Fractal quantitative endoscopic evaluation of the upper airway in patients with obstructive sleep apnea syndrome

OBJECTIVE: To analyze computer-assisted quantitative measurements of endoscopic images taken during fiberoptic nasopharyngoscopy with Muller's maneuver that were processed with the use of fractal geometry and correlate them with the apnea-hypopnea index of patients with obstructive sleep apnea syndrome as well as habitual snorers. STUDY DESIGN: Diagnostic test assessment for the analysis of images acquired through fiberoptic nasopharyngoscopy with Muller's maneuver and processed with the use of fractal geometry. SETTING: Sleep laboratory in a tertiary care private hospital in Athens, Greece. SUBJECTS AND METHODS: A total of 42 subjects (25 patients, 17 control subjects) underwent fiberoptic endoscopic nasopharyngoscopy with Muller's maneuver and polysomnographic overnight testing. Endoscopic images were digitally analyzed, and fractal analysis software was used to determine fractal dimensions of postpalatal airway contour during quiet inspiration and Muller's maneuver. Findings were correlated with the apnea-hypopnea index. RESULTS: Fractal dimension of collapsed airway and airway collapsibility was shown to be correlated with the apnea-hypopnea index (r = -0.481 and 0.518). We used 9.5 percent airway collapsibility on the basis of fractal dimensions as a cutoff (decided according to a receiver operating characteristic curve) and found that the predictability of obstructive sleep apnea had a sensitivity of 92.0 percent (95% confidence interval 75.03-97.78) and a specificity of 82.4 percent (95% confidence interval 58.97-93.8), with a positive predictive value of 88.5 percent and a negative predictive value of 87.5 percent (P < 0.0001). CONCLUSION: Fractal analysis provides an additional measure of objectivity for the interpretation of computer-assisted quantitative endoscopic evaluation of patients with obstructive sleep apnea. Nevertheless, as confidence intervals show, some uncertainty remains regarding the real population estimate, and therefore additional studies involving larger population groups are encouraged. (C) 2010 American Academy of Otolaryngology-Head and Neck Surgery Foundation. All rights reserved.