Abstract
STUDY OBJECTIVES: This is a feasibility study designed to evaluate the accuracy of thermal infrared imaging (TIRI) as a noncontact method to monitor airflow during polysomnography and to ascertain the chance-corrected agreement (K) between TIRI and conventional airflow channels (nasal pressure [Pn], oronasal thermistor and expired CO2 [P(E)CO2]) in the detection of apnea and hypopnea. DESIGN: Subjects were recruited to undergo polysomnography for 1 to 2 hours, during which simultaneous recordings from electroencephalography, electrooculography, electromyography, respiratory impedance plethysmography, conventional airflow channels, and TIRI were obtained. SETTING: University-affiliated, American Academy of Sleep Medicine-accredited sleep disorders center. PATIENTS OR PARTICIPANTS: Fourteen volunteers without a history of sleep disordered breathing and 13 patients with a history of obstructive sleep apnea were recruited. MEASUREMENTS AND RESULTS: In the detection of apnea and hypopnea, excellent agreement was noted between TIRI and thermistor (kappa = 0.92, Bayesian Credible Interval [BCI] 0.86, 0.96; pkappa = 0.99). Good agreement was noted between TIRI and Pn (kappa = 0.83, BCI 0.70, 0.90; pkappa = 0.98) and between TIRI and P(E)CO2 (kappa = 0.80, BCI 0.66, 0.89; pkappa = 0.94). CONCLUSIONS: TIRI is a feasible noncontact technology to monitor airflow during polysomnography. In its current methodologic incarnation, it demonstrates a high degree of chance-corrected agreement with the oronasal thermistor in the detection of apnea and hypopneas but demonstrates a lesser degree of chance-corrected agreement with Pn. Further overnight validation studies must be performed to evaluate its potential in clinical sleep medicine.