Publications

Background: The identification of obstructive and central hypopneas is considered challenging in clinical practice. Presently, obstructive and central hypopneas are usually not differentiated or scores lack reliability due to the technical limitations of standard polysomnography. Esophageal pressure measurement is the gold-standard for identifying these events but its invasiveness deters its usage in daily practice. Objectives: To determine the feasibility and efficacy of an automatic noninvasive analysis method for the differentiation of obstructive and central hypopneas based solely on a single-channel nasal airflow signal. The obtained results are compared with gold-standard esophageal pressure scores. Methods: A total of 41 patients underwent full night polysomnography with systematic esophageal pressure recording. Two experts in sleep medicine independently differentiated hypopneas with the gold-standard esophageal pressure signal. Features were automatically extracted from the nasal airflow signal of each annotated hypopnea to train and test the automatic analysis method. Interscorer agreement between automatic and visual scorers was measured with Cohen's kappa statistic (κ). Results: A total of 1,237 hypopneas were visually differentiated. The automatic analysis achieved an interscorer agreement of κ = 0.37 and an accuracy of 69% for scorer A, κ = 0.40 and 70% for scorer B and κ = 0.41 and 71% for the agreed scores of scorers A and B. Conclusions: The promising results obtained in this pilot study demonstrate the feasibility of noninvasive single-channel hypopnea differentiation. Further development of this method may help improving initial diagnosis with home screening devices and offering a means of therapy selection and/or control.

JTD Keywords: Central sleep hypopnea, Esophageal pressure, Home monitoring, Obstructive sleep hypopnea, Sleep disordered breathing

The automatic differentiation of obstructive and central respiratory events is a major challenge in the diagnosis of sleep-disordered breathing. Esophageal pressure (Pes) measurement is the gold-standard method to identify these events. This study presents a new classifier that automatically differentiates obstructive and central hypopneas with the Pes signal and a new approach for an automatic noninvasive classifier with nasal airflow. An overall of 28 patients underwent night polysomnography with Pes recording, and a total of 769 hypopneas were manually scored by human experts to create a gold-standard annotation set. Features were automatically extracted from the Pes signal to train and test the classifiers (discriminant analysis, support vector machines, and adaboost). After a significantly (p < 0.01) higher incidence of inspiratory flow limitation episodes in obstructive hypopneas was objectively, invasively assessed compared to central hypopneas, the feasibility of an automatic noninvasive classifier with features extracted from the airflow signal was demonstrated. The automatic invasive classifier achieved a mean sensitivity, specificity, and accuracy of 0.90 after a 100-fold cross validation. The automatic noninvasive feasibility study obtained similar hypopnea differentiation results as a manual noninvasive classification algorithm. Hence, both systems seem promising for the automatic differentiation of obstructive and central hypopneas.

JTD Keywords: Automatic differentiation, Central hypopnea, Esophageal pressure (Pes), Inspiratory flow limitation (IFL), Noninvasive classification, Obstructive hypopnea

The differentiation of obstructive and central respiratory events is a major challenge in the diagnosis of sleep disordered breathing. Esophageal pressure (Pes) measurement is the gold-standard method to identify these events but its invasiveness deters its usage in clinical routine. Flattening patterns appear in the airflow signal during episodes of inspiratory flow limitation (IFL) and have been shown with invasive techniques to be useful to differentiate between central and obstructive hypopneas. In this study we present a new method for the automatic non-invasive differentiation of obstructive and central hypopneas solely with nasal airflow. An overall of 36 patients underwent full night polysomnography with systematic Pes recording and a total of 1069 hypopneas were manually scored by human experts to create a gold-standard annotation set. Features were automatically extracted from the nasal airflow signal to train and test our automatic classifier (Discriminant Analysis). Flattening patterns were non-invasively assessed in the airflow signal using spectral and time analysis. The automatic non-invasive classifier obtained a sensitivity of 0.71 and an accuracy of 0.69, similar to the results obtained with a manual non-invasive classification algorithm. Hence, flattening airflow patterns seem promising for the non-invasive differentiation of obstructive and central hypopneas.

JTD Keywords: Practical, Experimental/ biomedical measurement, Feature extraction, Flow measurement, Medical disorders, Medical signal processing, Patient diagnosis, Pneumodynamics, Pressure measurement, Signal classification, Sleep, Spectral analysis/ automatic noninvasive differentiation, Obstructive hypopnea, Central hypopnea, Inspiratory flow limitation, Nasal airflow, Esophageal pressure, Polysomnography, Feature extraction, Discriminant analysis, Spectral analysis

New techniques for automatic invasive and noninvasive identification of inspiratory flow limitation (IFL) are presented. Data were collected from 11 patients with full nocturnal polysomnography and gold-standard esophageal pressure (Pes) measurement. A total of 38,782 breaths were extracted and automatically analyzed. An exponential model is proposed to reproduce the relationship between Pes and airflow of an inspiration and achieve an objective assessment of changes in upper airway obstruction. The characterization performance of the model is appraised with three evaluation parameters: mean-squared error when estimating resistance at peak pressure, coefficient of determination, and assessment of IFL episodes. The model's results are compared to the two best-performing models in the literature. The obtained gold-standard IFL annotations were then employed to train, test, and validate a new noninvasive automatic IFL classification system. Discriminant analysis, support vector machines, and Adaboost algorithms were employed to objectively classify breaths noninvasively with features extracted from the time and frequency domains of the breaths' flowpatterns. The results indicated that the exponential model characterizes IFL and subtle relative changes in upper airway obstruction with the highest accuracy and objectivity. The new noninvasive automatic classification system also succeeded in identifying IFL episodes, achieving a sensitivity of 0.87 and a specificity of 0.85.

JTD Keywords: Esophageal pressure, Exponential model, Inspiratory flow limitation, Noninvasive, Classification, Upper airway obstruction