Automated Prediction of the Apnea-Hypopnea Index from Nocturnal Oximetry Recordings

• Published in: IEEE transactions on biomedical engineering Vol. 59; no. 1; pp. 141 - 149
• Main Authors: Marcos, J. Víctor, Hornero, Roberto, Álvarez, Daniel, Aboy, Mateo, Campo, Félix Del
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.01.2012; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Apnea-hypopnea index; Applied sciences; Approximation methods; Artificial intelligence; Biological and medical sciences; Complexity theory; Computer science; control theory; systems; Connectionism. Neural networks; Diagnosis, Computer-Assisted - methods; Exact sciences and technology; Feature extraction; Female; Humans; Indexes; Male; Medical sciences; Middle Aged; multilayer perceptron neural networks; multiple linear regression; Oximetry - methods; Oxygen - blood; oxygen saturation; Pattern Recognition, Automated - methods; Pneumology; Regression analysis; Reproducibility of Results; Respiratory system : syndromes and miscellaneous diseases; Sensitivity and Specificity; Severity of Illness Index; Sleep; Sleep Apnea Syndromes - blood; Sleep Apnea Syndromes - diagnosis; sleep apnea-hypopnea syndrome; Studies; Time domain analysis; Training; Blood gas; Oxymetry; Nervous system diseases; Oxygen; Sleep apnea syndrome; Respiratory disease; Saturation; Apnea-hypopnea index; multilayer perceptron neural networks; oxygen saturation; Regression analysis; Neural network; Multiple regression; sleep apnea-hypopnea syndrome; Recording; Predictive factor; multiple linear regression; Biomedical engineering
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/TBME.2011.2167971

Nocturnal polysomnography (PSG) is the gold-standard for sleep apnea-hypopnea syndrome (SAHS) diagnosis. It provides the value of the apnea-hypopnea index (AHI), which is used to evaluate SAHS severity. However, PSG is costly, complex, and time-consuming. We present a novel approach for automatic estimation of the AHI from nocturnal oxygen saturation (SaO 2 ) recordings and the results of an assessment study designed to characterize its performance. A set of 240 SaO 2 signals was available for the assessment study. The data were divided into training (96 signals) and test (144 signals) sets for model optimization and validation, respectively. Fourteen time-domain and frequency-domain features were used to quantify the effect of SAHS on SaO 2 recordings. Regression analysis was performed to estimate the functional relationship between the extracted features and the AHI. Multiple linear regression (MLR) and multilayer perceptron (MLP) neural networks were evaluated. The MLP algorithm achieved the highest performance with an intraclass correlation coefficient (ICC) of 0.91. The proposed MLP-based method could be used as an accurate and cost-effective procedure for SAHS diagnosis in the absence of PSG.