Cross approximate entropy analysis of nocturnal oximetry signals in the diagnosis of the obstructive sleep apnea syndrome

Conf Proc IEEE Eng Med Biol Soc. 2006:2006:6149-52. doi: 10.1109/IEMBS.2006.260540.

Abstract

This study is focused on the analysis of blood oxygen saturation (SaO(2)) and heart rate (HR) from nocturnal oximetry using cross approximate entropy (Cross-ApEn). We assessed its usefulness in screening obstructive sleep apnea (OSA) syndrome. We applied Cross-ApEn(m,r,N) to quantify the asynchrony between paired SaO(2) and HR records of 74 patients (44 with a positive OSA diagnosis and 30 with a negative OSA diagnosis). Cross-ApEn values were significantly lower in the OSA positive group compared with those obtained in the OSA negative group. A receiver operating characteristic (ROC) analysis showed that the best results, in terms of diagnostic accuracy, were achieved with m = 2 and r = 0.6. With these input parameters, the optimum decision threshold was found at 1.7, where we achieved 95.5% sensitivity, 73.3% specificity and 86.5% accuracy. Further analyses should be carried out with new and larger data sets to test the usefulness of our methodology prospectively.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Algorithms
  • Entropy
  • Female
  • Humans
  • Male
  • Middle Aged
  • Models, Statistical
  • Oximetry / instrumentation*
  • Oximetry / methods
  • Pattern Recognition, Automated
  • Polysomnography / instrumentation*
  • Polysomnography / methods
  • Predictive Value of Tests
  • ROC Curve
  • Reproducibility of Results
  • Signal Processing, Computer-Assisted
  • Sleep Apnea, Obstructive / diagnosis*