Purpose:Clinical studies focusing on the association between the gut microbiota and obstructive sleep apnea (OSA) are limited. This study aimed to explore the relationship between intermittent hypoxia and the composition of gut microbiota in adults by analyzing the differences in the characteristics and functional distribution of gut microbiota between patients with different severities of OSA and healthy individuals. Patients and Methods:A cohort of 113 individuals from the First Affiliated Hospital of Sun Yat-sen University underwent overnight polysomnography from July 2019 to August 2021. The individuals included 16 healthy controls and 97 patients with OSA, categorized by the apnea-hypopnea index into mild, moderate, and severe groups. Fecal samples were analyzed using high-throughput sequencing of the 16S rRNA V3-V4 region to assess gut microbiota composition and function. Correlation analysis was used to evaluate the association between clinical indicators and microbiota markers. Results:In patients with OSA, the gut microbiota diversity and the abundance of specific microbes that produce short-chain fatty acids decreased (P<0.05). The phyla Verrucomicrobia and Candidatus Saccharibacteria, genera Gemmiger and Faecalibacterium, and the species Gemmiger formicilis exhibited decreasing abundance with increasing OSA severity. Correlation analysis revealed a robust association between the proportion of total sleep time, characterized by nighttime blood oxygen saturation below 90%, and the alterations in the gut microbiota, demonstrating that elevated levels of desaturation are correlated with pronounced microbiota dysbiosis (P<0.05). Conclusion:Compared to the control group, the intermittent hypoxia exhibited by patients with OSA may be related to alterations in the composition and structure of the gut microbiota. Our results demonstrate the importance of monitoring hypoxia indicators in future clinical practice.

Arousal detection is of great importance to measure sleep quality. Traditional diagnosis of arousal that relies on polysomnography (PSG) signals is cumbersome and costly, requiring an overnight stay in a sleep laboratory with multiple sensors attached to the subjects, which limits widespread adoption. Conversely, electrocardiogram (ECG) signal offers a convenient and affordable alternative due to its ease of integration into portable devices. However, most of methods for automatic arousal detection were based on PSG signals. Besides, they ignored the importance of learning neighborhood information of physiological signals. To address these issues, we propose a Multi-Scale Attention Network called MSANet to detect sleep arousal using single-channel ECG. Specifically, it consists of encoder-decoder structure, multi-scale channel attention module and temporal module. The encoder-decoder structure as the foundational framework is to extract ECG signal features across multiple time scales. The multi-scale channel attention module aims to help the model focus on essential features. In the temporal module, we utilize dilated convolutions to capture neighborhood information in the ECG signal. Besides, considering the rarity of sleep arousal events during the sleep period, we apply focal loss to address class imbalance issue. Experiment results on four publicly available sleep datasets show that our proposed model can achieve a competitive performance for arousal detection.

The most prevalent sleep-disordered breathing condition is Obstructive Sleep Apnea (OSA), which has been linked to various health consequences, including cardiovascular disease (CVD) and even sudden death. Therefore, early detection of OSA can effectively help patients prevent the diseases induced by it. However, many existing methods have low accuracy in detecting hypopnea events or even ignore them altogether. According to the guidelines provided by the American Academy of Sleep Medicine (AASM), two modal signals, namely nasal pressure airflow and pulse oxygen saturation (SpO2), offer significant advantages in detecting OSA, particularly hypopnea events. Inspired by this notion, we propose a bimodal feature fusion CNN model that primarily comprises of a dual-branch CNN module and a feature fusion module for the classification of 10-second-long segments of nasal pressure airflow and SpO2. Additionally, an Efficient Channel Attention mechanism (ECA) is incorporated into the second module to adaptively weight feature map of each channel for improving classification accuracy. Furthermore, we design an OSA Severity Assessment Framework (OSAF) to aid physicians in effectively diagnosing OSA severity. The performance of both the bimodal feature fusion CNN model and OSAF is demonstrated to be excellent through per-segment and per-patient experimental results, based on the evaluation of our method using two real-world datasets consisting of polysomnography (PSG) recordings from 450 subjects.

