連續正壓呼吸器（CPAP）為阻塞型睡眠呼吸中止症（OSA）的第一線臨床治療首選，惟經常受限於患者順從性不佳及個體療效差異。如何於治療前準確預測療效，為個人化睡眠醫學的重要課題。本研究提出一套整合新穎訊號特徵擷取與機器學習之連續鼻壓訊號分析框架，用以預測 OSA 患者的 CPAP 療效。本框架將連續鼻壓訊號切分為 30 秒片段，並擷取三大類呼吸動態特徵：（1）拓撲特徵：運用拓撲資料分析（TDA）之持續同調技術，分別對原始訊號與相空間點雲資料進行子層集與 Vietoris–Rips 過濾，提取統計、熵及持續性曲線特徵；（2）圖論特徵：透過加權可視圖轉化為網路結構特徵；（3）時域特徵：作為基準對照。模型建構採用兩階段特徵選擇，先依共線性移除冗餘特徵，再以遞迴式特徵消除法（RFE）篩選最佳特徵子集，並結合支持向量機（SVM）建立預測模型，最後透過 SHAP 分析評估特徵貢獻度。實驗結果顯示，鼻壓訊號之結構性特徵能顯著提升預測效能。單一特徵中以拓撲特徵表現最佳（AUROC = 0.923），優於圖論（0.864）與時域特徵（0.784）。特徵組合中，結合拓撲與圖論特徵之模型表現最優（AUROC = 0.954）。本研究證實鼻壓訊號之非線性動態結構資訊與 CPAP 療效具高度關聯，所建構之框架具備發展為臨床决策支援系統之潛力。

Continuous positive airway pressure (CPAP) is usually the first-line clinical treatment for obstructive sleep apnea (OSA); however, its efficacy is often limited by poor patient compliance and individual therapeutic variations. Accurately predicting treatment outcomes prior to intervention has thus become a crucial challenge in personalized sleep medicine. This study proposes a continuous nasal pressure signal analysis framework that integrates novel signal feature extraction with machine learning to predict CPAP treatment efficacy in OSA patients. The proposed framework segments continuous nasal pressure signals into 30-second epochs and extracts three major categories of respiratory dynamic features: (1) Topological features: persistent homology techniques from topological data analysis (TDA) are applied, utilizing sublevel set and Vietoris–Rips filtrations on the raw signals and phase-space point cloud data, respectively, to extract statistical, entropy-based, and persistence curve features; (2) Graph-theoretic features: network structural features are transformed via weighted visibility graphs; and (3) Time-domain features: used as baseline comparisons. Model construction employs a two-stage feature selection strategy, where redundant features are first removed based on collinearity, followed by the recursive feature elimination (RFE) method to select the optimal feature subset. A support vector machine (SVM) classifier is then used to build the predictive model, and SHAP analysis is finally applied to evaluate feature contributions. Experimental results demonstrate that the structural features extracted from nasal pressure signals significantly improve prediction performance. Among single feature groups, topological features achieve the best performance (AUROC = 0.923), outperforming graph-theoretic (0.864) and time-domain features (0.784). Regarding feature combinations, the model combining topological and graph-theoretic features yields the highest performance (AUROC = 0.954). This study confirms that the non-linear dynamic structural information embedded in nasal pressure signals is highly correlated with CPAP treatment outcomes, and the constructed framework holds strong potential for development into a clinical decision support system.

[1] J. F. Garvey, M. F. Pengo, P. Drakatos, and B. D. Kent, “Epidemiological aspects of obstructive sleep apnea,” Journal of thoracic disease, vol. 7, no. 5, pp. 920–9, 2015.
[2] A. V. Benjafield et al., “Estimation of the global prevalence and burden of obstructive sleep apnoea: a literature-based analysis,” Lancet Respir. Med., vol. 7, no. 8, pp. 687–698, Aug. 2019.
[3] I. Fietze et al., “Prevalence and association analysis of obstructive sleep apnea with gender and age differences – Results of SHIP‐Trend,” J. Sleep Res., vol. 28, no. 5, p. e12770, Oct. 2019.
