Intraoperative Laser Speckle Contrast Imaging For Real-Time Visualization of Cerebral Blood Flow in Cerebrovascular Surgery: Results From Pre-Clinical Studies

Cerebrovascular surgery can benefit from an intraoperative system that conducts continuous monitoring of cerebral blood flow (CBF). Such a system must be handy, non-invasive, and directly integrated into the surgical workflow. None of the currently available techniques, considered alone, meets all these criteria. Here, we introduce the SurgeON™ system: a newly developed non-invasive modular tool which transmits high-resolution Laser Speckle Contrast Imaging (LSCI) directly onto the eyepiece of the surgical microscope. In preclinical rodent and rabbit models, we show that this system enabled the detection of acute perfusion changes as well as the recording of temporal response patterns and degrees of flow changes in various microvascular settings, such as middle cerebral artery occlusion, femoral artery clipping, and complete or incomplete cortical vessel cautery. During these procedures, a real-time visualization of vasculature and CBF was available in high spatial resolution through the eyepiece as a direct overlay on the live morphological view of the surgical field. Upon comparison with indocyanine green angiography videoangiography (ICG-VA) imaging, also operable via SurgeON, we found that direct-LSCI can produce greater information than ICG-VA and that continuous display of data is advantageous for performing immediate LSCI-guided adjustments in real time.


Imaging process
For each stack of acquired speckle image frames, the speckle contrast (K), defined as the ratio of standard deviation of pixel intensities to the mean pixel intensity within a spatio -temporal neighborhood of pixels around every pixel P0 (Eq. S1), was computed.
where ℕ( 0 ) and ℕ( 0 ) are the standard deviation and mean, respectively, in the intensity of all pixels on a defined local neighborhood ℕ( 0 ). Traditionally, K(P0) values are calculated such that ℕ( 0 ) is chosen exclusively in a single image frame, that is, exclusively in the spatial domain; 53 however, this strategy compromises the spatial resolution of the estimated blood flow information. Preserving the spatial resolution is possible by calculating K(P0) values for every pixel across the stack of image frames, that is, exclusively in the temporal domain. 54 However, such a strategy would require at least 40 frames for a robust estimation of blood flow leading to significant temporal averaging and increased latency of the video output. Therefore, to optimize the spatio-temporal resolution and image acquisition times, the SurgeON System calculates speckle contrast using a spatio-temporal pixel-neighborhood of 5 pixels x 5 pixels (spatial window, S) x 5 frames (temporal window, N) around every pixel P0 (Eq. S2) for contrast calculation: 55 The correlation time of intensity fluctuations observed in speckle dynamics over the exposure time of the camera, is known to be inversely proportional to the velocity of the moving scatterers, therefore, providing a means to estimate a blood flow velocity index (BFVI). BFVI can be mathe matically computed from K at every pixel at a given exposure time T of the camera using Eq. S3. 56,57 The SurgeON displays a pseudo-color representation of BFVI values obtained through the use of look up tables.

SurgeON System calibration
We set up an in vitro microfluidic system to test the technical specifications of the SurgeON. The flow was varied across a wide range (4 mm/s -160 mm/s) to cover the practical range of velocities encountered in-vivo. Imaging was done as per the protocol described below and repeated at different exposures (0.5 ms through 8 ms). Care was taken to not disturb or move the setup when changing the exposure times. A syringe filled with rat blood was fitted into a syringe pump (PHD2000, Harvard Apparatus, MA) and connected to a polyethylene tube (internal diameter 1/32"). Blood was infused into the tube at different rates, in accordance with practically observed vascular velocities. Once the flow was steady (about 1 minute after the flow was started), the tube was imaged using the LSCI modality of the SurgeON System prototype. At each flow rate, LSCI was carried out at multiple exposure times of the camera for calibration.
As reported, the coefficient of determination (R-squared) for the logarithmic best fit at each

Supplementary Table 1: BFVI in cortical vessels and percentage of infarcted area
The table show that the CBF (1/tauc) values pre and post cautery in the cortical vessel and the % of the infarcted area at 24 hours post MCAO of all five rats. The % of the infarcted area is significantly higher the rats #2, #4, and #5 compared to rats #1 and #2, and it correlates with relative lower BFVI at 24hours post-cautery in these rats.