Objective. Explore a network architecture that can efficiently perform single-lead electrocardiogram (ECG) sleep apnea (SA) detection by utilizing the beneficial information of extended ECG segments and reducing the impact of their noisy information. Approach. We propose an effective deep-shallow fusion network (EDSFnet). The deeper residual network is used to extract high-level features with stronger semantics and less noise from the original ECG segments. The shallower convolutional neural network is used to extract lower-level features with higher resolution containing more detailed neighborhood information from the extended ECG segments. These two types of features are then fused using Effective Channel Attention, implementing automatic weight assignment to take advantage of their complementary nature. Main results. The performance of EDSFnet is evaluated on the Apnea-ECG dataset and the FAH-ECG dataset. In the Apnea-ECG dataset with 35 subjects as the training set and 35 subjects as the test set, the accuracy of EDSFnet was 92.6% and 100% for per-segment and per-recording test, respectively. In the FAH-ECG dataset with 348 subjects as the training set and 88 subjects as the test set, the accuracy of EDSFnet was 89.0% and 93.2% for per-segment and per-recording test, respectively. EDSFnet has achieved state-of-the-art results in both experiments using the publicly available Apnea-ECG dataset and subject-independent experiments using the FAH-ECG clinical dataset. Significance. The success of EDSFnet in handling SA detection underlines its robustness and adaptability. By achieving superior results across different datasets, EDSFnet offers promise in advancing the cost-effective and efficient detection of SA through single-lead ECG, reducing the burden on patients and healthcare systems alike.

Over the past few decades, researchers have attempted to simplify and accelerate the process of sleep stage classification through various approaches; however, only a few such approaches have gained widespread acceptance. Artificial intelligence technology, particularly deep learning, is promising for earning the trust of the sleep medicine community in automated sleep-staging systems, thus facilitating its application in clinical practice and integration into daily life. We aimed to comprehensively review the latest methods that are applying deep learning for enhancing sleep staging efficiency and accuracy. Starting from the requisite “data” for constructing deep learning algorithms, we elucidated the current landscape of this domain and summarized the fundamental modeling process, encompassing signal selection, data pre-processing, model architecture, classification tasks, and performance metrics. Furthermore, we reviewed the applications of automated sleep staging in scenarios such as sleep-disorder screening, diagnostic procedures, and health monitoring and management. Finally, we conducted an in-depth analysis and discussion of the challenges and future in intelligent sleep staging, particularly focusing on large-scale sleep datasets, interdisciplinary collaborations, and human-computer interactions.

To investigate the consistency between the reflux symptom score (RSS) and the multitemporal salivary pepsin test in screening for laryngopharyngeal reflux (LPR) and the screening value of the RSS for LPR by simultaneously administering daytime multitemporal salivary pepsin test and RSS to patients. This was a single-center prospective observational study. All included patients underwent simultaneous daytime multitemporal salivary pepsin testing and RSS. A participant was considered to have LPR when one or more positive salivary pepsin test results or RSS score > 13 were obtained. The consistency between the multitemporal salivary pepsin test and the RSS was compared by the weighted Cohen’s kappa statistic. The screening value of the RSS for LPR was investigated by receiver operating characteristic (ROC) analysis. A total of 67 patients were included. The positivity rate of LPR was 71.64

We evaluated the validity of a squeeze-and-excitation and multiscaled fusion network (SE-MSCNN) using single-lead electrocardiogram (ECG) signals for obstructive sleep apnea detection and classification. Overnight polysomnographic data from 436 participants at the Sleep Center of the First Affiliated Hospital of Sun Yat-sen University were used to generate a new FAH-ECG dataset comprising 260, 88, and 88 single-lead ECG signal recordings for training, validation, and testing, respectively. The SE-MSCNN was employed for detection of apnea-hypopnea events from the acquired ECG segments. Sensitivity, specificity, accuracy, and F1 scores were assigned to assess algorithm performance. We also used the SE-MSCNN to estimate the apnea-hypopnea index, classify obstructive sleep apnea severity, and compare the agreement between 2 sleep technicians. The SE-MSCNN’s accuracy, sensitivity, specificity, and F1 score on the FAH-ECG dataset were 86.6

