Method For Obstructive Sleep Apnea Research Articles

A novel deep learning model for obstructive sleep apnea diagnosis: hybrid CNN-Transformer approach for radar-based detection of apnea-hypopnea events.

The demand for cost-effective and accessible alternatives to polysomnography (PSG), the conventional diagnostic method for obstructive sleep apnea (OSA), has surged. In this study, we have developed and validated a deep learning model for detecting apnea-hypopnea events using radar data. We conducted a single-center prospective cohort study, dividing participants with suspected sleep-disordered breathing into development and temporally independent test sets. Utilizing a hybrid CNN-Transformer architecture, we performed fivefold cross-validation on the development set to develop and subsequently validate the model. Evaluation metrics included sensitivity for event detection, mean absolute error (MAE), intraclass correlation coefficient (ICC), and Pearson correlation coefficient (r) for apnea-hypopnea index (AHI) estimation. Linearly weighted kappa statistics (κ) assessed OSA severity. The development set comprised 54 participants (July 2021-May 2022), while the test set included 35 participants (June 2022-June 2023). In the test set, our model achieved an event detection sensitivity of 67.2% (95% CI = 65.8% to 68.5%) and demonstrated a MAE of 7.54 (95% CI = 5.36 to 9.72), indicating good agreement (ICC = 0.889 [95% CI = 0.792 to 0.942]) and a strong correlation (r = 0.892 [95% CI = 0.795 to 0.945]) with the ground truth for AHI estimation. Furthermore, OSA severity estimation showed substantial agreement (κ = 0.780 [95% CI = 0.658 to 0.903]). Our study highlights radar sensors and advanced AI models' potential to improve OSA diagnosis, paving the path for future radar-based diagnostic models in sleep medicine research.

Ten years of hypoglossal nerve stimulation in obstructive sleep apnea: a systematic literature review

To show the importance of hypoglossal nerve stimulation (HGNS) as a treatment method for obstructive sleep apnea (OSA) in the German healthcare context and to better assess the way patients who do not receive adequate care could benefit from HGNS. A systematic literature review in the Medline and Cochrane Library literature database was conducted, including publications using different stimulation technologies for HGNS. The efficacy of HGNS was assessed based on patient-relevant outcomes (daytime sleepiness, quality of life), treatment adherence and the apnea-hypopnea index (AHI) and oxygen desaturation index (ODI). The safety of the treatment method was assessed based on adverse events (AEs). Inclusion and analysis of 33 publications: 2 randomized controlled trials (RCTs, level Ib), 1 level IIb trial (n = 1) and 30 level IV trials with a study duration of up to 60 months. The RCTs showed better values for daytime sleepiness and quality of life when using HGNS than in the control group. AHI and ODI showed a deterioration under placebo stimulation or therapy withdrawal in the RCTs. Consistently high adherence was also reported in the long-term course. Severe AEs under HGNS were rare and could usually be resolved by repositioning electrodes or replacing device components. Other AEs were mostly transient or could be resolved by non-invasive measures. All investigated parameters showed similar results in the evaluated studies. The results of different stimulation systems are comparable in type and extent. The comprehensive review of the literature shows consistent data that highlight the importance of HGNS as an effective and safe treatment for OSA after unsuccessful CPAP treatment. The evaluation also shows that the different stimulation systems make it possible to better tailor the therapy to the patient's individual requirements. A future systematic evaluation of real-world data on the use of HGNS would help gain additional insights into the relevance of the method in routine clinical practice.

Nocturnal pulse oximetry as a possible screening method for obstructive sleep apnea in infants with laryngomalacia

