Automatic Segmentation and Quantification of Upper Airway Anatomic Risk Factors for Obstructive Sleep Apnea on Unprocessed Magnetic Resonance Images

Soft palate angle and basihyoid depth increase with tongue size and with body condition score in horses

BACKGROUND

Obesity has been associated with human obstructive sleep apnoea and canine brachycephalic obstructive airway syndrome. The effect of body condition score (BCS) on structures of the oropharynx, nasopharynx and upper airway of the horse has not been investigated.

OBJECTIVES

To investigate the effect of BCS on tongue measurements, soft palate angle and basihyoid depth in horses.

STUDY DESIGN

Retrospective, analytical, cross-sectional.

METHODS

Computed tomographic (CT) images of the head of 58 horses were assessed. DICOM viewing software was used to measure head length, basihyoid-skin depth, soft palate angle (SPA), midline tongue area, dorsoventral height (DVH) of the tongue in two locations and head angle. BCS were assigned during CT examinations. Associations between measurements were tested and following initial calculations, further associations with tongue measurements as a ratio of head length were assessed.

RESULTS

For initial measurements, 44 horses met the inclusion criteria. Addition of head length ratios to tongue measurements resulted in 24 of 44 horses meeting the inclusion criteria for the second set of calculations. Increased BCS led to an increased mean SPA (mean difference = 2.56° $$ {}^{{}^{\circ}} $$ ; p = 0.02) and increased median basihyoid depth (mean difference = 0.246 cm; p = 0.006). Following adjustments made for the effect of head length on tongue measures, significant correlation was identified between SPA and tongue area (Spearman's r = 0.544; p = 0.007); SPA and DVH of the tongue at the level of the hard palate (Spearman's r = 0.562; p = 0.004) and SPA and DVH of the tongue at the lingual process of the basihyoid bone (Spearman's r = 0.690; p < 0.001). No significant correlation was identified between variables and sex.

MAIN LIMITATIONS

The sample size was small and the effect of breed on measures was not studied. Measurements were acquired on a single sagittal CT plane. The investigator collecting CT measures was not blinded to BCS. All horses were sedated for the CT procedure which may have affected measures obtained.

CONCLUSIONS

Increased BCS increases SPA and basihyoid bone depth. Increases in tongue size measurements increase SPA. Results from this study warrant further investigation into the clinical significance of the effects of BCS on the upper airways of the horse.

Knowledge, attitude, and practice of inpatients with cardiovascular disease regarding obstructive sleep apnea

There is a significant interrelationship between cardiovascular disease and obstructive sleep apnea (OSA), as they share common risk factors and comorbidities. This study aimed to investigate the knowledge, attitude, and practice (KAP) of inpatients with cardiovascular disease towards OSA. This cross-sectional study was conducted between January, 2022 and January, 2023 at Zhongda Hospital Affiliated to Southeast University among inpatients with cardiovascular disease using a self-administered questionnaire. A self-designed questionnaire was used to assess KAP, and the STOP-Bang questionnaire was applied to evaluate participants' OSA risk. Spearman correlation and path analyses were conducted to explore relationships among KAP scores and high OSA risk. Subgroup analyses were conducted within the high-risk population identified by the STOP-Bang questionnaire. In a study analyzing 591 questionnaires, 66.33% were males. Mean scores were 6.81 ± 4.903 for knowledge, 26.84 ± 4.273 for attitude, and 14.46 ± 2.445 for practice. Path analysis revealed high risk of OSA positively impacting knowledge (β = 2.351, P < 0.001) and practice (β = 0.598, P < 0.001) towards OSA. Knowledge directly affected attitude (β = 0.544) and practice (β = 0.139), while attitude influenced practice (β = 0.266). Among high OSA risk individuals, knowledge directly impacted attitude (β = 0.645) and practice (β = 0.133). Knowledge indirectly influenced practice via attitude (β = 0.197). Additionally, attitude directly affected practice (β = 0.305). These findings provide insights into the interplay between OSA risk, knowledge, attitude, and practice. Inpatients with cardiovascular disease demonstrated inadequate knowledge, moderate attitude, and practice towards OSA. The findings highlighting the need for targeted educational interventions to improve awareness and management of OSA.

Evaluating deep learning techniques for identifying tongue features in subthreshold depression: a prospective observational study

Objective

This study aims to evaluate the potential of using tongue image features as non-invasive biomarkers for diagnosing subthreshold depression and to assess the correlation between these features and acupuncture treatment outcomes using advanced deep learning models.

