Towards Patient-centered Diagnosis of Pediatric Obstructive Sleep Apnea—A Review of Biomedical Engineering Strategies

A wearable ring oximeter for detection of sleep disordered breathing

INTRODUCTION

Wearable devices have been developed that can continuously monitor physiologic variables. One device, Circul® (Bodimetrics Corp, Los Angeles, CA), with a form factor of a ring, measures several variables (SpO2, movement, heart rate).

OBJECTIVES

To evaluate the potential utility of this device in detecting sleep breathing disorders.

MATERIALS AND METHODS

OSA severity measured by the oxygen desaturation Index (ODI) using a wearable oximetric Circul® ring (ODI as defined by 3 % desaturation - cODI3 %) was compared with the gold standard, polysomnography (PSG) in patients suspected of having sleep-disordered breathing. The Circul data was autoscored by the device's software; a technician scored the PSG data. The sensitivity (S), and specificity (E) for the different thresholds for cODI3 % compared to the PSG AHI and PSG ODI, were calculated.

RESULTS

164 patients (age = 44.8 years + 12.3 (SD)) were enrolled. Using the PSG-derived AHI as the reference for classification, the best cut-off points were: OSA = AHI ≥ 5: cODI3 % ≥ 4.3 (S 87.8 %, E 93.8 %); OSA = AHI ≥ 15: cODI3 % ≥ 13.1 (S 76 %, E 100 %); OSA = AHI ≥ 30 = : cODI3 % ≥ 16.2 (S 85.7 %, E 92 %); Using the ODI from the PSG as the reference for classification, the respective cut-off points were: OSA=ODI ≥ 5: cODI3 % ≥ 4.3 (S 93.4 %, E 88.9 %); OSA=ODI ≥ 15:cODI3 % ≥ 13.1 (S85.2 %, E98.4 %); OSA=ODI ≥ 30: cODI3 % ≥ 18.7 (S 98.4 %, E 92.2 %).

CONCLUSIONS

Circul® oximetry demonstrated good diagnostic accuracy compared to the gold standard in determining OSA severity. cODI3 % greater than 13 suggests that significant OSA might be present.

Preoperative Anxiety and Information Needs Among Patients in the Preoperative Holding Area

PURPOSE

This study aims to evaluate the current state of preoperative anxiety and the informational needs of patients undergoing surgery in a preoperative holding area.

DESIGN

Descriptive and Pre post study.

METHODS

A total of 655 elective surgery patients awaiting surgery were selected from November 2021 to March 2022. Multiple linear regression analyses were conducted to examine the factors associated with preoperative anxiety and informational needs in patients exhibiting shadow responses.

FINDINGS

The patients' mean anxiety scores were 10.33 ± 3.25. Among them, 51 patients had mean anxiety scores greater than or equal to 12, indicating a state of anxiety. The mean informational needs scores were 7.81 ± 2.66. A total of 71 patients had mean informational needs scores greater than or equal to 5, reflecting a moderate or higher level of informational needs. Multiple linear regression analysis identified gender, age, surgical table, type of surgery, quality of sleep before surgery, surgical experience, and anesthesia experience as the primary factors influencing preoperative anxiety in patients awaiting surgery. Age, surgical experience, and anesthesia experience were identified as the main factors affecting informational needs in the preoperative holding area.

CONCLUSIONS

Patients undergoing surgery in the preoperative holding area exhibit heightened levels of anxiety and informational needs. Nurses must provide enhanced psychological support interventions for these patients, particularly focusing on those who are older, female, undergoing repeated operations, gynecological surgeries, experiencing poor sleep quality before surgery, or have had distressing surgical or anesthesia experiences.