ObjectiveLaryngomalacia can be an important cause of obstructive sleep apnea (OSA) in infants. Nocturnal oximetry is a cheap and safe method in comparison to polysomnography for the detection of sleep-disordered breathing. The aim of this study is to evaluate the validity of nocturnal oximetry as a diagnostic tool for OSA in infants with laryngomalacia. MethodsThis retrospective study included infants with laryngomalacia and a clinical suspicion of OSA who underwent a polysomnography at the Antwerp University Hospital. The oximetry was rescored manually, blinded to the polysomnography results, according to four different scoring methods. An obstructive apnea–hypopnea index (oAHI) ≥ 2/h on polysomnography was used to define OSA. ResultsThis study included 53 patients with laryngomalacia (51% boys, mean age 3.72 ± 0.26 months). A diagnosis of OSA was established in 46 patients (87%) by polysomnography. Among the four different scoring methods, the scoring according to Brouillette et al., yielded the highest diagnostic accuracy with a sensitivity and specificity of 91% and 25% respectively and with a negative and positive predictive value of 25% and 91%, respectively. Correlations and the Bland–Altman plot showed a wide limit of agreement for laboratory polysomnography oAHI and nocturnal oximetry ODI. ConclusionOur data show that overnight pulse oximetry has a high sensitivity and PPV to diagnose OSA in infants with laryngomalacia. However, the low specificity and NPV indicate that PSG is still needed to exclude OSA in cases with normal oximetry.

Polygraphy as a new screening method for obstructive sleep apnea in patients with atrial fibrillation?

Abstract Background Obstructive sleep apnea (OSA) is a modifiable risk factor of atrial fibrillation (AF) but is underdiagnosed in these patients. According to the 2020 European Society of Cardiology guidelines for the diagnosis and management of AF, it seems reasonable to test for OSA in symptomatic AF patients before initiation of rhythm control therapy. A good OSA screening pathway is lacking in the guidelines and in daily clinical practice in most centers today. Polysomnography (PSG) is the gold standard for diagnosing OSA but it is not an ideal screening method, due to its expensiveness, labour-intensiveness and limited availability. Purpose The aim of this study was to validate and examine the performance of different (cardiorespiratory) polygraphy (PG) devices, used with their automated algorithm to detect OSA in an AF cohort, compared with PSG as gold standard. Methods Prospective study with three PGs (ApneaLink Air (ALA), SOMNOtouch RESP (STR) and SpiderSAS (SpS) combined with ResMed Airview-, SomnoMedics DOMINO- and Microport SYNESCOPE software respectively) in consecutive AF patients who were referred for a PSG evaluation. Patients received the three devices with instructions for use after their PSG and were requested to use each one during three consecutive nights at home. Severity of OSA was classified according to the Apnea-Hypopnea Index (AHI) with <5 events/hour= no OSA, 5–14= mild-, 15–30= moderate-, >30= severe OSA. Results A total of 100 AF patients presenting for PSG were included (64.0±8.6 years, 73% male, 87% non-permanent AF, mean Body Mass Index 30.6±5.9 kg/m2, mean CHA2DS2-VASc score 2.4±1.7, mEHRA≥2 in 64%; mean AF history 5.4±5.6 years). PSG diagnosed ≥mild OSA in 90% of patients, ≥moderate in 69%, and severe OSA in 33%. Successful PG execution to determine the AHI, i.e. ≥4 hours of data on automated analysis, was obtained in 72.0%, 73.0% and 79.0% of patients with the ALA, STR and SpS respectively. In screening for clinically relevant OSA (i.e. at least moderate OSA), the PGs had an area under the curve of 0.802, 0.761 and 0.803 for the ALA, STR and SpS respectively (Figure 1). Conclusions This first analysis shows that PGs combined with an automated algorithm for AHI determination, can be used as OSA screening tools in AF patients before referral for PSG. Further optimisation is likely by choosing an appropriate AHI cut-off value for each PG. Funding Acknowledgement Type of funding sources: Public Institution(s). Main funding source(s): This study is supported by the Antwerp University Cardiology Research Fund, and is part of Limburg Clinical Research Center, supported by the foundation Limburg Sterk Merk, province of Limburg, Flemish government, Hasselt University, Ziekenhuis Oost-Limburg and Jessa Hospital. Figure 1

