A comparison of regularized logistic regression and random forest machine learning models for daytime diagnosis of obstructive sleep apnea

A novel machine learning model for screening the risk of obstructive sleep apnea using craniofacial photography with questionnaires

STUDY OBJECTIVES

Undiagnosed or untreated moderate-to-severe obstructive sleep apnea (OSA) increases cardiovascular risks and mortality. Early and efficient detection is critical, given its high prevalence. We aimed to develop a practical and efficient approach for OSA screening, using simple facial photography and sleep questionnaires.

METHODS

We retrospectively included 748 participants who completed polysomnography, sleep questionnaires (STOP-BANG), and facial photographs at a university hospital between 2012 and 2023. Owing to class imbalance, we randomly undersampled the participants, categorized into the moderate/severe or no/mild OSA group, based on an apnea-hypopnea index of 15 events/h. Using a validated convolutional neural network, we extracted the OSA probability scores from photographs, which were used as the input for the questionnaires. Four machine learning models were employed to classify the moderate/severe vs no/mild groups and evaluated in the test dataset.

RESULTS

We analyzed 426 participants (213 each in the moderate/severe and no/mild groups). The mean (standard deviation) age was 44.6 (14.7) years; 80.8% were men. Logistic regression achieved the highest performance: the area under the receiver operator curve was 97.2%, and accuracy was 91.9%. Adding OSA probability, retrieved from facial photographs, to the questionnaires improved performance, compared with using questionnaires or photographs alone (the area under the receiver operating characteristic curve 97.2% using both, 85.7% for photographs alone, and 64% and 79.1% for questionnaire threshold STOP-BANG scores of 3 and 4, respectively).

CONCLUSIONS

Using simple facial photographs and sleep questionnaires, a 2-stage approach (convolutional neural network + machine learning) accurately classified OSA into moderate/severe vs no/mild OSA groups. This method may facilitate optimal OSA treatment and avoid unnecessary costly evaluations.

CITATION

Park J-Y, Shin H-R, Kim MH, et al. A novel machine learning model for screening the risk of obstructive sleep apnea using craniofacial photography with questionnaires. J Clin Sleep Med. 2025;21(5):843-854.

Advancements in Obstructive Sleep Apnea Diagnosis and Screening Through Artificial Intelligence: A Systematic Review

BACKGROUND

Obstructive sleep apnea (OSA) is a prevalent yet underdiagnosed condition associated with a major healthcare burden. Current diagnostic tools, such as full-night polysomnography (PSG), pose a limited accessibility to diagnosis due to their elevated costs. Recent advances in Artificial Intelligence (AI), including Machine Learning (ML) and deep learning (DL) algorithms, offer novel potential tools for an accurate OSA screening and diagnosis. This systematic review evaluates articles employing AI-powered models for OSA screening and diagnosis in the last decade.

METHODS

A comprehensive electronic search was performed on PubMed/MEDLINE, Google Scholar, and SCOPUS databases. The included studies were original articles written in English, reporting the use of ML algorithms to diagnose and predict OSA in suspected patients. The last search was performed in June 2024. This systematic review is registered in PROSPERO (Registration ID: CRD42024563059).

RESULTS

Sixty-five articles, involving data from 109,046 patients, met the inclusion criteria. Due to the heterogeneity of the algorithms, outcomes were analyzed into six sections (anthropometric indexes, imaging, electrocardiographic signals, respiratory signals, and oximetry and miscellaneous signals). AI algorithms demonstrated significant improvements in OSA detection, with accuracy, sensitivity, and specificity often exceeding traditional tools. In particular, anthropometric indexes were most widely used, especially in logistic regression-powered algorithms.

CONCLUSIONS

The application of AI algorithms to OSA diagnosis and screening has great potential to improve patient outcomes, increase early detection, and lessen the load on healthcare systems. However, rigorous validation and standardization efforts must be made to standardize datasets.

