Optical coherence tomography for the quantitative study of cerebrovascular physiology

doi:10.1038/jcbfm.2011.19

J Cereb Blood Flow Metab. 2011 June; 31(6): 1339–1345.

Published online 2011 March 2. doi: 10.1038/jcbfm.2011.19

PMCID: PMC3130321

Optical coherence tomography for the quantitative study of cerebrovascular physiology

Vivek J Srinivasan,¹^* Dmitriy N Atochin,² Harsha Radhakrishnan,¹ James Y Jiang,³ Svetlana Ruvinskaya,¹ Weicheng Wu,¹ Scott Barry,³ Alex E Cable,³ Cenk Ayata,^4,⁵ Paul L Huang,² and David A Boas¹

¹Department of Radiology, MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, USA

²Cardiology Division and Cardiovascular Research Center, Massachusetts General Hospital, Charlestown, Massachusetts, USA

³Thorlabs Inc., Newton, New Jersey, USA

⁴Department of Radiology, Stroke and Neurovascular Regulation Laboratory, Massachusetts General Hospital, Charlestown, Massachusetts, USA

⁵Department of Neurology, Stroke Service and Neuroscience Intensive Care Unit, Massachusetts General Hospital, Boston, Massachusetts, USA

^*Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Building 149, Room 2280, 13th Street, Charlestown, MA 02129, USA. E-mail: vjsriniv/at/nmr.mgh.harvard.edu

Received August 17, 2010; Revised November 22, 2010; Accepted December 9, 2010.

This article has been cited by other articles in PMC.

Abstract

Doppler optical coherence tomography (DOCT) and OCT angiography are novel methods to investigate cerebrovascular physiology. In the rodent cortex, DOCT flow displays features characteristic of cerebral blood flow, including conservation along nonbranching vascular segments and at branch points. Moreover, DOCT flow values correlate with hydrogen clearance flow values when both are measured simultaneously. These data validate DOCT as a noninvasive quantitative method to measure tissue perfusion over a physiologic range.

Keywords: hydrogen clearance, optical coherence tomography, optical imaging, vascular biology

Introduction

For the study of cerebrovascular physiology, an ideal cerebral blood flow (CBF) measurement technique should provide high spatial and temporal resolution. The gold standard for determination of regional blood flow is autoradiography (Sakurada et al, 1978). Although autoradiographic methods provide absolute, spatially resolved blood flow measurements, they contain no information about the temporal evolution of CBF changes. The hydrogen clearance method (Aukland et al, 1964) allows in situ measurement of blood flow, but requires insertion of an electrode into tissue and has limited spatial resolution. Methods of flow measurement based on magnetic resonance imaging (Calamante et al, 1999) and positron emission tomography (Heiss et al, 1994) provide maps of CBF, but are limited in their spatial and temporal resolution. Therefore, a simple method that provides real-time three-dimensional, spatially resolved, and absolute CBF measurements would aid in experimental studies of brain functional activation and cerebrovascular pathophysiology.

Doppler optical coherence tomography (DOCT) is a path length-resolved optical imaging technique that can measure velocity axial projections in well-defined microscopic volumes (Chen et al, 1997; Leitgeb et al, 2003). Doppler OCT velocity axial projections can be used to perform absolute measurements of blood flow either by estimating vessel angles (Wang et al, 2007b) or by processing volumetric data (Srinivasan et al, 2010b). Optical coherence tomography angiography has also been developed to visualize red blood cell (RBC) perfusion (Wang et al, 2007a). Variations of this method have been shown to visualize perfusion with improved sensitivity and resolution (Srinivasan et al, 2010a; Vakoc et al, 2009). In this study, we show the self-consistency of DOCT flow measurements by showing flow conservation along nonbranching vascular segments and at branch points. Furthermore, we show the suitability of DOCT for interanimal comparisons by correlating DOCT flow with hydrogen clearance flow measured concurrently in the same brain.