Background Laryngopharyngeal cancer (LPC) includes laryngeal and hypopharyngeal cancer, whose early diagnosis can significantly improve the prognosis and quality of life of patients. Pathological biopsy of suspicious cancerous tissue under the guidance of laryngoscopy is the gold standard for diagnosing LPC. However, this subjective examination largely depends on the skills and experience of laryngologists, which increases the possibility of missed diagnoses and repeated unnecessary biopsies. We aimed to develop and validate a deep convolutional neural network-based Laryngopharyngeal Artificial Intelligence Diagnostic System (LPAIDS) for real-time automatically identifying LPC in both laryngoscopy white-light imaging (WLI) and narrow-band imaging (NBI) images to improve the diagnostic accuracy of LPC by reducing diagnostic variation among on-expert laryngologists. Methods All 31,543 laryngoscopic images from 2382 patients were categorised into training, verification, and test sets to develop, validate, and internal test LPAIDS. Another 25,063 images from five other hospitals were used as external tests. Overall, 551 videos were used to evaluate the real-time performance of the system, and 200 randomly selected videos were used to compare the diagnostic performance of the LPAIDS with that of laryngologists. Two deep-learning models using either WLI (model W) or NBI (model N) images were constructed to compare with LPAIDS. Results LPAIDS had a higher diagnostic performance than models W and N, with accuracies of 0·956 and 0·949 in the internal image and video tests, respectively. The robustness and stability of LPAIDS were validated in external sets with the area under the receiver operating characteristic curve values of 0·965–0·987. In the laryngologist-machine competition, LPAIDS achieved an accuracy of 0·940, which was comparable to expert laryngologists and outperformed other laryngologists with varying qualifications. Conclusions LPAIDS provided high accuracy and stability in detecting LPC in real-time, which showed great potential for using LPAIDS to improve the diagnostic accuracy of LPC by reducing diagnostic variation among on-expert laryngologists.

Sleep apnea (SA) is a common sleep-related breathing disorder, which would lead to damage of multiple systemic organs or even sudden death. In clinical practice, portable device is an important tool to monitor sleep conditions and detect SA events by using physiological signals. However, SA detection performance is still limited due to physiological signals with time-variability and complexity. In this paper, we focus on SA detection with single lead ECG signals, which can be easily collected by a portable device. Under this context, we propose a restricted attention fusion network called RAFNet for sleep apnea detection. Specifically, RR intervals (RRI) and R-peak amplitudes (Rpeak) are generated from ECG signals and divided into one-minute-long segments. To alleviate the problem of insufficient feature information of the target segment, we combine the target segment with two pre- and post-adjacent segments in sequence, (i.e. a five-minute-long segment), as the input. Meanwhile, by leveraging the target segment as the query vector, we propose a new restricted attention mechanism with cascaded morphological and temporal attentions, which can effectively learn the feature information and depress redundant feature information from the adjacent segments with adaptive assigning weight importance. To further improve the SA detection performance, the target and adjacent segment features are fused together with the channel-wise stacking scheme. Experiment results on the public Apnea-ECG dataset and the real clinical FAH-ECG dataset with sleep apnea annotations show that the RAFNet greatly improves SA detection performance and achieves competitive results, which are superior to those achieved by the state-of-the-art baselines.