SleepECG-Net: Explainable Deep Learning Approach With ECG for Pediatric Sleep Apnea Diagnosis

Obstructive sleep apnea (OSA) in children is a prevalent and serious respiratory condition linked to cardiovascular morbidity. Polysomnography, the standard diagnostic approach, faces challenges in accessibility and complexity, leading to underdiagnosis. To simplify OSA diagnosis, deep learning (DL) algorithms have been developed using cardiac signals, but they often lack interpretability. Our study introduces a novel interpretable DL approach (SleepECG-Net) for directly estimating OSA severity in at-risk children. A combination of convolutional and recurrent neural networks (CNN-RNN) was trained on overnight electrocardiogram (ECG) signals. Gradient-weighted Class Activation Mapping (Grad-CAM), an eXplainable Artificial Intelligence (XAI) algorithm, was applied to explain model decisions and extract ECG patterns relevant to pediatric OSA. Accordingly, ECG signals from the semi-public Childhood Adenotonsillectomy Trial (CHAT, n = 1610) and Cleveland Family Study (CFS,n = 64), and the private University of Chicago (UofC, n = 981) databases were used. OSA diagnostic performance reached 4-class Cohen's Kappa of 0.410, 0.335, and 0.249 in CHAT, UofC, and CFS, respectively. The proposal demonstrated improved performance with increased severity along with heightened cardiovascular risk. XAI findings highlighted the detection of established ECG features linked to OSA, such as bradycardia-tachycardia events and delayed ECG patterns during apnea/hypopnea occurrences, focusing on clusters of events. Furthermore, Grad-CAM heatmaps identified potential ECG patterns indicating cardiovascular risk, such as P, T, and U waves, QT intervals, and QRS complex variations. Hence, SleepECG-Net approach may improve pediatric OSA diagnosis by also offering cardiac risk factor information, thereby increasing clinician confidence in automated systems, and promoting their effective adoption in clinical practice.

Development and Validation of a Nomogram for Predicting Obstructive Sleep Apnea Severity in Children

Purpose

The clinical presentation of Obstructive Sleep Apnea (OSA) in children is insidious and harmful. Early identification of children with OSA, particularly those at a higher risk for severe symptoms, is essential for making informed clinical decisions and improving long-term outcomes. Therefore, we developed and validated a risk prediction model for severity in Chinese children with OSA to effectively identify children with moderate-to-severe OSA in a clinical setting.

Patients and Methods

From June 2023 to September 2023, we retrospectively analyzed the medical records of 367 Children diagnosed with OSA through portable bedside polysomnography (PSG). Predictor variables were screened using the least absolute shrinkage and selection operator (LASSO) and logistic regression techniques to construct nomogram to predict the severity of OSA. Receiver operating characteristic curve (ROC), calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC) were used to determine the discrimination, calibration, and clinical usefulness of the nomogram.

Results

A total of 367 children with a median age of 84 months were included in this study. Neck circumference, ANB, gender, learning problem, and level of obstruction were identified as independent risk factors for moderate-severe OSA. The consistency indices of the nomogram in the training and validation cohorts were 0.841 and 0.75, respectively. The nomogram demonstrated a strong concordance between the predicted probabilities and the observed probabilities for children diagnosed with moderate-severe OSA. With threshold probabilities ranging from 0.1 to 1.0, the predictive model demonstrated strong predictive efficacy and yielded improved net benefit for clinical decision-making. ROC analysis was employed to classify the children into high and low-risk groups, utilizing the Optimal Cutoff value of 0.39.

Conclusion

A predictive model using LASSO regression was developed and validated for children with varying levels of OSA. This model identifies children at risk of developing OSA at an early stage.

ECG-based convolutional neural network in pediatric obstructive sleep apnea diagnosis

