WO2010015954A1

WO2010015954A1 - Detection, visualization, and quantification of microvasculature using confocal microscopy

Info

Publication number: WO2010015954A1
Application number: PCT/IB2009/053054
Authority: WO
Inventors: Lyubomir Zagorchev; Mary J. Mulligan-Kehoe
Original assignee: Koninklijke Philips Electronics, N.V.
Priority date: 2008-08-04
Filing date: 2009-07-14
Publication date: 2010-02-11

Abstract

When detecting and/or imaging microvasculature, images slices of a stained tissue sample are generated using a confocal microscope (12). The image slices are aggregated into vertical stacks, or z-stacks. A thresholding or ray casting technique is applied to the z-stacks to segment out stained microvessels (64). A 3D image representation of the z-stacks with segmented microvessels (64) is rendered for viewing.

Description

DETECTION, VISUALIZATION, AND QUANTIFICATION OF MICROVASCULATURE USING CONFOCAL MICROSCOPY

DESCRIPTION

The present innovation finds particular application in anatomic imaging systems, particularly involving confocal microscopy and the like. However, it will be appreciated that the described technique may also find application in other imaging systems, other imaging scenarios, other image analysis techniques, and the like.

Angiogenesis is the process of formation of new blood vessels as outgrowths of pre-existing ones. It occurs naturally during development, tissue repair, and abnormally in pathologic diseases. It is associated with proliferation of blood vessels that penetrate into abnormal tissue areas to supply nutrients and remove waste products. Tumor-induced angiogenesis is the proliferation of blood vessels that penetrate into cancerous growths to supply nutrients and remove waste products. It provides the important link between the quiescent phase of initial tumor growth and the more harmful vascular phase when the tumor is large enough to require extensive vascularization for nutrient supply. The process starts with chemotaxis and locomotion of endothelial cells and is caused by abnormal cells that send signals to surrounding normal tissue. The signaling process is complex, but evidence indicates that the most important part is played by a protein known as vascular endothelial growth factor (VEGF). VEGF is secreted by abnormal cells and diffused in the surrounding tissue. The VEGF protein binds to VEGF receptors on nearby endothelial cells and signals them to produce protease enzymes to degrade the basal lamina of their parent capillary or venule. Once free, the endothelial cells begin to migrate toward the source of the signal and form small tubular sprouts recruited from the parent vessel. The sprouts grow in length due to migration, proliferation, and recruitment of new endothelial cells, and continue to move toward the abnormal cell (tumor, etc.) directed by the motion of the leading endothelial cell at the tip of the sprout. New growth extensions may occur if the endothelial cells of the newly formed sprout's wall begin to proliferate.

Atherosclerosis is a chronic disease of large and medium size arteries and is the most frequent cause of coronary, peripheral and carotid artery disease. Atherosclerotic plaque is an occlusive vascular disease that affects coronary, peripheral and carotid arteries. As the plaque continues to develop, it can become unstable and lead to potential life-threatening rupture, thrombosis, and/or compromised lumen of the artery. The vasa vasorum is the main conduit for nutrient supplies to the arterial wall due to the lack of vessels in the inner media and intima. Angiogenesis is the process of formation of new blood vessels as outgrowths of pre-existing ones. Angiogenic vasa vasorum are associated with more advanced stages of human atherosclerosis and with increased lesion size in hypercholesterolemic animal models. It has been shown that anti-angiogenic molecules can inhibit neovascularization in the vasa vasorum and reduce plaque progression. Neo-angiogenesis associated with more advanced stages of human atherosclerosis is found in plaque and the vasa vasorum, the microvasculature in the adventitial layer of large arteries that provides arterial blood supply to the arterial wall. The presence and extent of vasa vasorum correlate with atherosclerotic lesion size in hypercholesterolemic animal models. In atherosclerosis, vasa vasorum are considered to be the conduit for nutrient supplies to the plaque. It has also been demonstrated that inhibition of neovascularization in the vasa vasorum is associated with reduced plaque progression. At the same time, there is little available information about the origin of plaque vasculature and the role of vasa vasorum in plaque growth. Furthermore, there is no conclusive evidence that angiogenesis in vasa vasorum promotes plaque development. Techniques for detection and visualization of angiogenic vasa vasorum are yet to be developed.

The physical resolution of small animal imaging scanners is not high enough to allow successful 3-D imaging of angiogenic vasa vasorum, or of sprouts, since the size of the sprouts is in the 1-3 micron range. The typical high resolution of micro- CT scanners allows for imaging of objects with size of 5-10 microns, which therefore excludes the imaging of sprouts and vasa vasorum. In addition, contrast agents that can penetrate small sprouts currently do not exist. The capability of available contrast to visualize sprouts is limited by a large number of factors such as injection pressure, polymerization of contrast, size of contrast particles, etc. Despite those facts, quantitative assessment of sprout formation plays an important role in evaluation of angiogenesis inhibitors and tumor growth. The present application provides new and improved systems and methods for imaging angiogenic micro vasculature, which overcome the above -referenced problems and others.