[Sleep Innovation]

Advances in sleep medicine have driven significant improvements in the diagnosis and treatment of sleep disorders such as obstructive sleep apnea (OSA). This disorder affects one billion people worldwide and traditionally, diagnosis is based on polysomnography (PSG), a laborious method that requires specialized personnel. However, the integration of artificial intelligence (AI) in sleep medicine has made it possible to automate the analysis of sleep phases and respiratory events with high accuracy.Machine learning algorithms and neural networks have proven to be effective in automatic sleep coding, with hit rates comparable to those of human experts. These advances make it possible to improve the efficiency of sleep labs and to personalize OSA treatment. In addition, techniques such as cluster analysis are used to identify symptomatic patterns and phenotypes, which improves understanding of OSA pathophysiology and optimizes CPAP treatment.However, implementation of AI in hospitals faces technological, ethical, and legal barriers. Challenges include data quality, patient privacy, and the need for specialized personnel. Despite these obstacles, AI and Big Data have the potential to transform medical care for sleep disorders, improving both diagnosis and treatment adherence, provided regulatory and cultural barriers are overcome.

Predicting class switch recombination in B-cells from antibody repertoire data

Statistical and machine learning methods have proved useful in many areas of immunology. In this paper, we address for the first time the problem of predicting the occurrence of class switch recombination (CSR) in B-cells, a problem of interest in understanding antibody response under immunological challenges. We propose a framework to analyze antibody repertoire data, based on clonal (CG) group representation in a way that allows us to predict CSR events using CG level features as input. We assess and compare the performance of several predicting models (logistic regression, LASSO logistic regression, random forest, and support vector machine) in carrying out this task. The proposed approach can obtain an unweighted average recall of71 % $71\%$ with models based on variable region descriptors and measures of CG diversity during an immune challenge and, most notably, before an immune challenge.

Development and application of a machine learning-based predictive model for obstructive sleep apnea screening

Objective

To develop a robust machine learning prediction model for the automatic screening and diagnosis of obstructive sleep apnea (OSA) using five advanced algorithms, namely Extreme Gradient Boosting (XGBoost), Logistic Regression (LR), Support Vector Machine (SVM), Light Gradient Boosting Machine (LightGBM), and Random Forest (RF) to provide substantial support for early clinical diagnosis and intervention.

Methods

We conducted a retrospective analysis of clinical data from 439 patients who underwent polysomnography at the Affiliated Hospital of Xuzhou Medical University between October 2019 and October 2022. Predictor variables such as demographic information [age, sex, height, weight, body mass index (BMI)], medical history, and Epworth Sleepiness Scale (ESS) were used. Univariate analysis was used to identify variables with significant differences, and the dataset was then divided into training and validation sets in a 4:1 ratio. The training set was established to predict OSA severity grading. The validation set was used to assess model performance using the area under the curve (AUC). Additionally, a separate analysis was conducted, categorizing the normal population as one group and patients with moderate-to-severe OSA as another. The same univariate analysis was applied, and the dataset was divided into training and validation sets in a 4:1 ratio. The training set was used to build a prediction model for screening moderate-to-severe OSA, while the validation set was used to verify the model's performance.

Results

Among the four groups, the LightGBM model outperformed others, with the top five feature importance rankings of ESS total score, BMI, sex, hypertension, and gastroesophageal reflux (GERD), where Age, ESS total score and BMI played the most significant roles. In the dichotomous model, RF is the best performer of the five models respectively. The top five ranked feature importance of the best-performing RF models were ESS total score, BMI, GERD, age and Dry mouth, with ESS total score and BMI being particularly pivotal.

Conclusion

Machine learning-based prediction models for OSA disease grading and screening prove instrumental in the early identification of patients with moderate-to-severe OSA, revealing pertinent risk factors and facilitating timely interventions to counter pathological changes induced by OSA. Notably, ESS total score and BMI emerge as the most critical features for predicting OSA, emphasizing their significance in clinical assessments. The dataset will be publicly available on my Github.

Obstructive sleep apnea detection during wakefulness: a comprehensive methodological review

Obstructive sleep apnea (OSA) is a chronic condition affecting up to 1 billion people, globally. Despite this spread, OSA is still thought to be underdiagnosed. Lack of diagnosis is largely attributed to the high cost, resource-intensive, and time-consuming nature of existing diagnostic technologies during sleep. As individuals with OSA do not show many symptoms other than daytime sleepiness, predicting OSA while the individual is awake (wakefulness) is quite challenging. However, research especially in the last decade has shown promising results for quick and accurate methodologies to predict OSA during wakefulness. Furthermore, advances in machine learning algorithms offer new ways to analyze the measured data with more precision. With a widening research outlook, the present review compares methodologies for OSA screening during wakefulness, and recommendations are made for avenues of future research and study designs.