Materials and methods

Sprague–Dawley rats (n=17, weighing 350±30

g) were prepared under isoflurane anesthesia (1.5% to 2.5% in 25% oxygen, 75% air). After tracheotomy and femoral arterial and venous catheterization, the head was fixed in a stereotaxic frame, the scalp was retracted, and a ~5

mm × 5

mm area of the skull over the somatosensory cortex was thinned to translucency. In five rats, OCT imaging was performed through a closed cranial window after removing the skull and dura, using a preparation described previously (Srinivasan et al, 2010b). In the other 12 rats, a small, ~250-μm diameter hole in the skull was made in a parietal cortex area devoid of large pial vessels for insertion of the hydrogen clearance electrode. A thin layer of artificial cerebrospinal fluid (pH=7.4, 37°C) was applied to the skull immediately before flow measurements. All OCT imaging procedures were performed in 2

minutes to avoid significant evaporation. Average blood gas values were Pa_CO2=41.3±2.7

Hg, Pa_O2=122.0±7.3

Hg, and pH=7.37±0.02. Body temperature was maintained between 36.5°C and 37.5°C. All animal procedures were reviewed and approved by the Subcommittee on Research Animal Care at the Massachusetts General Hospital, where these experiments were performed.

A 1,310

nm spectral/Fourier domain OCT microscope (described in Supplementary Information) was used for these experiments. The axial (depth) resolution was 4.7

μm in air (3.5

μm in tissue, full-width at half-maximum), and the transverse resolution was 7.2

μm (full-width at half-maximum). Figure 1 and Supplementary Figure 2 show imaging through a closed cranial window, whereas all other figures show data obtained from animals with thinned skull preparations. For quantitative flow measurements, a DOCT scan protocol and an OCT angiography scan protocol, requiring ~1

minute each, were performed. The DOCT scan generated a three-dimensional map of the axial (z) projection of velocity (v_z(x,y,z)) (Srinivasan et al, 2010b), whereas the angiography scan generated a three-dimensional angiogram showing perfusion (Srinivasan et al, 2010a). An area of 1.04

mm × 1.04

mm was imaged for absolute CBF estimation.

Figure 1

OCT angiography and Doppler OCT (DOCT) in a rat closed cranial window preparation. (A) OCT angiogram. (B) Zoom of boxed region showing arterioles (solid circles) and venules (dotted circles). (C) Cross-sectional image across vessel location showing a (more ...)

Flow at specific locations in vessels was obtained from DOCT velocity axial projections by calculating the velocity flux through the en face (also known as transverse or xy) plane using the following expression (Srinivasan et al, 2010b),

A mathematical equation, expression, or formula.
Object name is jcbfm201119e1.jpg

where integration was performed over a two-dimensional region of interest (ROI) corresponding to the vessel cross-section in the transverse plane at a particular depth (axial location) of z₀. To determine the depth of DOCT velocity images, the inner surface of the dura was assigned a depth of z₀=0

μm. (However, it should be emphasized that z₀ strictly refers to axial location and not to cortical depth. Ideally, the cortical surface should be aligned approximately perpendicular to the axial direction; however, in general, the brain surface may be tilted or curved.)

A vessel location at a given depth z₀ can also be characterized by its mean velocity axial projection,

A mathematical equation, expression, or formula.
Object name is jcbfm201119e2.jpg

and by its cross-sectional area in the transverse plane,

A mathematical equation, expression, or formula.
Object name is jcbfm201119e3.jpg

From the above equations, it follows that flow in a vessel is the product of the vessel transverse cross-sectional area and the mean velocity axial projection.

A mathematical equation, expression, or formula.
Object name is jcbfm201119e4.jpg

The total draining flow at the cortical surface was approximated by summing flow over all ascending venules in the field of view, where ascending venules were identified based on the OCT angiogram. A cortical volume of 2.16

mm³ drained by this flow was calculated from the known cortical surface area (field of view) of 1.08

mm² and an assumed cortical thickness of 2

mm (Paxinos and Watson, 1982). The mass of tissue corresponding to this cortical volume was calculated by assuming a brain tissue density of 1.05

g/mL (Franceschini et al, 2007; Weaver et al, 2001). Thus,

A mathematical equation, expression, or formula.
Object name is jcbfm201119e5.jpg

This represents a novel method of absolute CBF estimation based on flow measurements in individual vessels, which does not rely on Kety-Schmidt methods (Kety, 1951). More details of this procedure are shown in Supplementary Figure 3.