Obstructive sleep apnea (OSA) is a prevalent respiratory condition in children and is characterized by partial or complete obstruction of the upper airway during sleep. The respiratory events in OSA induce transient alterations of the cardiovascular system that ultimately can lead to increased cardiovascular risk in affected children. Therefore, a timely and accurate diagnosis is of utmost importance. However, polysomnography (PSG), the standard diagnostic test for pediatric OSA, is complex, uncomfortable, costly, and relatively inaccessible, particularly in low-resource environments, thereby resulting in substantial underdiagnosis. Here, we propose a novel deep-learning approach to simplify the diagnosis of pediatric OSA using raw electrocardiogram tracing (ECG). Specifically, a new convolutional neural network (CNN)-based regression model was implemented to automatically predict pediatric OSA by estimating its severity based on the apnea-hypopnea index (AHI) and deriving 4 OSA severity categories. For this purpose, overnight ECGs from 1,610 PSG recordings obtained from the Childhood Adenotonsillectomy Trial (CHAT) database were used. The database was randomly divided into approximately 60%, 20%, and 20% for training, validation, and testing, respectively. The diagnostic performance of the proposed CNN model largely outperformed the most accurate previous algorithms that relied on ECG-derived features (4-class Cohen's kappa coefficient of 0.373 versus 0.166). Specifically, for AHI cutoff values of 1, 5, and 10 events/hour, the binary classification achieved sensitivities of 84.19%, 76.67%, and 53.66%; specificities of 46.15%, 91.39%, and 98.06%; and accuracies of 75.92%, 86.96%, and 91.97%, respectively. Therefore, pediatric OSA can be readily identified by our proposed CNN model, which provides a simpler, faster, and more accessible diagnostic test that can be implemented in clinical practice.

MicroNAPS: A Novel Classification for Infants with Micrognathia, Robin Sequence, and Tongue-based Airway Obstruction

Background

Robin sequence (RS) describes a heterogeneous population with micrognathia, glossoptosis, and upper airway obstruction (UAO). Workup, treatment, outcomes assessment, and research inclusion are widely variable. Despite several classifications and algorithms, none is broadly endorsed. The objective of this investigation was to develop and trial a novel classification system designed to enhance clinical communication, treatment planning, prognostication, and research.

Methods

This is a retrospective cross-sectional study. A classification system was developed with five elements: micrognathia, nutrition, airway, palate, syndrome/comorbidities (MicroNAPS). Definitions and a framework for "stage" assignment (R0-R4) were constructed. Stage "tongue-based airway obstruction" (TBAO) was defined for infants with glossoptosis and UAO without micrognathia. MicroNAPS was applied to 100 infants with at least 1-year follow-up. Clinical course, treatment, airway, and feeding characteristics were assessed. Descriptive and analytic statistics were calculated and a P value less than 0.05 was considered significant.

Results

Of the 100 infants, 53 were male. Mean follow-up was 5.0 ± 3.6 years. R1 demonstrated feeding-predominant mild RS for which UAO was managed nonoperatively but gastrostomy tubes were prevalent. R2 was characterized by airway-predominant moderate RS, typically managed with mandibular distraction or tongue-lip adhesion, with few gastrostomy tubes and short lengths-of-stay. R3 denoted severe RS, with similar UAO treatment to R2, but with more surgical feeding tubes and longer admissions. R4 represented a complex phenotype with 33% tracheostomies, protracted hospitalizations, and delayed palatoplasty. R0 ("at risk") and TBAO groups displayed the most variability.

Conclusions

MicroNAPS is easy to use and associated with relevant disease characteristics. We propose its adoption in clinical and research settings.

Thinking Beyond the Device: An Overview of Human- and Equity-Centered Approaches for Health Technology Design

A shift in the traditional technocentric view of medical device design to a human-centered one is needed to bridge existing translational gaps and improve health equity. To ensure the successful and equitable adoption of health technology innovations, engineers must think beyond the device and the direct end user and must seek a more holistic understanding of broader stakeholder needs and the intended context of use early in a design process. The objectives of this review article are (a) to provide rationale for the need to incorporate meaningful stakeholder analysis and contextual investigation in health technology development and biomedical engineering pedagogy, (b) to review existing frameworks and human- and equity-centered approaches to stakeholder engagement and contextual investigation for improved adoption of innovative technologies, and (c) to present case studyexamples of medical device design that apply these approaches to bridge the gaps between biomedical engineers and the contexts for which they are designing.

