Oxygen Saturation Sample Entropy, a Novel Diagnostic Tool in Sleep Disordered Breathing

Area of Science:

• Physiological signal analysis
• Network Physiology
• Sleep Medicine

Background:

• Sleep disordered breathing (SDB) diagnosis can be complex.
• Current diagnostic methods may require extensive data.
• A simplified approach using physiological signals is needed.

Purpose of the Study:

• To evaluate Network Physiology methods (Sample Entropy, Transfer Entropy) for diagnosing sleep disordered breathing (SDB).
• To determine if these methods can simplify the diagnostic process using sleep study data.
• To assess the potential of using simple physiological signals for SDB classification.

Main Methods:

• Retrospective analysis of sleep study data (2017-2021).
• Calculation of Sample Entropy and Transfer Entropy for heart rate, respiratory rate, oxygen saturation, and nasal flow.
• Validation using Receiver Operating Characteristic (ROC) analysis on separate training and testing datasets.

Main Results:

• Sample Entropy of oxygen saturation significantly differed between SDB and normal sleep.
• ROC analysis showed high accuracy in distinguishing SDB types (AUC 0.943 and 0.785).
• Oxygen saturation Sample Entropy achieved 100% sensitivity for diagnosing obstructive sleep apnoea in the test set.

Conclusions:

• Network Physiology mapping of oxygen saturation aids in distinguishing SDB types.
• This approach has potential for simplified sleep apnoea diagnosis using oximetry.
• Further development could streamline global diagnostic pathways for sleep disorders.