Pediatric obstructive sleep apnea diagnosis: leveraging machine learning with linear discriminant analysis

Objective

The objective of this study was to investigate the effectiveness of a machine learning algorithm in diagnosing OSA in children based on clinical features that can be obtained in nonnocturnal and nonmedical environments.

Patients and methods

This study was conducted at Beijing Children's Hospital from April 2018 to October 2019. The participants in this study were 2464 children aged 3-18 suspected of having OSA who underwent clinical data collection and polysomnography(PSG). Participants' data were randomly divided into a training set and a testing set at a ratio of 8:2. The elastic net algorithm was used for feature selection to simplify the model. Stratified 10-fold cross-validation was repeated five times to ensure the robustness of the results.

Results

Feature selection using Elastic Net resulted in 47 features for AHI ≥5 and 31 features for AHI ≥10 being retained. The machine learning model using these selected features achieved an average AUC of 0.73 for AHI ≥5 and 0.78 for AHI ≥10 when tested externally, outperforming models based on PSG questionnaire features. Linear Discriminant Analysis using the selected features identified OSA with a sensitivity of 44% and specificity of 90%, providing a feasible clinical alternative to PSG for stratifying OSA severity.

Conclusions

This study shows that a machine learning model based on children's clinical features effectively identifies OSA in children. Establishing a machine learning screening model based on the clinical features of the target population may be a feasible clinical alternative to nocturnal OSA sleep diagnosis.

Developing a prediction model of children asthma risk using population-based family history health records

BACKGROUND

Identifying children at high risk of developing asthma can facilitate prevention and early management strategies. We developed a prediction model of children's asthma risk using objectively collected population-based children and parental histories of comorbidities.

METHODS

We conducted a retrospective population-based cohort study using administrative data from Manitoba, Canada, and included children born from 1974 to 2000 with linkages to ≥1 parent. We identified asthma and prior comorbid condition diagnoses from hospital and outpatient records. We used two machine-learning models: least absolute shrinkage and selection operator (LASSO) logistic regression (LR) and random forest (RF) to identify important predictors. The predictors in the base model included children's demographics, allergic conditions, respiratory infections, and parental asthma. Subsequent models included additional multiple comorbidities for children and parents.

RESULTS

The cohort included 195,666 children: 51.3% were males and 17.7% had asthma diagnosis. The base LR model achieved a low predictive performance with sensitivity of 0.47, 95% confidence interval (0.45-0.48), and specificity of 0.67 (0.66-0.67) using a predicted probability threshold of 0.20. Sensitivity significantly improved when children's comorbidities were included using LASSO LR: 0.71 (0.69-0.72). Predictive performance further improved by including parental comorbidities (sensitivity = 0.72 [0.70-0.73], specificity = 0.69 [0.69-0.70]). We observed similar results for the RF models. Children's menstrual disorders and mood and anxiety disorders, parental lipid metabolism disorders and asthma were among the most important variables that predicted asthma risk.

CONCLUSION

Including children and parental comorbidities to children's asthma prediction models improves their accuracy.

Application of artificial intelligence in the diagnosis of sleep apnea

STUDY OBJECTIVES

Machine learning (ML) models have been employed in the setting of sleep disorders. This review aims to summarize the existing data about the role of ML techniques in the diagnosis, classification, and treatment of sleep-related breathing disorders.

METHODS

A systematic search in Medline, EMBASE, and Cochrane databases through January 2022 was performed.

RESULTS

Our search strategy revealed 132 studies that were included in the systematic review. Existing data show that ML models have been successfully used for diagnostic purposes. Specifically, ML models showed good performance in diagnosing sleep apnea using easily obtained features from the electrocardiogram, pulse oximetry, and sound signals. Similarly, ML showed good performance for the classification of sleep apnea into obstructive and central categories, as well as predicting apnea severity. Existing data show promising results for the ML-based guided treatment of sleep apnea. Specifically, the prediction of outcomes following surgical treatment and optimization of continuous positive airway pressure therapy can be guided by ML models.

CONCLUSIONS

The adoption and implementation of ML in the field of sleep-related breathing disorders is promising. Advancements in wearable sensor technology and ML models can help clinicians predict, diagnose, and classify sleep apnea more accurately and efficiently.