In accordance with one aspect, a microvasculature detection and visualization system includes a processor that receives image data corresponding to a plurality of image slices of a tissue sample stained with a first contrast agent, generates one or more z-stacks from the plurality of image slices, and executes a thresholding process on the one or more z-stacks to segment out micro vessels from other tissue. The system further includes a memory that stores computer-executable instructions for generating the one or more z-stacks, and for segmenting the one or more z-stacks to identify the microvessels.

In accordance with another aspect, a method of detecting and quantifying microvessels in a tissue sample includes generating image slices of the tissue sample, aggregating the image slices to generate one or more 3D z-stacks, segmenting the one or more z-stacks to identify the microvessels, and rendering a graphical representation of the microvessels on a display.

In accordance with another aspect, an apparatus for detecting and quantifying microvessels in a lumen-stained tissue sample includes means for generating image slices of the tissue sample, means for stacking the image slices into one or more z- stacks, and for segmenting the one or more z-stacks to identify microvessels, and means for rendering a 3D representation of the microvessels. The image slices are generated at sub-micron resolution. In accordance with another aspect, a method of determining a volume or surface area of microvessels comprises aggregating tissue image slices to generate one or more z- stacks, segmenting the one or more z-stacks to identify microvessels, and comparing the segmented images to the tissue image slices to obtain the volume or surface area of the microvessels. One advantage is that microvessels are imaged at sub-micron resolution

Another advantage resides in cost savings relative to other imaging techniques. Another advantage resides in imaging microvessels using more than one contrast agent or stain.

Still further advantages of the subject innovation will be appreciated by those of ordinary skill in the art upon reading and understand the following detailed description.

The innovation may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating various aspects and are not to be construed as limiting the invention.

FIGURE 1 illustrates a system that facilitates detecting and quantifying angiogenic microvessels using confocal microscopy and fluorescently labeled samples.

FIGURE 2 is a screenshot of a single z-stack showing segmented micro vasculature such as is generated via execution by the processor of the various algorithms stored in the memory, and displayed on the display.

FIGURE 3 is a screenshot of a plurality of aggregated z-stacks showing aggregated microvasculature in a 3D representation.

FIGURE 4 is a screenshot of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in the adventitia.

FIGURE 5 is a screenshot of detected and visualized microvessels in the smooth muscle after processing by the processor and execution of the computer- executable instructions stored in the memory of the system. FIGURE 6 is a screenshot of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in a blood vessel wall.

FIGURE 7 is a screenshot of detected and visualized microvessels in the vessel wall after processing by the processor and execution of the computer-executable instructions stored in the memory of the system. FIGURE 8 is a screenshot of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in an atherosclerotic plaque.

FIGURE 9 is a screenshot of detected and visualized microvessels in the plaque after processing by the processor and execution of the computer-executable instructions stored in the memory of the system.

FIGURE 1 illustrates a system 10 that facilitates detecting, quantifying, and visualizing angiogenic microvasculature using confocal microscopy and fluorescently labeled samples. Although terms "microvessel" and "microvasculature" are used generally in various examples herein to describe angiogenic vasa vasorum and/or tumor-induced angiogenic vessel sprouts, it will be appreciated that the described systems and methods are applicable to any microvascular tissue and/or structure. For instance, combined with a method for quantification, confocal microscopy presents a unique opportunity for analytical assessment and quantitative studies of sprout growth and proliferation. Additionally, when combined with methods for detection and visualization of blood vessels, confocal microscopy presents a unique opportunity for quantitative studies of angiogenic vasa vasorum in atherosclerotic plaque and provides information about the origin of plaque vasculature and the role of vasa vasorum in plaque growth and development.

The system 10 includes a confocal microscope 12 that is coupled to a user interface 14. The confocal microscope 12 generates image data of fluorescently labeled samples placed therein, and the image data is received at the user interface for processing. A processor 16 executes one or more algorithms (e.g., computer-executable instructions) stored in a memory 18 to generate three-dimensional images, comprising one or more "z- stacks," of the fluorescently labeled samples. In one embodiment, the confocal microscope 12 employs a spatial pinhole that eliminates or reduces out-of-focus light or flare in samples that are thicker than a focal plane of the microscope. For instance, the microscope 12 starts at the top of a tissue sample and takes sequential slice images at different focal depths, which are then stacked to form a z-stack (e.g., a vertical stack of image slices in the z-direction). A plurality of such z-stacks is vertically stacked to generate a 3D image, when more than one z-stack is generated.

In one embodiment the system 10 analyzes the z-stacks and segments out the microvasculature using ray casting, thresholding, or some other suitable volume rendering technique. The system calculates the density of the rendered microvasculature, and optionally calculates the surface area thereof. For instance, a sample of the tissue of interest is treated with a complex, but known technique to stain the microvessels. The confocal microscope is used to take a series of very high resolution images at each of a plurality of depths. This data is digitized by the processor and formed into a 3D digital image representation of the tissue sample. A threshold segmentation is performed to separate the dyed or stained microvessels from other tissue. The volume of the microvessels and the volume of the sample are calculated. From this information, a ratio or other measure of the density of microvessels is determined.