An increasing number of people are being influenced by sleep apnea (SA), a disease that brings complications such as hypertension and arrhythmia. Single-lead electrocardiogram (ECG) SA detection based on portable devices is of increased interest as an alternative to the expensive and time-consuming polysomnography (PSG) performed in hospitals, because of its low cost and the use of light-weight portable devices. In single-lead ECG SA detection, researchers have found that considering the neighborhood information of the ECG segment can improve detection performance. However, existing work fails to reduce the impact of noise in the neighborhood information. To address this issue, we propose a deep-shallow fusion network, DSF-SANet. We use a deeper residual network to extract higher-level features of the original ECG segments, which are semantically strong and contain less noise, and lower-level features with high resolution, containing more detailed neighborhood information from ECG segments. Effective channel attention (ECA) is then used to fuse these two types of features to exploit their complementary nature. DSF-SANet is state-of-the-art on the PhysioNet Apnea-ECG dataset, with accuracy of 92.6% per-segment, and 100% per-recording. DSF-SANet achieved competitive results in subject-independent experiments on the FAH-ECG clinical dataset. The DSF-SANet source code is available at https://github.com/lipanhub/DSF-SANet.

Background For recurrent laryngeal cancer, the feasibility of salvage transoral laser microsurgery (TLM) remains controversial. This study compared the efficacy of TLM and open partial laryngectomy (OPL) for treatment of early local recurrence of glottic squamous cell cancer (GSCC) and confirm the effectiveness of salvage TLM as a treatment option. Methods This retrospective study involved 55 patients with early local recurrent GSCC treated with TLM, and the oncologic outcomes, functional outcomes, hospitalization time and complications were compared with a group of 40 recurrent GSCC patients matched for clinical variables of TLM group, treated by OPL by the same team of surgeons. Results The 5-year overall survival and disease-specific survival rates were 65.8% and 91.5%, respectively, for 55 patients with rT is –rT 2 stage treated by TLM and 77.1% and 94.7%, respectively, for 40 patients with rT is –rT 2 stage treated by OPL (OPL group). In the TLM and OPL groups, the local control rates after 5 years were 77.5% and 79.3%, respectively, and the laryngeal preservation rates were 94.4% and 83.6%, respectively ( p > 0.05). Compared with the OPL group, the complication rate (1.82%) and hospitalization duration (5.42 ± 2.26 days) were significantly lower in the TLM group ( p < 0.05). Compared with the OPL group, postsurgical health-related quality of life and quality of voice were significantly better in the TLM group ( p < 0.001). Conclusion Salvage TLM can be used as an effective treatment option for suitable patients after a full, comprehensive, and careful assessment of the characteristics of early locally recurrent glottic carcinoma. Graphical Abstract

Sleep apnea (SA) is a common breathing disease, with clinical manifestations of sleep snoring at night with apnea and daytime sleepiness. It could lead to ischemic heart disease, stroke, or even sudden death. SpO 2 signal is highly related to SA, and many automatic SA detection methods have been proposed. However, extant work focuses on small datasets with relatively few subjects (less than 100) and is unaware of SA syndromes occurring about 5 seconds prior to the SpO 2 change. This study proposes an automatic SA detector called DSCNN using a single-lead SpO 2 signal with a dual-scale convolutional neural network. To solve the time-delayed problem of SpO 2 changes, we enlarge the target SpO 2 segment information by combining its subsequent segment information. To utilize neighbouring segments information and further facilitate the SA detection performance, a dual-scale neural network with the fusing information of the prolonged target segment and its two surrounding segments is proposed. Three datasets from multiple centres are employed to verify the generic performance of DSCNN. Here, we must point out that we use two datasets as external datasets, and one of them is collected from the First Affiliated Hospital of Sun Yat-sen University with a large sample size (450 subjects). Extensive experiment results show that DSCNN can achieve promising results which are superior to the existing state-of-the-art methods.