CITATION

Bazoukis G, Bollepalli SC, Chung CT, et al. Application of artificial intelligence in the diagnosis of sleep apnea. J Clin Sleep Med. 2023;19(7):1337-1363.

Identification of intestinal microbiome associated with lymph-vascular invasion in colorectal cancer patients and predictive label construction

OBJECTIVE

To identify differences between the composition, abundance, and biological function of the intestinal microbiome of patients with and without lymph-vascular invasion (LVI) colorectal cancer (CRC) and to construct predictive labels to support accurate assessment of LVI in CRC.

METHOD

134 CRC patients were included, which were divided into two groups according to the presence or absence of LVI, and their intestinal microbiomes were sequenced by 16SrRNA and analyzed for differences. The transcriptome sequencing data of 9 CRC patients were transformed into immune cells abundance matrix by CIBERSORT algorithm, and the correlation among LVI-associated differential intestinal microbiomes, immune cells, immune-related genes and LVI-associated differential GO items and KEGG pathways were analyzed. A random forest (RF) and eXtreme Gradient Boosting (XGB) model were constructed to predict the LVI of CRC patients based on the differential microbiome.

RESULT

There was no significant difference in α-diversity and β-diversity of intestinal microbiome between CRC patients with and without LVI (P > 0.05). Linear discriminant analysis Effect Size (LEfSe) analysis showed 34 intestinal microbiomes enriched in CRC patients of the LVI group and 5 intestinal microbiomes were significantly enriched in CRC patients of the non-lymph-vascular invasion (NLVI) group. The RF and XGB prediction models constructed with the top 15% of the LVI-associated differential intestinal microbiomes ranked by feature significance had good efficacy.

CONCLUSIONS

There are 39 intestinal flora with significantly different species abundance between the LVI and NLVI groups. g:Alistipes.s:Alistipes_indistinctus is closely associated with colorectal cancer vascular invasion. LVI-associated differential intestinal flora may be involved in regulating the infiltration of immune cells in CRC and influencing the expression of immune-related genes. LVI-associated differential intestinal flora may influence the process of vascular invasion in CRC through a number of potential biological functions. RF prediction models and XGB prediction models constructed based on microbial markers of gut flora can be used to predict CRC-LVI conditions.

Application of various machine learning techniques to predict obstructive sleep apnea syndrome severity

As the incidence of obstructive sleep apnea syndrome (OSAS) increases worldwide, the need for a new screening method that can compensate for the shortcomings of the traditional diagnostic method, polysomnography (PSG), is emerging. In this study, data from 4014 patients were used, and both supervised and unsupervised learning methods were used. Clustering was conducted with hierarchical agglomerative clustering, K-means, bisecting K-means algorithm, Gaussian mixture model, and feature engineering was carried out using both medically researched methods and machine learning techniques. For classification, we used gradient boost-based models such as XGBoost, LightGBM, CatBoost, and Random Forest to predict the severity of OSAS. The developed model showed high performance with 88%, 88%, and 91% of classification accuracy for three thresholds for the severity of OSAS: Apnea-Hypopnea Index (AHI) [Formula: see text] 5, AHI [Formula: see text] 15, and AHI [Formula: see text] 30, respectively. The results of this study demonstrate significant evidence of sufficient potential to utilize machine learning in predicting OSAS severity.

A wearable device for at-home obstructive sleep apnea assessment: State-of-the-art and research challenges

In the last 3 years, almost all medical resources have been reserved for the screening and treatment of patients with coronavirus disease (COVID-19). Due to a shortage of medical staff and equipment, diagnosing sleep disorders, such as obstructive sleep apnea (OSA), has become more difficult than ever. In addition to being diagnosed using polysomnography at a hospital, people seem to pay more attention to alternative at-home OSA detection solutions. This study aims to review state-of-the-art assessment techniques for out-of-center detection of the main characteristics of OSA, such as sleep, cardiovascular function, oxygen balance and consumption, sleep position, breathing effort, respiratory function, and audio, as well as recent progress in the implementation of data acquisition and processing and machine learning techniques that support early detection of severe OSA levels.