To validate this method of absolute flow estimation, 12 rats were imaged under either 1.5% isoflurane anesthesia (n=6) or 40

mg/kg per h α-chloralose anesthesia (n=6). A breathing mixture of 20% O₂ and 80% N₂ was used. A platinum electrode for hydrogen clearance with a diameter of 100

μm, sampling from a calculated tissue volume of 0.58

mm³, was used (Demchenko et al, 2005). The electrode was inserted to 1

mm depth (Feng et al, 1988) into the cortex. A reference Ag–AgCl electrode was fixed at the base of the tail. Measurements were performed at least 30

minutes after electrode insertion. Optical coherence tomography imaging protocols were performed through the thinned skull in the vicinity of the hydrogen clearance electrode, simultaneously with acquisition of hydrogen clearance curves. To the extent possible, the OCT field of view was chosen to encompass the electrode insertion point while avoiding large pial vessels. Although the axial image resolution was 3.5

μm, we averaged velocity data such that the effective axial resolution of the DOCT velocity image was 10.5

μm. Animals were switched to a breathing mixture containing air and 2.5% hydrogen for 2

minutes, and then switched back to the initial breathing mixture of 20% O₂ and 80% N₂. Hydrogen clearance curves from the polarographic amplifier were digitized using a separate computer with OCT acquisition triggers. Clearance curves were converted to a logarithmic scale, and the linear portion of the curve was fitted to determine the decay time constant which was converted to blood flow (Young, 1980).

Results

An OCT angiogram of the rat somatosensory cortex, obtained as described previously through a closed cranial window (Srinivasan et al, 2010a), revealed the microvasculature with a high spatial resolution (Figure 1A). Microvascular density was conspicuously lower in areas near the diving arterioles (solid circles), but not near the ascending venules (dotted circles) of comparable size (Figure 1B). This result strongly suggests oxygen delivery from the diving arterioles to the tissue. This finding was reproduced in all five rats imaged through closed cranial windows. The angiograms also showed a distinct ‘stripe' pattern down the center of large vessels. When viewed in depth cross-section, vessels showed a backscattering pattern corresponding to the stripe seen in the en face view, with higher backscattering at the top and bottom of the vessel, and lower backscattering near the sides (Figure 1C). This pattern is a consequence of two phenomena: first, the elongation and preferential orientation of the RBC flat face approximately parallel to the vessel wall (Fischer, 1980; Fischer et al, 1978) caused by shear flow and second, an orientation dependence of RBC backscattering (Nilsson et al, 1998). Thus, RBCs in the upper and lower quadrants in the vessel backscatter more than those in the right and left quadrants because of their orientation. We observed that this stripe pattern disappeared with stoppage of shear flow (data not shown).

Doppler OCT enables high-resolution mapping of velocity axial projections (Figure 1D). Distinct regions of positive flow (solid circles) or negative flow (dotted circles) are associated with descending arterioles or ascending venules, respectively (Figures 1E to 1H). Capillaries are not, in general, visualized in the DOCT image, although they are visualized in the OCT angiogram (Figure 1A, Supplementary Figure 2). The magnitude of the velocity vector, or speed, is greater than that of the axial projection of velocity; therefore, the velocity axial projections measured by DOCT are smaller in magnitude than speeds measured previously (Lipowsky, 2005).

To validate DOCT in a thinned skull preparation, we investigated conservation of DOCT flow along nonbranching vascular segments (Figure 2). We selected two transverse locations in a nonbranching vessel (Figure 2A), obtaining two en face DOCT images at two depths corresponding to the two vessel locations (Figure 2B). By selecting an ROI around the vessel at each of the two depths, we obtained the mean velocity axial projection over the ROI, ROI transverse cross-sectional area, and flow, defined as the velocity flux through the ROI. Importantly, as the angle between the vessel long axis and the axial direction decreases (i.e., the vessel orientation becomes more vertical), the transverse cross-sectional area is decreased, whereas the mean velocity axial projection is increased (Figure 2B). Therefore, flow (calculated as the product of area and velocity) is the same at both locations. In Figures 2C to 2E, data from 24 vessels are color coded either red or blue depending on whether the magnitude of the mean velocity axial projection is greater at location 1 (superficial) or at location 2 (deep), respectively. For 22 of 24 vessels, the magnitude of the mean velocity axial projection and the transverse cross-sectional area changed in opposite directions from location 1 to location 2 (Figure 2D). Therefore, flow was highly conserved along nonbranching vascular segments (Figure 2E). Indeed, flow was better conserved in nonbranching vessels than the mean velocity axial projection (P<0.001, Wilcoxon's signed-rank test on magnitude of fractional error between locations).