The 5-year survival rate of laryngeal cancer continues to decline, and the laryngeal particularity of the anatomy adversely affects the patient's quality of life. Emerging evidence suggests that long noncoding RNAs (lncRNAs) are closely correlated to key steps in the malignant progression of cancer cells. In this study, we report the role of lncRNA SBF2-AS1/miR-302b-3p/TGFBR2 interactions in the metastasis of laryngeal squamous cell carcinoma (LSCC). We verified that SBF2-AS1 was significantly downregulated in LSCC tissues and cell lines using qRT-PCR analysis. Its low expression was correlated to lymph node metastasis and an advanced clinical stage. More importantly, LSCC patients with low expression of SBF2-AS1 tended to have a poor prognosis. Based on this, we performed gain-of-function and loss-of-function experiments in LSCC cell lines. The results confirmed that knocking down SBF2-AS1 can promote the metastasis of LSCC cells and enhance epithelial-mesenchymal transition phenotype, while the upregulation of SBF2-AS1 expression resulted in the opposite. Our in vivo model verified that SBF2-AS1 overexpression could inhibit LSCC cell metastasis. Subsequent mechanistic studies revealed that SBF2-AS1 acted as a competing endogenous RNA that upregulated the expression of TGFBR2 by endogenous sponging for miR-302b-3p in LSCC cell lines. Moreover, miR-302b-3p overexpression reversed the inhibitory effects on LSCC metastasis induced by upregulation of SBF2-AS1 expression, and inhibition of TGFBR2 expression reversed the effect of SBF2-AS1 on metastasis. Our study proposes SBF2-AS1 as a biomarker to predict the prognosis of LSCC patients and a novel potential therapeutic target.

Abstract Background Deep neck space abscess (DNSA) is a serious infection in the head and neck. Antibiotic therapy is an important treatment in patients with DNSA. However, the results of bacterial culture need at least 48 h, and the positive rate is only 30–50%, indicating that the use of empiric antibiotic treatment for most patients with DNSA should at least 48 h or even throughout the whole course of treatment. Thus, how to use empiric antibiotics has always been a problem for clinicians. This study analyzed the distribution of bacteria based on disease severity and clinical characteristics of DNSA patients, and provides bacteriological guidance for the empiric use of antibiotics. Methods We analyzed 433 patients with DNSA who were diagnosed and treated at nine medical centers in Guangdong Province between January 1, 2015, and December 31, 2020. A nomogram for disease severity (mild/severe) was constructed using least absolute shrinkage and selection operator–logistic regression analysis. Clinical characteristics for the Gram reaction of the strain were identified using multivariate analyses. Results 92 (21.2%) patients developed life-threatening complications. The nomogram for disease severity comprised of seven predictors. The area under the receiver operating characteristic curves of the nomogram in the training and validation cohorts were 0.951 and 0.931, respectively. In the mild cases, 43.2% (101/234) had positive culture results (49% for Gram-positive and 51% for Gram-negative strains). The positive rate of cultures in the patients with severe disease was 63% (58/92, 37.9% for Gram-positive, and 62.1% for Gram-negative strains). Diabetes mellitus was an independent predictor of Gram-negative strains in the mild disease group, whereas gas formation and trismus were independent predictors of Gram-positive strains in the severe disease group. The positivity rate of multidrug-resistant strains was higher in the severe disease group (12.1%) than in the mild disease group (1.0%) (P < 0.001). Metagenomic sequencing was helpful for the bacteriological diagnosis of DNSA by identifying anaerobic strains (83.3%). Conclusion We established a DNSA clinical severity prediction model and found some predictors for the type of Gram-staining strains in different disease severity cases. These results can help clinicians in effectively choosing an empiric antibiotic treatment.