Tracheal sound analysis for detection of sleep disordered breathing

This study aimed to investigate whether the combination of tracheal sounds and nasal pressure sensor (TS-NP) allows the same detection of apneas and hypopneas as the recommended American Academy of Sleep Medicine (AASM) method for obstructive sleep apnea (OSA) diagnosis, i.e., the combination of an oronasal thermistor and nasal pressure sensor (Therm-NP). Polysomnographic recordings of 33 patients (6 females; mean age 52.9 ± 10.3 years; mean body mass index, BMI, 30.0 ± 5.2 kg/m2) were analyzed to compare detection of apneas by two different methods of combined airflow signals: a) Therm-NP and b) TS-NP. The two scoring methods were performed randomly and independently of each other. Respiratory event detection results and apnea–hypopnea index (AHI) obtained using Therm-NP signals as a reference were compared to those obtained by TS-NP signals. The number of respiratory events scored using Therm-NP was 7268, while it was 7329 with TS-NP. The sensitivity and positive predictive value of TS-NP with respect to Therm-NP were 93.0 and 90.6, respectively. The average AHI for all patients was 29.8 ± 22.9 for Therm-NP and 30.0 ± 22.6/h for TS-NP. With the Therm-NP as a reference detection, a kappa statistic value of 0.86 for TS-NP revealed a high agreement for classifying OSA into the severity classes mild, moderate, and severe. However, 55% of total events scored by Therm-NP were hypopneas while 59% of events scored by TS-NP were apneas. These results show that TS-NP reliably detects the same number of respiratory events as Therm-NP, despite a difference in the apnea to hypopnea ratio between the two methods. The use of TS does not modify AHI calculations and guarantees accurate sleep apnea diagnosis and severity assessment. Thus, TS can be used as a substitute for an oronasal thermistor in sleep recording systems.

Obstructive Sleep Apnea Screening Using a Piezo-Electric Sensor.

In this study, we propose a novel method for obstructive sleep apnea (OSA) detection using a piezo-electric sensor. OSA is a relatively common sleep disorder. However, more than 80% of OSA patients remain undiagnosed. We investigated the feasibility of OSA assessment using a single-channel physiological signal to simplify the OSA screening. We detected both snoring and heartbeat information by using a piezo-electric sensor, and snoring index (SI) and features based on pulse rate variability (PRV) analysis were extracted from the filtered piezo-electric sensor signal. A support vector machine (SVM) was used as a classifier to detect OSA events. The performance of the proposed method was evaluated on 45 patients from mild, moderate, and severe OSA groups. The method achieved a mean sensitivity, specificity, and accuracy of 72.5%, 74.2%, and 71.5%; 85.8%, 80.5%, and 80.0%; and 70.3%, 77.1%, and 71.9% for the mild, moderate, and severe groups, respectively. Finally, these results not only show the feasibility of OSA detection using a piezo-electric sensor, but also illustrate its usefulness for monitoring sleep and diagnosing OSA.

Sleep Apnea Screening by Autoregressive Models From a Single ECG Lead

This paper presents a method for obstructive sleep apnea (OSA) screening based on the electrocardiogram (ECG) recording during sleep. OSA is a common sleep disorder produced by repetitive occlusions in the upper airways and this phenomenon can usually be observed also in other peripheral systems such as the cardiovascular system. Then the extraction of ECG characteristics, such as the RR intervals and the area of the QRS complex, is useful to evaluate the sleep apnea in noninvasive way. In the presented analysis, 50 recordings coming from the apnea Physionet database were used; data were split into two sets, the training and the testing set, each of which was composed of 25 recordings. A bivariate time-varying autoregressive model (TVAM) was used to evaluate beat-by-beat power spectral densities for both the RR intervals and the QRS complex areas. Temporal and spectral features were changed on a minute-by-minute basis since apnea annotations where given with this resolution. The training set consisted of 4950 apneic and 7127 nonapneic minutes while the testing set had 4428 apneic and 7927 nonapneic minutes. The K-nearest neighbor (KNN) and neural networks (NN) supervised learning classifiers were employed to classify apnea and non apnea minutes. A sequential forward selection was used to select the best feature subset in a wrapper setting. With ten features the KNN algorithm reached an accuracy of 88%, sensitivity equal to 85%, and specificity up to 90%, while NN reached accuracy equal to 88%, sensitivity equal to 89% and specificity equal to 86%. In addition to the minute-by-minute classification, the results showed that the two classifiers are able to separate entirely (100%) the normal recordings from the apneic recordings. Finally, an additional database with eight recordings annotated as normal or apneic was used to test again the classifiers. Also in this new dataset, the results showed a complete separation between apneic and normal recordings.