Figure 2

Investigation of DOCT flow conservation along nonbranching vascular segments in a rat thinned skull preparation. (A) OCT angiogram showing two transverse locations (1 and 2) in a nonbranching vessel. (B) DOCT en face images at two depths corresponding (more ...)

Flow was also found to be conserved before and after the branch point of a vessel. A single ROI was selected before, and two ROIs were selected after a branch point (Figures 3A and 3B). Branch sums and main trunk values of mean velocity axial projection (Figure 3C), transverse cross-sectional area (Figure 3D), and flow (Figure 3E) were compared. A similar analysis showed that flow before a branch point was equal to the sum of flows in the branches after the branch point (n=13; Figure 3E). Again, flow was better conserved than the mean velocity axial projection at branch points (P<0.001, Wilcoxon's signed-rank test on magnitude of fractional error between the main trunk and the branch sum).

Figure 3

Investigation of DOCT flow conservation at branch points in a rat thinned skull preparation. (A) OCT angiogram showing three locations in a branching vessel. (B) DOCT en face images at two depths corresponding to the main trunk location (1) and two branch (more ...)

To validate DOCT absolute flow values, we compared these values with those simultaneously measured by the hydrogen clearance method in the vicinity of the OCT imaging field (Figure 4A). Absolute CBF was calculated from DOCT, as shown in Supplementary Figure 3, using an average of 20 ascending venules per animal (also see equation (5)). By using different anesthetic regimens to vary the resting CBF, we found a linear correlation between CBF measured by DOCT and hydrogen clearance techniques (R²=0.76, least-squares linear fit, P<0.001; Figure 4B). However, the statistically significant linear correlation does not necessarily imply that the two methods agree. Importantly, further analysis revealed that DOCT yielded higher flow values than did hydrogen clearance in 11 of 12 rats. On average, DOCT absolute CBF measurements exceeded hydrogen clearance CBF measurements by 45%.

Figure 4

Validation of DOCT absolute CBF using simultaneous hydrogen clearance in a rat thinned skull preparation. (A) OCT angiogram showing the outline of the hydrogen clearance electrode and field of view for absolute flow measurements as a box. The dural and (more ...)

To further explore this apparent discrepancy, DOCT absolute flow values in concentric annuli are shown versus distance from the point of electrode insertion (Figure 4C). Mean flows and s.e., obtained from measurements across all animals, are shown. If flow data from a particular annulus were not available for one experiment because of the lack of overlap with the OCT field of view, then this annulus from this experiment was excluded from mean and s.e. calculation. This bar graph suggests that the 100-μm diameter electrode causes a local flow reduction over a region with a radius of ~200

μm.

Discussion and conclusions

We demonstrated that DOCT flow measurements obey the expected steady-state properties of blood flow, including conservation along nonbranching vascular segments and at branch points. Importantly, variations in the average velocity axial projection caused by changes in vessel angle are balanced by equal and opposite changes in the transverse cross-sectional area, yielding flow values that do not depend on the vessel angle. It should be noted that to achieve the most reliable measurements, we limited our measurements to within 150

μm of the pial surface, where multiple scattering effects are minimized and signal-to-noise ratio is the highest. Furthermore, we measured only vessels with an axial orientation sufficient to generate a detectable Doppler shift.

