In vivo visualization of the renal microvascular perfusion in mouse was obtained through the subcostal flank incision. Data were acquired over a region of 2.5 mm by 2.5 mm on the exposed mouse kidney. By applying UHS-OMAG algorithm on these data, the cross-sectional OCT structure image and the corresponding-OMAG blood flow can be obtained as show in
. From the structural image, the capsule and renal cortex can be observed. Notably, the maximal penetration depth is around ~1 mm in renal tissue. This limited imaging depth is in part due to 1) the high scattering property of renal cortex which prevents the light from penetrating deeper in the tissue, and 2) the dense microvasculature network into which the blood moves within the cortex that attenuate the penetrating light. Luckily, within this imaging depth plenty of capillary vessels (peritubular capillaries) could be detected as shown in , where all small bright dots represent capillary vessels as pointed by small arrows. Thinner cortical tissue would make the arcuate artery visible as pointed by the open arrow.
Fig. 2 Typical in vivo UHS-OMAG imaging results of the renal microcirculation in adult mouse. (A) Shows one typical cross-sectional OCT image of the mouse kidney structure, and (B) is corresponding UHS-OMAG blood flow image of (A). Small white arrows point to (more ...)
Rendered with 3D visualization software Amira 5.01 (Visage Imaging, Inc.), 3D volumetric perfusion image of renal microvasculature were shown in . We interpret the dark holes to be uriniferous tubules which surrounded by capillary networks. We also provided 3D reconstruction of the micro-perfusion map merged with the structure image as shown in . Single capillary vessel can be resolved clearly from the dense micro-perfusion map. In order to better show the network of the capillaries, an en face image of the microvasculature at the depth of 100 µm below the surface was presented in , where capillary vessels are in loop structures. Most of them are peritubular capillaries in the cortical nephron. shows a 3D view of the renal microvasculature. Notably, we can see some luminescent areas which are supposed to be the renal tubules. To the best of our knowledge, this is the first demonstration of using OCT to visualize the detailed capillary network within renal cortex in vivo.
One of the advantages of OMAG versus other imaging techniques like LSI is its capability of giving depth-resolved information. To show the vessel distribution along the depth direction better, we segmented and rendered the perfusion maps in different depths.
is the 3D perfusion map that shows vessels in all depths of a kidney sample. shows the en face perfusion map within 0-200 μm, 200-400μm, 400-600 μm, 600 + μm, respectively. By looking at vessels within these four depths, we found that the superficial layer contains mostly capillaries, while the deeper layer contains some bigger vessels. These big vessels are supposed to be interlobular arterioles and venules. Using the depth-resolved perfusion map would benefit the identification of abnormal renal function in depth.
Fig. 3 Typical image of renal microvasculature at different depth (distance from surface): (a) 3D volumetric perfusion map, maximal intensity projection map at depth of (b) 0-200 um, (c) 200-400um, (d) 400-600 um, (e) > 600 um. The imaged region is ~1.5 (more ...)
are the cross-sectional structure and corresponding blood flow images before renal ischemia, while (c) and (d) are the structure and blood flow images right after renal ischemia captured at the same cross-section. Obvious change can be seen in the blood flow image since in almost no blood flow was detected while the structure images kept almost unchanged. The functional vessel density was estimated from the cross-sectional blood flow image. In doing so, the flow images were first transformed into a binary format by setting a fixed intensity threshold. The percentage of pixels with binary value of 1s versus pixel numbers of the whole image was calculated as the estimation of vessel density. shows the relative vessel density (normalized to the baseline that was captured at the normal state) change during the progress with 30 seconds interval between the time points. The functional blood vessel density reduced gradually when the renal artery was blocked until almost no blood flow could be detected which indicates almost the complete renal ischemia. When the artery was released, abrupt reperfusion occurs as the blood vessel density rise back to around 80% of the initial stage.
Fig. 4 Monitor of the process of renal ischemia using UHS-OMAG. (a, c) are the cross-sectional structure images captured before and after renal ischemia, and (b, d) are the corresponding UHS-OMAG flow images. (e) The normalized vessel density change during the (more ...)
The changes of the 3D rendered functional microvasculature were shown in
, where the functional vessel density reduction could be observed directly after renal ischemia. Within one minute, renal ischemia resulted in an obvious reduction of capillary flow within cortex. Meanwhile, we noticed that the capillaries stop flowing sooner than the interlobular arterioles or venules [as can be seen from , where almost no capillary left and only afferent arteriole/ venules and interlobular artery/vein can be seen]. At the third minute, almost all the blood flow stopped; only several vessels could be seen from which represents severe renal ischemia. Reperfusion turned out to be faster than ischemia. Once the loop was released, an abrupt reperfusion happened within the first half minute . The functional microvasculature at this time point shows a vessel density that is similar to that at the baseline (0 min). And one minute later no more further change was observed as shown in .
Representative 3D renal microcirculation perfusion maps show the process of renal ischemia and reperfusion. The imaged region shown here is ~1.5 x 1.5 mm2.
To examine in more detail about the dynamic change of microvasculature in response to ischemia, we resolved the phase difference between A-lines within subsequent B frames of the UHS-OMAG flow images to calculate the flow velocity. In order to eliminate the influence of background phase noise, phase differences were only calculated when the structural signal is 15dB above the noise floor. Longitudinal quantitative analysis was done on the blood flow velocity of single capillary vessel and the result was shown in
. We choose a typical capillary vessel using both the 2D cross-sectional flow image and the 3D capillary network map as shown in panel (a). Then we applied the phase analysis method discussed above to calculate the velocity of this capillary and the velocity change during ischemia process as shown in . The blood flow velocity in the capillary started to decrease immediately after the renal artery was clamped. Then it decreased almost linearly and stopped flowing in ~6 seconds because at this time point the velocity fell to the range of noise level (~6 µm/s). Our UHS-OMAG imaging system which has an imaging speed of 150 fps is capable of capturing this fast response in capillaries. And this result confirms the disappearing of small capillary vessels in at 0.5 min point after the clamp of renal artery.
Fig. 6 (a) Cross-sectional blood flow image (upper) showing the capillary vessel used for velocity evaluation as located by the circle and arrow. This capillary vessel was also pointed out in the bottom 3D capillary map. (b) Blood flow velocity change of the (more ...)
Our method has ultrahigh sensitivity to the slow blood flow as we increased the time interval used for the phase calculation. But one drawback for this is that the unambiguous detectable velocity range is decreased, which means phase wrapping would happen for blood vessels with axial velocity components larger than this range. In our case, the maximal unambiguous detectable velocity is 72 µm/s. In order to avoid the problem of phase wrapping, we managed to find single capillary vessels with proper Doppler angle that given axial velocity component within this range for velocity evaluation.