The calibration and validation of a phase-modulated near-infrared cerebral oximeter

This study was undertaken to compare the cerebral oxygenation measured by an experimental phase-modulated near-infrared (NIR) spectroscopy system with capillary saturation estimated from jugular venous oxygen saturation.

Jugular venous catheters were placed in 30 patients undergoing carotid endarterectomy and 194 measurements of venous oxygen saturation were obtained intra operatively. Simultaneous measurement of optical path length at 754, 785, and 816 nm was performed using a phase-modulated near-infrared spectroscopy system. Optical calibration was performed using both an optical bench and a scattering mold. Hemoglobin saturation was calculated from NIR measurements using equations derived from diffusion theory. Capillary saturation was calculated from the arterial and venous saturations.

Jugular venous saturations ranged from 41 to 92%. When calibrated using the optical bench, the NIR estimates of hemoglobin saturation deviated from estimated capillary values by an average of 2.6% bias and 4.3% deviation. No systematic bias was noted. NIR values derived from mold calibration were less accurate and precise (4.6% bias and 6.9% deviation.) Use of the initial venous sample as an in vivo calibration improved the accuracy of the mold calibration but did not alter the performance of the bench calibration.

Under the conditions tested, an experimental phase-modulated near-infrared spectroscopy system calibrated using an optical bench agreed with capillary saturation estimated from jugular venous samples. Further work is necessary to demonstrate valid performance of the system under other conditions.