gms | German Medical Science

Objective: Intraoperative thermography in neurosurgery can be used to correlate changes in brain surface temperature distribution to changes in regional cerebral blood flow (rCBF). This work shows an online framework for cortical perfusion imaging that enables intraoperative evaluation of rCBF by detecting and analyzing an intravenously applied cold bolus.

Methods: Measurements were performed as a proof of concept on five patients that underwent a decompressive hemicraniectomy for malignant MCA stroke. The exposed cortical surface was recorded with an uncooled infrared imaging system providing resolutions up to 30 mK and 125 µm per pixel at 30 cm object distance at a maximum acquisition rate of 60 Hz. During recording of changes in brain surface temperature distribution, a cold bolus of regular saline was applied via either a central or peripheral venous line. Real time visualization, which displayed the cold signal passing the superficial vessels, was realized by a combination of frequency filtering and signal magnification. Detailed evaluation of intensity, duration and arrival time of the cold bolus and tissue perfusion status was accomplished shortly after data acquisition by nonlinear regression analysis. To recognize small temperature gradients besides technical artefacts, noise and interferences from external sources, it is necessary to include appropriate preprocessing algorithms in the analysis workflow.

Results: Although the cerebral infarction zone could be spotted approximately in raw data images, processing steps were necessary for a statement about time-resolved tissue perfusion analysis. Previous studies yielded a low decrease of blood vessel temperature after cold bolus injection followed by an increase to baseline. This behavior was tracked in real time. Detailed analysis revealed quantitative results as the cold signal arrived 17.5 s (± 2.54 s) after injection in large arterial vessels and 21.8 s (± 4.04 s) in the parenchyma. These findings concerning arrival, duration and intensity of the cold bolus were analyzed for all pixels and represented by cortical flow maps matching to the anatomy of the associated regions.

Conclusion: The investigations provide evidence that thermography has the capability to perform intraoperative online analysis of cortical blood flow in a non-invasive and marker-free manner. It might turn into a useful imaging tool to target otherwise indefinite interventions by identifying temporarily occluded vessels.