PSG) in a group of well-characterized OSA patients.

METHODS

Diagnostic PSG was performed in 56 patients (29 M, 37 F; age 48.9 ± 14.6 years; body mass index [BMI] 37.5 ± 9.0 kg/m(2); apnea-hypopnea index-events/h-[AHI] 32.8 ± 22.9). The AS was applied and positioned to detect nasal and oral airflow. The AS gives a "positive" result for AHI ≥ 15. We examined the sensitivity and specificity of the AS against three thresholds derived from PSG: AHI ≥ 5, AHI ≥ 15 (company recommendation), and AHI ≥ 30.

RESULTS

For PSG AHI ≥ 15, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the AS were 80, 54.5, 87.8, and 40 %, respectively. For PSG AHI ≥ 5, the values were 75.1, 66.7, 97.1, and 13.3 %, respectively. For PSG AHI ≥ 30, the values were 86.9, 36.2, 48.8, and 80 %, respectively. There were no significant modifying effects of age, BMI, gender, hypertension, diabetes, lung disease, and heart disease.

CONCLUSION

The AS has a high sensitivity for detection of OSA with AHI ≥ 15, but only modest specificity. The AS could be a useful component of an OSA screening program; however, negative results should be interpreted cautiously.