Abstract Objective This article aims to propose a new surgical method for the treatment of pyriform fistula, especially for the complex pyriform fistula. Methods A total of 36 patients with pyriform fistula underwent the procedure between August 2017 to October 2020. Surgery was performed by the senior authors using the same technique at the same clinical center for all patients. The median follow‐up time was 33 months. Meantime, we collected information on patients with pyriform fistula using traditional surgical methods in our hospital from April 2015 to November 2018 for comparison. Results The surgery was successfully completed in 36 patients. In all, 32 patients had a history of multiple incisions and drainage, 16 patients had a history of surgical resections, and two patients had a history of cauterization of the internal fistula. Compared with traditional surgical methods, our new surgical method greatly shortens the length of the surgical incision (4.3 vs. 5.5, p < 0.0001), reduces the operation time (8.1 vs. 27.1, p < 0.0001), and reduces the blood loss (103.2 vs. 196.8, p < 0.0001). None of the 36 patients in this study had complications such as pharyngeal fistula, recurrent laryngeal nerve paralysis, or hypothyroidism. The mean follow‐up duration after the excision of the lesion was 34.1 months. To date, no patients have relapsed. Conclusion Our experience showed that this surgical technique could be used to completely remove the fistula, and it was easier to perform than the conventional strategies. These treatment options result in less trauma and reliable results, especially for complex pyriform fistulas. Level of evidence IV

Background Airway management, including noninvasive endotracheal intubation or invasive tracheostomy, is an essential treatment strategy for patients with deep neck space abscess (DNSA) to reverse acute hypoxia, which aids in avoiding acute cerebral hypoxia and cardiac arrest. This study aimed to develop and validate a novel risk score to predict the need for airway management in patients with DNSA. Methods Patients with DNSA admitted to 9 hospitals in Guangdong Province between January 1, 2015, and December 31, 2020, were included. The cohort was divided into the training and validation cohorts. The risk score was developed using the least absolute shrinkage and selection operator (LASSO) and logistic regression models in the training cohort. The external validity and diagnostic ability were assessed in the validation cohort. Results A total of 440 DNSA patients were included, of which 363 (60 required airway management) entered into the training cohort and 77 (13 required airway management) entered into the validation cohort. The risk score included 7 independent predictors (p < 0.05): multispace involvement (odd ratio [OR] 6.42, 95% confidence interval [CI] 1.79–23.07, p < 0.001), gas formation (OR 4.95, 95% CI 2.04–12.00, p < 0.001), dyspnea (OR 10.35, 95% CI 3.47–30.89, p < 0.001), primary region of infection, neutrophil percentage (OR 1.10, 95% CI 1.02–1.18, p = 0.015), platelet count to lymphocyte count ratio (OR 1.01, 95% CI 1.00–1.01, p = 0.010), and albumin level (OR 0.86, 95% CI 0.80–0.92, p < 0.001). Internal validation showed good discrimination, with an area under the curve (AUC) of 0.951 (95% CI 0.924–0.971), and good calibration (Hosmer–Lemeshow [HL] test, p = 0.821). Application of the clinical risk score in the validation cohort also revealed good discrimination (AUC 0.947, 95% CI 0.871–0.985) and calibration (HL test, p = 0.618). Decision curve analyses in both cohorts demonstrated that patients could benefit from this risk score. The score has been transformed into an online calculator that is freely available to the public. Conclusions The risk score may help predict a patient’s risk of requiring airway management, thus advancing patient safety and supporting appropriate treatment.

Purpose This study evaluated a novel approach for diagnosis and classification of obstructive sleep apnea (OSA), called Obstructive Sleep Apnea Smart System (OSASS), using residual networks and single-channel nasal pressure airflow signals. Methods Data were collected from the sleep center of the First Affiliated Hospital, Sun Yat-sen University, and the Integrative Department of Guangdong Province Traditional Chinese Medical Hospital. We developed a new model called the multi-resolution residual network (Mr-ResNet) based on a residual network to detect nasal pressure airflow signals recorded by polysomnography (PSG) automatically. The performance of the model was assessed by its sensitivity, specificity, accuracy, and F1-score. We built OSASS based on Mr-ResNet to estimate the apnea‒hypopnea index (AHI) and to classify the severity of OSA, and compared the agreement between OSASS output and the registered polysomnographic technologist (RPSGT) score, assessed by two technologists. Results In the primary test set, the sensitivity, specificity, accuracy, and F1-score of Mr-ResNet were 90.8%, 90.5%, 91.2%, and 90.5%, respectively. In the independent test set, the Spearman correlation for AHI between OSASS and the RPSGT score determined by two technologists was 0.94 (p < 0.001) and 0.96 (p < 0.001), respectively. Cohen’s Kappa scores for classification between OSASS and the two technologists’ scores were 0.81 and 0.84, respectively. Conclusion Our results indicated that OSASS can automatically diagnose and classify OSA using signals from a single-channel nasal pressure airflow, which is consistent with polysomnographic technologists’ findings. Thus, OSASS holds promise for clinical application.