Methods

We employed five advanced deep learning models-DenseNet169, MobileNetV3Small, SEResNet101, SqueezeNet, and VGG19_bn-to analyze tongue image features in individuals with subthreshold depression. These models were assessed based on accuracy, precision, recall, and F1 score. Additionally, we investigated the relationship between the best-performing model's predictions and the success of acupuncture treatment using Pearson's correlation coefficient.

Results

Among the models, SEResNet101 emerged as the most effective, achieving an impressive 98.5% accuracy and an F1 score of 0.97. A significant positive correlation was found between its predictions and the alleviation of depressive symptoms following acupuncture (Pearson's correlation coefficient = 0.72, p<0.001).

Conclusion

The findings suggest that the SEResNet101 model is highly accurate and reliable for identifying tongue image features in subthreshold depression. It also appears promising for assessing the impact of acupuncture treatment. This study contributes novel insights and approaches to the auxiliary diagnosis and treatment evaluation of subthreshold depression.

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.

Achieving Better Understanding of Obstructive Sleep Apnea Treatment Effects on Cardiovascular Disease Outcomes through Machine Learning Approaches: A Narrative Review

Obstructive sleep apnea (OSA) affects almost a billion people worldwide and is associated with a myriad of adverse health outcomes. Among the most prevalent and morbid are cardiovascular diseases (CVDs). Nonetheless, randomized controlled trials (RCTs) of OSA treatment have failed to show improvements in CVD outcomes. A major limitation in our field is the lack of precision in defining OSA and specifically subgroups with the potential to benefit from therapy. Further, this has called into question the validity of using the time-honored apnea-hypopnea index as the ultimate defining criteria for OSA. Recent applications of advanced statistical methods and machine learning have brought to light a variety of OSA endotypes and phenotypes. These methods also provide an opportunity to understand the interaction between OSA and comorbid diseases for better CVD risk stratification. Lastly, machine learning and specifically heterogeneous treatment effects modeling can help uncover subgroups with differential outcomes after treatment initiation. In an era of data sharing and big data, these techniques will be at the forefront of OSA research. Advanced data science methods, such as machine-learning analyses and artificial intelligence, will improve our ability to determine the unique influence of OSA on CVD outcomes and ultimately allow us to better determine precision medicine approaches in OSA patients for CVD risk reduction. In this narrative review, we will highlight how team science via machine learning and artificial intelligence applied to existing clinical data, polysomnography, proteomics, and imaging can do just that.

Oxygen desaturation index, lowest arterial oxygen saturation and time spent below 90% oxygen saturation as diagnostic markers for obstructive sleep apnea

BACKGROUND

Obstructive sleep apnea (OSA) syndrome is associated with a high mortality, and blood oxygen indexes play an important role in evaluating this disease. The objective of this study was to explore the value of blood oxygen indexes, including minimum oxygen saturation (LSpO2), oxygen reduction index (ODI) and time spent with oxygen saturation below 90% (TS 90%), as diagnostic markers for OSA syndrome.

METHODS

In this retrospective study, 320 patients with OSA treated in Ningbo First Hospital from June 2018 to June 2021 were included and divided into mild, moderate, and severe groups according to the severity of the condition (n = 104, 92, and 124, respectively). The blood oxygen indexes as well as the apnea-hypopnea index (AHI) were compared. The Spearman correlation analysis was performed to explore the relationship between the parameters. Receiver operating characteristic curves were generated to evaluate the diagnostic value of the blood oxygen indexes for OSA syndrome.

RESULTS

There were significant differences in body weight, body mass index, and blood pressure before and after sleep among the groups (P < 0.05). LSpO2 levels followed a pattern with the severe group showing the lowest values, followed by the moderate group, and then the mild group, whereas ODI and TS 90% levels showed the opposite (P < 0.05). Spearman correlation analysis showed that AHI, ODI, TS 90% were positively correlated with severity of OSA, whereas LSpO2 was negatively correlated with severity of OSA. ODI showed a high diagnostic value for OSA (area under curve (AUC) = 0.823, 95% CI: 0.730-0.917). TS 90% showed a high diagnostic value for OSA (AUC = 0.872, 95% CI: 0.794-0.950). LSpO2 showed high accuracy in diagnostic value for OSA (AUC = 0.716, 95% CI: 0.596-0.835). The combination of the 3 indexes demonstrated a high diagnostic value for OSA (AUC = 0.939, 95% CI: 0.890-0.989). The diagnostic value of the combined signature was found to be significantly higher compared to the value of individual indexes (P < 0.05).

CONCLUSION

The evaluation of the severity of OSA should not rely solely on a single observation index, but rather on a combination of ODI, LSpO2 and TS 90%. This combined diagnostic signature can provide a more comprehensive assessment of the patient's condition and serve as an alternative diagnostic basis to ensure timely diagnosis and appropriate clinical treatment for OSA.