Our results show a correlation between absolute flow measurements by DOCT and hydrogen clearance, suggesting that DOCT flowmetry can be used to compare flow between animals. We used ascending venules to measure the total flow draining a given area of the cortex (Srinivasan et al, 2010b), and we normalized this quantity to the estimated tissue mass to obtain an absolute flow measurement. Venules were chosen rather than arterioles because ascending venules are more numerous, flow is less pulsatile, and the axial projection of velocity virtually always falls within the Doppler detection range of our system (0.7 to 11.4

mm/sec). We frequently observed aliasing in arterioles.

Our results show that DOCT absolute CBF measurements exceed hydrogen clearance CBF measurements by 45%, on average. This may be caused by damage introduced by the platinum electrode insertion, causing a local reduction in flow near the electrode. This is supported by Figure 4C, which shows an apparent local reduction in flow within 200

μm of the point of electrode insertion. Other plausible explanations for the discrepancy are deviations in cortical thickness from the assumed value and overestimation of the extent of the flow profile with DOCT. The flow values of DOCT may depend on instrument resolution (axial and transverse), data processing algorithms to obtain flow measurements, and details such as location of optical focus and skull thickness. Supplementary Figure 4 shows that choice of algorithm parameters for processing DOCT data can affect the scale of absolute flow measurements (absolute flow values ranging from 53.3 to 155.2

mL/100

g per min), although DOCT flow remained linearly correlated with hydrogen clearance flow (0.55 [less-than-or-eq, slant]

R²

0.82, P<0.01) regardless of parameter choices. For the results presented in Figure 4B, we chose the DOCT algorithm parameters to yield velocity profiles consistent with vessel calibers determined from the OCT angiogram. In separate validation experiments, vessel calibers in the OCT angiogram agreed qualitatively with those determined by two-photon microscopy, as shown in Supplementary Figure 2.

Finally, we note that DOCT flow measurements require scattering from RBCs. Thus, DOCT measurements may be inaccurate at extremely low hematocrits. In addition, at very low flow rates, it is possible that venular velocity axial projections may not exceed the threshold for detection. Therefore, we cannot assert the accuracy of DOCT flow measurements under all conditions. The correlation between DOCT and hydrogen clearance (Figure 4B) supports the assertion that DOCT CBF values can be compared between animals under physiologic conditions. However, the scale of absolute flow measurements may change depending on a number of experimental or algorithm details. Therefore, we strongly suggest that any future studies using DOCT flowmetry should include controls to account for these effects.

In conclusion, DOCT flowmetry and OCT angiography are suitable for noninvasive and quantitative investigation of cerebrovascular physiology with good spatiotemporal resolution. Doppler OCT flow obeys properties expected of steady-state CBF, including conservation along nonbranching vascular segments and at branch points. Moreover, DOCT absolute CBF values correlate with those obtained by simultaneous hydrogen clearance in the rat cortex. Our DOCT CBF estimate is based on the topology of the rat cortical blood supply and the principle of flow conservation, and does not rely on Kety-Schmidt methods. Therefore, DOCT is a novel method to compare CBF across animals over a physiologically relevant range.

Notes

JYJ, SB, and AEC are employees of Thorlabs Inc.

Footnotes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http://www.nature.com/jcbfm)

This study was supported by the NIH (R01-NS057476, R01-EB001954, R01-NS033335, R01-NS061505, P01NS055104, P50NS010828, and K99NS067050), the AHA (SDG 0835344N and 11IRG5440002), and the AFOSR (MFEL FA9550-07-1-0101).