Alternatives to Polysomnography for the Diagnosis of Pediatric Obstructive Sleep Apnea

Diagnosis of obstructive sleep apnea (OSA) in children with sleep-disordered breathing (SDB) requires hospital-based, overnight level I polysomnography (PSG). Obtaining a level I PSG can be challenging for children and their caregivers due to the costs, barriers to access, and associated discomfort. Less burdensome methods that approximate pediatric PSG data are needed. The goal of this review is to evaluate and discuss alternatives for evaluating pediatric SDB. To date, wearable devices, single-channel recordings, and home-based PSG have not been validated as suitable replacements for PSG. However, they may play a role in risk stratification or as screening tools for pediatric OSA. Further studies are needed to determine if the combined use of these metrics could predict OSA.

Evaluation and diagnosis of pediatric obstructive sleep apnea—An update

Purpose

Formal overnight polysomnography (PSG) is required to diagnose obstructive sleep apnea (OSA) in children with sleep disordered breathing (SDB). Most clinical guidelines do not recommend home-based tests for pediatric OSA. However, PSG is limited by feasibility, cost, availability, patient discomfort, and resource utilization. Additionally, the role of PSG in evaluating disease impact may need to be revised. There is a strong need for alternative testing that can stratify the need for PSG and improve the time to diagnosis of OSA. This narrative review aims to evaluate and discuss innovative approaches to pediatric SDB diagnosis.

Findings

Methods to evaluate pediatric SDB outside of PSG include validated questionnaires, single-channel recordings, incorporation of telehealth, home sleep apnea testing (HSAT), and predictive biomarkers. Despite the promise, no individual metric has been found suitable to replace standard PSG. In addition, their use in combination to diagnose OSA diagnosis still needs to be defined.

Summary

When combined with adjunct assessments, HSAT advancements may accurately evaluate SDB in children and thus minimize the need for overnight in-laboratory PSG. Further studies are required to confirm diagnostic validity vis-à-vis PSG as a reference standard.

Features of Obstructive Sleep Apnea in Children with and without Comorbidities

BACKGROUND

To compare the clinical and polysomnographic features of obstructive sleep apnea (OSA) in children with adenotonsillar hypertrophy (Group A) and comorbidities (Group B).

METHODS

A five-year prospective study using nocturnal polysomnography before and after treatment.

RESULTS

We included 168 patients: 121 in Group A and 47 in Group B, with differences in age (6.5 ± 3 vs. 8.6 ± 4 years; p < 0.001), body mass index (BMI) (18 ± 4 vs. 20 ± 7 kg/m2; p < 0.05), neck circumference (28 ± 4 vs. 30 ± 5 cm; p < 0.05), and obesity (17% vs. 30%; p < 0.05). Group B patients were more likely to have facial anomalies (p < 0.001), macroglossia (p < 0.01), dolichocephaly (p < 0.01), micrognathia (p < 0.001), and prognathism (p < 0.05). Adenotonsillectomy was performed in 103 Group A patients (85%) and 28 Group B patients (60%) (p < 0.001). In B, 13 children (28%) received treatment with continuous positive airway pressure (CPAP) and 2 (4%) with bilevel positive airway pressure (BiPAP), compared with 7 in Group A (6%) (p < 0.001). Maxillofacial surgery was more common in Group B (p < 0.01). Clinical and polysomnography parameters improved significantly in both groups after therapeutic intervention, though Group A showed better results.

CONCLUSIONS

Obesity and facial anomalies are more frequent in childhood OSA patients with comorbidities, who often require combination therapy, such as ventilation and surgery.

An automated heart rate-based algorithm for sleep stage classification: Validation using conventional polysomnography and an innovative wearable electrocardiogram device

Background

The rapid advancement in wearable solutions to monitor and score sleep staging has enabled monitoring outside of the conventional clinical settings. However, most of the devices and algorithms lack extensive and independent validation, a fundamental step to ensure robustness, stability, and replicability of the results beyond the training and testing phases. These systems are thought not to be feasible and reliable alternatives to the gold standard, polysomnography (PSG).