[4] K. A. Franklin, C. Sahlin, H. Stenlund, and E. Lindberg, “Sleep apnoea is a common occurrence in females,” Eur. Respir. J., vol. 41, no. 3, pp. 610–615, Mar. 2013.
[5] P. E. Peppard, T. Young, J. H. Barnet, M. Palta, E. W. Hagen, and K. M. Hla, “Increased Prevalence of Sleep-Disordered Breathing in Adults,” Am. J. Epidemiol., vol. 177, no. 9, pp. 1006–1014, May 2013.
[6] R. G. Bilyukov et al., “Cognitive Impairment and Affective Disorders in Patients With Obstructive Sleep Apnea Syndrome,” Front. Psychiatry, vol. 9, p. 357, Aug. 2018.
[7] E. K. Stranks and S. F. Crowe, “The Cognitive Effects of Obstructive Sleep Apnea: An Updated Meta-analysis,” Arch. Clin. Neuropsychol., p. acv087, Jan. 2016.
[8] Z. Li, S. Cai, Z. Wang, X. Ding, Q. Wang, and R. Chen, “Impact of excessive daytime sleepiness on attention impairment in obstructive sleep apnea: a cross-sectional observational study,” Eur. Arch. Otorhinolaryngol., vol. 281, no. 10, pp. 5503–5510, Oct. 2024.
[9] S. Javaheri et al., “Sleep Apnea: Types, Mechanisms, and Clinical Cardiovascular Consequences,” J. Am. Coll. Cardiol., vol. 69, no. 7, pp. 841–858, Feb. 2017.
[10] D. J. Gottlieb and N. M. Punjabi, “Diagnosis and Management of Obstructive Sleep Apnea: A Review,” JAMA, vol. 323, no. 14, p. 1389, Apr. 2020.
[11] J. M. Marin, S. J. Carrizo, E. Vicente, and A. G. Agusti, “Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study,” The Lancet, vol. 365, no. 9464, pp. 1046–1053, Mar. 2005.
[12] J. Li, L. Zeng, and T. Feng, “The Pathophysiological Relationship and Treatment Progress of Obstructive Sleep Apnea Syndrome, Obesity, and Metabolic Syndrome,” Exploratory Research and Hypothesis in Medicine, vol. 0, no. 0, Jan. 2025.
[13] S. P. Patil, I. A. Ayappa, S. M. Caples, R. J. Kimoff, S. R. Patel, and C. G. Harrod, “Treatment of Adult Obstructive Sleep Apnea with Positive Airway Pressure: An American Academy of Sleep Medicine Clinical Practice Guideline,” Journal of Clinical Sleep Medicine, vol. 15, no. 02, pp. 335–343, Feb. 2019.
[14] V. K. Kapur et al., “Clinical Practice Guideline for Diagnostic Testing for Adult Obstructive Sleep Apnea: An American Academy of Sleep Medicine Clinical Practice Guideline,” Journal of Clinical Sleep Medicine, vol. 13, no. 03, pp. 479–504, Mar. 2017.
[15] R. B. Berry et al., “Rules for Scoring Respiratory Events in Sleep: Update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events,” Journal of Clinical Sleep Medicine, vol. 8, no. 5, Oct. 2012.
[16] R. B. Berry et al., “AASM Scoring Manual Updates for 2017 (Version 2.4),” Journal of Clinical Sleep Medicine, vol. 13, no. 5, pp. 665–666, 2017.
[17] S. P. Patil, I. A. Ayappa, S. M. Caples, R. J. Kimoff, S. R. Patel, and C. G. Harrod, “Treatment of Adult Obstructive Sleep Apnea With Positive Airway Pressure: An American Academy of Sleep Medicine Systematic Review, Meta-Analysis, and GRADE Assessment,” J. Clin. Sleep Med., vol. 15, no. 02, pp. 301–334, Feb. 2019.
[18] T. E. Weaver et al., “Relationship Between Hours of CPAP Use and Achieving Normal Levels of Sleepiness and Daily Functioning,” Sleep, vol. 30, no. 6, pp. 711–719, Jun. 2007.