Machine Learning Approach to Predict In-Hospital Mortality in Patients Admitted for Peripheral Artery Disease in the United States

Background Peripheral artery disease (PAD) affects >10 million people in the United States. PAD is associated with poor outcomes, including premature death. Machine learning (ML) has been increasingly used on big data to predict clinical outcomes. This study aims to develop ML models to predict in-hospital mortality in patients hospitalized for PAD based on a national database. Methods and Results Inpatient hospitalization data were obtained from the 2016 to 2019 National Inpatient Sample. A total of 150 921 inpatients were identified with a primary diagnosis of PAD and PAD-related procedures using codes of the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) and International Classification of Diseases, Tenth Revision, Procedure Coding System (ICD-10-PCS). Four ML models, including logistic regression, random forest, light gradient boosting, and extreme gradient boosting models, were trained to predict the risk of in-hospital death based on a selection of variables, including patient characteristics, comorbidities, procedures, and hospital-related factors. In-hospital mortality occurred in 1.8% of patients. The performance of the 4 models was comparable, with the area under the receiver operating characteristic curve ranging from 0.83 to 0.85, sensitivity of 77% to 82%, and specificity of 72% to 75%. These results suggest adequate predictability for clinical decision-making. In all 4 models, the total number of diagnoses and procedures, age, endovascular revascularization procedure, congestive heart failure, diabetes, and diabetes with complications were critical predictors of in-hospital mortality. Conclusions This study demonstrates the feasibility of ML in predicting in-hospital mortality in patients with a primary PAD diagnosis. Findings highlight the potential of ML models in identifying high-risk patients for poor outcomes and guiding personalized intervention.

Differentiation Model for Insomnia Disorder and the Respiratory Arousal Threshold Phenotype in Obstructive Sleep Apnea in the Taiwanese Population Based on Oximetry and Anthropometric Features

Insomnia disorder (ID) and obstructive sleep apnea (OSA) with respiratory arousal threshold (ArTH) phenotypes often coexist in patients, presenting similar symptoms. However, the typical diagnosis examinations (in-laboratory polysomnography (lab-PSG) and other alternatives methods may therefore have limited differentiation capacities. Hence, this study established novel models to assist in the classification of ID and low- and high-ArTH OSA. Participants reporting insomnia as their chief complaint were enrolled. Their sleep parameters and body profile were accessed from the lab-PSG database. Based on the definition of low-ArTH OSA and ID, patients were divided into three groups, namely, the ID, low- and high-ArTH OSA groups. Various machine learning approaches, including logistic regression, k-nearest neighbors, naive Bayes, random forest (RF), and support vector machine, were trained using two types of features (Oximetry model, trained with oximetry parameters only; Combined model, trained with oximetry and anthropometric parameters). In the training stage, RF presented the highest cross-validation accuracy in both models compared with the other approaches. In the testing stage, the RF accuracy was 77.53% and 80.06% for the oximetry and combined models, respectively. The established models can be used to differentiate ID, low- and high-ArTH OSA in the population of Taiwan and those with similar craniofacial features.

ANALYSIS OF PHOTOPLETHYSMOGRAPHIC MORPHOLOGY IN SLEEP APNEA SYNDROME PATIENTS USING CURVE FITTING AND SUPPORT VECTOR MACHINE

This paper develops a time-saving, simple, and comfortable method for detecting Sleep Apnea Syndrome (SAS). Seventy SAS patients and 17 healthy persons were randomly selected in this study, and nine analytical parameters (i.e., [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text] of healthy persons and SAS patients during five sleep stages (i.e., W, R, N1, N2, and N3) were obtained to construct a SAS classification model based on logarithmic normal analytical parameters using the Support Vector Machine (SVM) method to fit Photoplethysmographic (PPG) signals. The results show that there were no statistical differences among the five sleep stages for either the healthy or SAS patients. However, there were significant differences in the measured logarithmic normal analytical parameters between the healthy persons and the SAS patients in each of the five sleep stages. The accuracies of the SAS classification model were 95.00%, 90.00%, 84.00%, 94.67%, and 90.77%, corresponding to the five sleep stages, respectively. The SAS classification model based on the SVM method of logarithmic normal analysis parameters can achieve higher classification accuracy for each of the five sleep stages. It can be considered to collect the patient’s pulse wave during the awake period, but not necessarily during the sleep period to classify and identify the SAS; it provides an idea for a convenient and comfortable SAS detection.