Materials and methods

This validation study highlights the accuracy and precision of the proposed heart rate (HR)-based deep-learning algorithm for sleep staging. The illustrated solution can perform classification at 2-levels (Wake; Sleep), 3-levels (Wake; NREM; REM) or 4- levels (Wake; Light; Deep; REM) in 30-s epochs. The algorithm was validated using an open-source dataset of PSG recordings (Physionet CinC dataset, n = 994 participants, 994 recordings) and a proprietary dataset of ECG recordings (Z3Pulse, n = 52 participants, 112 recordings) collected with a chest-worn, wireless sensor and simultaneous PSG collection using SOMNOtouch.

Results

We evaluated the performance of the models in both datasets in terms of Accuracy (A), Cohen's kappa (K), Sensitivity (SE), Specificity (SP), Positive Predictive Value (PPV), and Negative Predicted Value (NPV). In the CinC dataset, the highest value of accuracy was achieved by the 2-levels model (0.8797), while the 3-levels model obtained the best value of K (0.6025). The 4-levels model obtained the lowest SE (0.3812) and the highest SP (0.9744) for the classification of Deep sleep segments. AHI and biological sex did not affect scoring, while a significant decrease of performance by age was reported across the models. In the Z3Pulse dataset, the highest value of accuracy was achieved by the 2-levels model (0.8812), whereas the 3-levels model obtained the best value of K (0.611). For classification of the sleep states, the lowest SE (0.6163) and the highest SP (0.9606) were obtained for the classification of Deep sleep segment.

Conclusion

The results of the validation procedure demonstrated the feasibility of accurate HR-based sleep staging. The combination of the proposed sleep staging algorithm with an inexpensive HR device, provides a cost-effective and non-invasive solution deployable in the home environment and robust across age, sex, and AHI scores.

Reliability of machine learning to diagnose pediatric obstructive sleep apnea: Systematic review and meta-analysis

BACKGROUND

Machine-learning approaches have enabled promising results in efforts to simplify the diagnosis of pediatric obstructive sleep apnea (OSA). A comprehensive review and analysis of such studies increase the confidence level of practitioners and healthcare providers in the implementation of these methodologies in clinical practice.

OBJECTIVE

To assess the reliability of machine-learning-based methods to detect pediatric OSA.

DATA SOURCES

Two researchers conducted an electronic search on the Web of Science and Scopus using term, and studies were reviewed along with their bibliographic references.

ELIGIBILITY CRITERIA

Articles or reviews (Year 2000 onwards) that applied machine learning to detect pediatric OSA; reported data included information enabling derivation of true positive, false negative, true negative, and false positive cases; polysomnography served as diagnostic standard.

APPRAISAL AND SYNTHESIS METHODS

Pooled sensitivities and specificities were computed for three apnea-hypopnea index (AHI) thresholds: 1 event/hour (e/h), 5 e/h, and 10 e/h. Random-effect models were assumed. Summary receiver-operating characteristics (SROC) analyses were also conducted. Heterogeneity (I 2 ) was evaluated, and publication bias was corrected (trim and fill).

RESULTS

Nineteen studies were finally retained, involving 4767 different pediatric sleep studies. Machine learning improved diagnostic performance as OSA severity criteria increased reaching optimal values for AHI = 10 e/h (0.652 sensitivity; 0.931 specificity; and 0.940 area under the SROC curve). Publication bias correction had minor effect on summary statistics, but high heterogeneity was observed among the studies.

Feasibility of conducting type III home sleep apnoea test in children

OBJECTIVES

To test the feasibility of conducting unattended paediatric type III HSAT and to identify issues for improvements to optimize signal quality.

MATERIAL AND METHODS

Parents were instructed in setting up the unattended HSAT and reported their experiences. Signal quality and causes of signal failure of recordings were assessed.

RESULTS

Forty children were included. Mean age was 5.2 years. Predefined success criteria were met in 53% of recordings. Main causes of signal failure were nasal cannula, pulse-oximetry and battery failure. Sensor fixation techniques were developed and implemented during the study and hence signal quality improved. Seventeen (94%) parents reported HSAT to be either easy or medium hard to use.Conclusions and significance: Unattended paediatric type III HSAT can be conducted at home with acceptable signal quality. Signal quality improved considerably using simple sensor fixation techniques.