Supplementary Material

Supplementary Figures S1–S4

Click here for additional data file.^{(1.4M, pdf)}

Supplementary Methods

Click here for additional data file.^{(61K, doc)}

References

Aukland K, Berliner RW, Bower BF. Measurement of local blood flow with hydrogen gas. Circ Res. 1964;14:164–187. [PubMed]
Calamante F, Thomas DL, Pell GS, Wiersma J, Turner R. Measuring cerebral blood flow using magnetic resonance imaging techniques. J Cereb Blood Flow Metab. 1999;19:701–735. [PubMed]
Chen Z, Milner TE, Dave D, Nelson JS. Optical Doppler tomographic imaging of fluid flow velocity in highly scattering media. Opt Lett. 1997;22:64–66. [PubMed]
Demchenko IT, Luchakov YI, Moskvin AN, Gutsaeva DR, Allen BW, Thalmann ED, Piantadosi CA. Cerebral blood flow and brain oxygenation in rats breathing oxygen under pressure. J Cereb Blood Flow Metab. 2005;25:1288–1300. [PubMed]
Feng ZC, Roberts EL, Sick TJ, Rosenthal M. Depth profile of local oxygen-tension and blood-flow in rat cerebral-cortex, white matter and hippocampus. Brain Res. 1988;445:280–288. [PubMed]
Fischer TM. On the energy dissipation in a tank-treading human red blood cell. Biophys J. 1980;32:863–868. [PubMed]
Fischer TM, Stohr-Lissen M, Schmid-Schonbein H. The red cell as a fluid droplet: tank tread-like motion of the human erythrocyte membrane in shear flow. Science. 1978;202:894–896. [PubMed]
Franceschini MA, Thaker S, Themelis G, Krishnamoorthy KK, Bortfeld H, Diamond SG, Boas DA, Arvin K, Grant PE. Assessment of infant brain development with frequency-domain near-infrared spectroscopy. Pediatr Res. 2007;61:546–551. [PMC free article] [PubMed]
Heiss WD, Graf R, Wienhard K, Lottgen J, Saito R, Fujita T, Rosner G, Wagner R. Dynamic penumbra demonstrated by sequential multitracer PET after middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab. 1994;14:892–902. [PubMed]
Kety SS. The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev. 1951;3:1–41. [PubMed]
Leitgeb RA, Schmetterer L, Drexler W, Fercher AF, Zawadzki RJ, Bajraszewski T. Real-time assessment of retinal blood flow with ultrafast acquisition by color Doppler Fourier domain optical coherence tomography. Opt Exp. 2003;11:3116–3121. [PubMed]
Lipowsky HH. Microvascular rheology and hemodynamics. Microcirculation. 2005;12:5–15. [PubMed]
Nilsson AM, Alsholm P, Karlsson A, Andersson-Engels S. T-matrix computations of light scattering by red blood cells. Appl Opt. 1998;37:2735–2748. [PubMed]
Paxinos G, Watson C. The Rat Brain: In Stereotaxic Coordinates. Sydney: Academic Press; 1982.
Sakurada O, Kennedy C, Jehle J, Brown JD, Carbin GL, Sokoloff L. Measurement of local cerebral blood flow with iodo [14C] antipyrine. Am J Physiol. 1978;234:H59–H66. [PubMed]
Srinivasan VJ, Jiang JY, Yaseen MA, Radhakrishnan H, Wu W, Barry S, Cable AE, Boas DA. Rapid volumetric angiography of cortical microvasculature with optical coherence tomography. Opt Lett. 2010a;35:43–45. [PMC free article] [PubMed]
Srinivasan VJ, Sakadzic S, Gorczynska I, Ruvinskaya S, Wu W, Fujimoto JG, Boas DA. Quantitative cerebral blood flow with optical coherence tomography. Opt Exp. 2010b;18:2477–2494. [PMC free article] [PubMed]
Vakoc BJ, Lanning RM, Tyrrell JA, Padera TP, Bartlett LA, Stylianopoulos T, Munn LL, Tearney GJ, Fukumura D, Jain RK, Bouma BE. Three-dimensional microscopy of the tumor microenvironment in vivo using optical frequency domain imaging. Nat Med. 2009;15:1219–1223. [PMC free article] [PubMed]
Wang RK, Jacques SL, Ma Z, Hurst S, Hanson SR, Gruber A. Three dimensional optical angiography. Opt Exp. 2007a;15:4083–4097. [PubMed]
Wang Y, Bower BA, Izatt JA, Tan O, Huang D. In vivo total retinal blood flow measurement by Fourier domain Doppler optical coherence tomography. J Biomed Opt. 2007b;12:041215. [PubMed]
Weaver BM, Staddon GE, Mapleson WW. Tissue/blood and tissue/water partition coefficients for propofol in sheep. Br J Anaesth. 2001;86:693–703. [PubMed]
Young W. H-2 clearance measurement of blood-flow—a review of technique and polarographic principles. Stroke. 1980;11:552–564. [PubMed]

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of
Nature Publishing Group