Sleep apnea-hypopnea syndrome (SAHS), as a widespread respiratory sleep disorder, if left untreated, can lead to a series of pathological changes. By using Polysomnography (PSG), traditional SAHS diagnosis tends to be complex and costly. Nasal airflow (NA) is the most direct reflection of the severity of SAHS. Therefore, we try to take advantage of NA signals that can be easily recorded by wearable devices. In this paper, we present an automatic detection approach of SAH events based on single-channel signal. Through this approach, an enhanced frequency extraction network is designed, which factorizes the mixed feature maps by their frequencies. And the spatial resolution of low-frequency components is reduced so as to save spending. Besides, in our research, the vanilla convolution block of the high-frequency components are replaced by residual blocks and smaller groups of filters with bigger size kernels. And we use the spatial attention module to facilitate feature extraction. Compared with state-of-the-art networks in this field, the promising results reveal that the proposed network for SAH events multiclass classification shows outstanding performance with accuracy of 91.23%, sensitivity of 90.81% and specificity of 90.59%. Thus, we believe that our approach, as a low-cost and high-efficiency solution, shows a great potential for detecting SAH events.

*These authors contributed equally to this work Purpose: This study evaluated a novel approach for diagnosis and classification of obstructive sleep apnea (OSA), called Obstructive Sleep Apnea Smart System (OSASS), using residual networks and single-channel nasal pressure airflow signals. Methods: Data were collected from the sleep center of the First Affiliated Hospital, Sun Yat-sen University, and the Integrative Department of Guangdong Province Traditional Chinese Medical Hospital. We developed a new model called the multi-resolution residual network (Mr-ResNet) based on a residual network to detect nasal pressure airflow signals recorded by polysomnography (PSG) automatically. The performance of the model was assessed by its sensitivity, specificity, accuracy, and F1-score. We built OSASS based on MrResNet to estimate the apnea‒hypopnea index (AHI) and to classify the severity of OSA, and compared the agreement between OSASS output and the registered polysomnographic technologist (RPSGT) score, assessed by two technologists. Results: In the primary test set, the sensitivity, specificity, accuracy, and F1-score of MrResNet were 90.8%, 90.5%, 91.2%, and 90.5%, respectively. In the independent test set, the Spearman correlation for AHI between OSASS and the RPSGT score determined by two technologists was 0.94 (p < 0.001) and 0.96 (p < 0.001), respectively. Cohen’s Kappa scores for classification between OSASS and the two technologists’ scores were 0.81 and 0.84, respectively. Conclusion: Our results indicated that OSASS can automatically diagnose and classify OSA using signals from a single-channel nasal pressure airflow, which is consistent with polysomnographic technologists’ findings. Thus, OSASS holds promise for clinical application.

Obstructive sleep apnea (OSA) is the most common sleep related breathing disorder causing sleepiness and several chronic medical conditions, for which the gold-standard diagnostic test is polysomnography (PSG), however the analysis of PSG is a time-consuming and labor-intensive procedure. To address these issues, we use deep learning as a new method to detect sleep respiratory events which can provide effective and accurate OSA diagnosis. We present a network named Mr-ResNeXt improved from ResNeXt, in which the 3 × 3 filters was replaced by a new block containing multi-level group convolution. The first level group convolution is used to exchange information between groups and the second level group convolution contains filters of different sizes which are used to extract features of different resolutions. All group convolutions involve residual-like connections. All the above changes help to extract multi-resolution image features more easily. Firstly, the experimental results show that our network can achieve a nearly 3