[19] C. Ratneswaran, M. K. Sagoo, and J. Steier, “Preface for the 3rd Clinical Update Sleep, 23rd February 2018, Royal College of Physicians, London, UK: year in review,” Journal of Thoracic Disease, vol. 10, no. S1, pp. S1–S23, Jan. 2018.
[20] M.-C. Rotty et al., “Mask side-effects in long-term CPAP-patients impact adherence and sleepiness: the InterfaceVent real-life study,” Respiratory Research, vol. 22, no. 1, Jan. 2021.
[21] G. Balasubramani, T. G. S. Raja, and P. Pradeep J, “Adverse effects of Continuous Positive Airway Pressure (CPAP) Therapy: A Comprehensive Review,” International Journal of Respiratory Medicine, vol. 2, no. 1, pp. 1–11, Jan. 2025.
[22] B. W. Rotenberg, D. Murariu, and K. P. Pang, “Trends in CPAP adherence over twenty years of data collection: a flattened curve,” Journal of Otolaryngology - Head & Neck Surgery, vol. 45, no. 1, Aug. 2016.
[23] R. J. Schwab et al., “An Official American Thoracic Society Statement: Continuous Positive Airway Pressure Adherence Tracking Systems. The Optimal Monitoring Strategies and Outcome Measures in Adults,” American Journal of Respiratory and Critical Care Medicine, vol. 188, no. 5, pp. 613–620, Sep. 2013.
[24] G. Scioscia et al., “Machine learning-based prediction of adherence to continuous positive airway pressure (CPAP) in obstructive sleep apnea (OSA),” vol. 47, no. 3, pp. 274–282, Nov. 2021.
[25] K. Ramar et al., “Clinical Practice Guideline for the Treatment of Obstructive Sleep Apnea and Snoring with Oral Appliance Therapy: An Update for 2015,” Journal of Clinical Sleep Medicine, vol. 11, no. 7, Jul. 2015.
[26] J.-L. Pépin, P. Eastwood, and D. J. Eckert, “Novel avenues to approach non-CPAP therapy and implement comprehensive obstructive sleep apnoea care,” European Respiratory Journal, vol. 59, no. 6, p. 2101788, Nov. 2021.
[27] J.-E. Lo et al., “The combination of physiology and machine learning for prediction of CPAP pressure and residual AHI in OSA,” Journal of clinical sleep medicine : JCSM : official publication of the American Academy of Sleep Medicine, vol. 21, no. 5, pp. 775–782, Jan. 2025.
[28] K. Balakrishnan, K. T. James, and E. M. Weaver, “Predicting CPAP Use and Treatment Outcomes Using Composite Indices of Sleep Apnea Severity,” Journal of Clinical Sleep Medicine, vol. 12, no. 06, pp. 849–854, Jun. 2016.
[29] J. H. Choi, Y. Lee, S. Shin, T. K. Ha, and S. Suh, “Demographic and Sleep Study Factors Influencing Short-Term Adherence to Positive Airway Pressure Therapy in Obstructive Sleep Apnea,” Journal of Clinical Medicine, vol. 14, no. 11, p. 3988, Jun. 2025.
[30] S.-E. Cho et al., “Predictors of Continuous Positive Airway Pressure Adherence and Comparison of Clinical Factors and Polysomnography Findings Between Compliant and Non-Compliant Korean Adults With Obstructive Sleep Apnea,” Psychiatry Investigation, vol. 21, no. 2, pp. 200–207, Feb. 2024.
[31] Y. Fujita et al., “Variability of breathing during wakefulness while using CPAP predicts adherence,” Respirology, vol. 22, no. 2, pp. 386–393, Sep. 2016.
[32] A. M. May and J. E. Dalton, “Comparison of machine learning approaches for positive airway pressure adherence prediction in a veteran cohort,” Frontiers in Sleep, vol. 3, Feb. 2024.
[33] Takehiko Yamanashi et al., “Topological data analysis (TDA) enhances bispectral EEG (BSEEG) algorithm for detection of delirium,” vol. 11, no. 1, Jan. 2021.
[34] A. Sathyanarayana, S. Manjunath, and J. A. Perea, “Topological Data Analysis Based Characteristics of Electroencephalogram Signals in Children With Sleep Apnea,” Journal of sleep research, vol. 34, no. 6, p. e70017, Dec. 2025.
