Feasibility of near-infrared spectroscopic tomography for intraoperative functional cerebral monitoring: a primate study.

OBJECTIVE: The wide-ranging manipulations to the cardiovascular system that frequently occur during cardiac surgery can expose the brain to variations in its blood supply that could prove deleterious. As a first step to developing a resource suitable for monitoring such changes, we detected the hemodynamic events induced in the brain of a primate model, using high-density near-infrared spectroscopy combined with tomographic reconstruction methods and validated the findings using established radiologic and histologic techniques. METHODS: Continuous monitoring of the relative changes in the components of the cerebral hemoglobin signal was performed using high-density near-infrared spectroscopy (270 source-detector channel array) in anesthetized bonnet macaques with the brain exposed to induced ischemia and other acute events. A comparative analysis (exact binomial test) applied to reconstructed 3-dimensional images before and after the events and between cerebral hemispheres, combined with postprocedure magnetic resonance imaging, and postmortem histopathologic examination of the macaques' brains was performed to document and validate the spatial features revealed by the optical findings. RESULTS: Relative changes in the measured and calculated components of the hemoglobin signal, in response to the performed manipulations, revealed substantial concurrence among the reconstructed 3-dimensional images, magnetic resonance imaging of the macaques' brains, and postmortem histopathologic examination findings. Concurrence was seen when the manipulated hemoglobin concentration and associated oxygenation levels were either increased or decreased, and whether they were bilateral or restricted to a specified hemisphere. CONCLUSIONS: Continuous near-infrared spectroscopy tomography has been shown to accurately capture and localize cerebral ischemia, vasodilatation, and hemorrhage in primates in real time. These findings are directly applicable to clinical intraoperative functional cerebral monitoring.