Dilation of the oropharynx via selective stimulation of the hypoglossal nerve

The functional effects of selective hypoglossal nerve (HG) stimulation with a multi-contact peripheral nerve electrode were assessed using images of the upper airways and the tongue in anesthetized beagles. A biphasic pulse train of 50 Hz frequency and 2 s duration was applied through each one of the tripolar contact sets of the nerve electrode while the pharyngeal images were acquired into a computer. The stimulation current was limited to 20% above the activation threshold for maximum selectivity. The images showed that various contact sets could generate several different activation patterns of the tongue muscles resulting in medial and/or lateral dilation and closing of the airways at the tongue root. Some of these patterns translated into an increase in the oropharyngeal size while others did not have any effect. The pharyngeal sizes were not statistically different during stimulation either between the two different positions of the head (30° and 60°), or when the lateral contacts were compared with the medial ones. The contacts that had the least effect generated an average of 53 ± 15% pharyngeal dilation relative to the best contacts, indicating that the results are marginally sensitive to the contact position around the HG nerve trunk. These results suggest that selective HG nerve stimulation can be a useful technique to produce multiple tongue activation patterns that can dilate the pharynx. This may in turn increase the size of the patient population who can benefit from HG nerve stimulation as a treatment method for obstructive sleep apnea.

Radiofrequency of the lingual tonsil through a cervical approach

Objectives: To explore the feasibility of a new approach to reduce lingual tonsil hypertrophy using radiofrequency in a patient with obstructive sleep apnea Methods: A 50-year-old patient who had had 3 consecutive procedures to treat upper airway disorders (UPPP, bimaxillary advancement, and radiofrequency tongue base reduction) still complained of daytime sleepiness. His respiratory disturbance index was 41 events per hour. His clinical examination showed major lingual tonsil hypertrophy. Radiofrequency of the lingual tonsils trough a cervical approach was performed with a sonographic and fluoroscopic control and under potentialized local anesthesia. A retractable 2 cm active electrode mounted on a 5 cm mandrel was introduced through the skin midway between the hyoid bone and the mandible in the midline. Eleven thousand joules were delivered in several points in the lingual tonsils Results: The patient did not complain of any pain during the procedure. He could eat within 4 hours. His tongue mobility remained normal. Nevertheless, the patient expressed pain for the 2 following weeks. There was no sign of clinical infection. A tomodensitometry showed oedema but no abscess. Five months after the procedure, the patient was no longer somnolent. The volume of the lingual tonsils decreased. His respiratory disturbance index was 8. Conclusion: Lingual tonsil radiofrequency is possible using a cervical approach under local anesthesia thus avoiding a possible difficult intubation. The fluoroscopy helped determine the position of the electrode in the sagittal and horizontal planes. Lingual tonsil hypertrophy can induce obstructive sleep apnea. By reducing the lingual tonsil volume, a patient could be cured.

Closed-loop stimulation of hypoglossal nerve in a dog model of upper airway obstruction.

Electrical stimulation of upper airway (UAW) muscles has been under investigation as a treatment method for obstructive sleep apnea (OSA). Particular attention has been given to the electrical activation of the genioglossal muscle, either directly or via the stimulation of the hypoglossal nerve (HG), since the genioglossus is the main tongue protrusor muscle. Regardless of the stimulation site or method, an implantable electrical stimulation device for OSA patients will require a reliable method for detection of obstructive breaths to apply the stimulation when needed. In this paper, we test the hypothesis that the activity of the HG nerve can be used as a feedback signal for closed-loop stimulation of the HG nerve in an animal model of UAW obstruction where a force is applied on the submental region to physically narrow the airways. As an advantage, the method uses a single electrode for both recording and stimulation of the HG nerve. Simple linear filtering techniques were found to be adequate for producing the trigger signal for the electrical stimulation from the HG recordings. Esophageal pressure, which was used to estimate the size of the UAW passage, returned to the preloading values during closed-loop stimulation of the HG nerve. The data demonstrate the feasibility of the closed-loop stimulation of the HG nerve using its activity as the feedback signal.