[35] Y.-M. Chung, W. K. Huang, and H.-T. Wu, “Topological data analysis assisted automated sleep stage scoring using airflow signals,” Biomedical Signal Processing and Control, vol. 89, p. 105760, Mar. 2024.
[36] C.-T. Lin, “Characterization of polysomnogram signals for surgical response in obstructive sleep apnea using machine learning,” M.S. thesis, Department of Biomedical Engineering, National Cheng Kung University, Tainan, Taiwan, 2024.
[37] J.-C. Kang, “Machine Learning on Weighted Visibility Graph Features with Fuzzy Inference System for Surgical Outcome Prediction in OSA,” M.S. thesis, Department of Biomedical Engineering, National Cheng Kung University, Tainan, Taiwan, 2025.
[38] Y.-M. Chung, C.-S. Hu, Y.-L. Lo, and H.-T. Wu, “A Persistent Homology Approach to Heart Rate Variability Analysis With an Application to Sleep-Wake Classification,” Frontiers in Physiology, vol. 12, Mar. 2021.
[39] P. Bendich et al., “Topological and statistical behavior classifiers for tracking applications,” IEEE Transactions on Aerospace and Electronic Systems, Dec. 2016.
[40] H. Chintakunta, T. Gentimis, R. Gonzalez-Diaz, M.-J. Jimenez, and H. Krim, “An entropy-based persistence barcode,” Pattern Recognition, vol. 48, no. 2, pp. 391–401, Feb. 2015.
[41] Y.-M. Chung, M. Hull, A. Lawson, and N. Pritchard, “Gaussian Persistence Curves,” Journal of Statistical Theory and Practice, vol. 20, no. 1, Jan. 2026.
[42] C. Cortes and V. Vapnik, “Support-vector networks,” Machine Learning, vol. 20, no. 3, pp. 273–297, Sep. 1995.
[43] S. M. Lundberg and S.-I. Lee, "A unified approach to interpreting model predictions," Advances in Neural Information Processing Systems, vol. 30, 2017.
[44] B. Lam, “Craniofacial profile in Asian and white subjects with obstructive sleep apnoea,” Thorax, vol. 60, no. 6, pp. 504–510, Jun. 2005.
[45] J.-H. Byun, K. T. Kim, H. Moon, G. K. Motamedi, and Y. W. Cho, “The first night effect during polysomnography, and patients’ estimates of sleep quality,” Psychiatry Research, vol. 274, pp. 27–29, Apr. 2019.
[46] X. Qian et al., “A Review of Methods for Sleep Arousal Detection Using Polysomnographic Signals,” Brain Sciences, vol. 11, no. 10, p. 1274, Sep. 2021.
[47] T. E. Weaver and R. R. Grunstein, “Adherence to continuous positive airway pressure therapy: the challenge to effective treatment,” Proceedings of the American Thoracic Society, vol. 5, no. 2, pp. 173–178, Feb. 2008.
[48] J. C. Borel et al., “Type of Mask May Impact on Continuous Positive Airway Pressure Adherence in Apneic Patients,” PLoS ONE, vol. 8, no. 5, p. e64382, May 2013.
[49] G. Carlsson, “Topology and data,” Bulletin of the American Mathematical Society, vol. 46, no. 2, pp. 255–308, Jan. 2009.
[50] J. J. Hosselet, R. G. Norman, I. Ayappa, and D. M. Rapoport, “Detection of flow limitation with a nasal cannula/pressure transducer system,” American Journal of Respiratory and Critical Care Medicine, vol. 157, no. 5 Pt 1, pp. 1461–1467, May 1998.
[51] Sushmita Pamidi et al., “An Official American Thoracic Society Workshop Report: Noninvasive Identification of Inspiratory Flow Limitation in Sleep Studies.,” PubMed, vol. 14, no. 7, pp. 1076–1085, Jul. 2017.
[52] R. Guido, S. Ferrisi, D. Lofaro, and D. Conforti, “An Overview on the Advancements of Support Vector Machine Models in Healthcare Applications: A Review,” Information, vol. 15, no. 4, p. 235, Apr. 2024.