In another embodiment, the tissue sample is stained in such a manner that dead-end microvessels are stained with a first color and flowing microvessels are stained with a different color. This enables a further refinement in the data in which the relative volume of dead-end and flow through lumens are both determined.

According to an example, to prepare for microscopy, subjects (e.g., mice in this example) receive an injection of 0.1 ml ketamine per 30 grams of weight. The subjects are euthanized, the chest is opened, and the subjects are perfused with phosphate buffered saline (PBS) containing calcium and magnesium at a pressure of 100-120 mm/Hg. The subjects are then perfused with 1% paraformaldehyde containing 0.5% glutaraldehyde in PBS, pH 7.4, for 5 minutes under 100-120 mm/Hg pressure followed by two one-minute perfusions consisting of 50 ml PBS, then 50 ml PBS containing 1% bovine serum albumin. Finally, the subjects are perfused with 20 mg biotinylated lectin in 50 ml PBS for one minute.

The target areas are surgically removed, mounted on a slide and imaged on the confocal microscope 12. In one embodiment, the confocal microscope is a Zeiss LSM-510 META point scanning confocal microscope, and ten 2-micron-thick Z-stacks are collected at 4OX objective, 0.7 scan zoom, and 471 μm pinhole aperture size, approximately. In another embodiment, if a target area (e.g., a tumor or the like) is too thick to be imaged by the confocal microscope, then it is manually segmented. The processor 16 executes one or more sample reconstruction algorithms to reconstruct images of the z-stack samples. Additionally or alternatively, the processor executes a sample aggregation algorithm 22 that aggregates the z-stacks into a 3-D image volume. In one embodiment, the processor executes a ray casting or thresholding algorithm 24 on the z-stacks to segment out the microvessels. The processor executes a density calculation algorithm 26 to determine the density and/or surface area of the microvessels in the z-stacks. For instance, given the total volume of a z-stack, Vz, and the volume of the microvessels, Vsv, microvessel density can be calculated as the ratio of blood vessel volume over the volume of interest: Vsv/Vz- The surface area of the segmented microvessels, extracted directly from the segmentation, can be used as an additional quantitative measure.

The processor 16 executes one or more rendering algorithms 28 to present a 3D image volume representation to a user, as well as the calculated density and/or surface area information, on a display 30. By collecting tissue samples and generating images over time, growth or regression of the microvessels can be catalogued. This technique is applicable to any highly vascularized tissues, such as the lung, kidney, cancers, etc. In addition or alternatively to a confocal microscope, this technique can also be performed using a micro-CT scanner, preferably one with 1 -micron resolution, although any type of imager with suitable resolution may be employed in conjunction with the various embodiments described herein.

It will be appreciated that the quantification of microvasculature as described herein may be employed for tubular structures in general, and is not limited to sprouts or vasa vasorum. For instance, the above-described quantification technique can be used for quantification of fluorescently labeled angiogenic vessels or any other structures that exhibit tubular shapes. Moreover, once quantified, the microvasculature volume and/or surface area information can be used for therapy planning, as input for other calculations, functions, comparisons, (e.g., to determine or evaluate a physiological parameter such as blood flow rate or volume through the microvessels, etc.).

In another embodiment, microvessel migration and growth in tissues culture can be quantified by using time stamped z-stacks and digital subtraction. For example, if the there are two z-stacks, one acquired at time tl and the other at time t2, microvessel migration, growth, or reduction is quantified by subtracting the segmented micro vessels at tl from the segmented micro vessels at t2. Such information can be employed to evaluate the efficacy of a current therapy regimen, for future therapy planning, etc.

In another embodiment, 2D or 3D region-of-interest delineation functionality is provided, such as in cases where standard automatic segmentation techniques may not work well due to background scatter or other factors.

According to another example, descending aorta cross sections probed for smooth muscle actin and lectin are imaged on the confocal microscope 12. In one embodiment, the microscope 12 is a Zeiss LSM-510 META point scanning confocal microscope at 63x resolution. Z-stacks comprising, e.g., 15 slices with dimensions of 512x512 pixels at physical resolution of approximately 2.54 μm are acquired. The effective spacing between slices is, for instance, 0.5 μm. The processor 16 executes the sample aggregation algorithm 22 to align the slices into a 3-D volumetric image.

To increase the resolution of the volumetric data, the processor 16 executes a tri-linear interpolation algorithm 32, yielding a volume with an isotropic voxel size of approximately 0.254 μm. In this manner, approximately .25 μm resolution can be achieved at a cost that is much less than that associated with, for example, a 1 μm resolution micro-CT device.

To visualize blood vessels in the plaque, adventitia, and vessel wall, the reconstructed z-stacks can be manually segmented, e.g. using a user input device 34 (e.g., a mouse, stylus, keypad, etc.), by drawing two dimensional regions of interest representing co-localized stains in consecutive axial slices. Only blood vessels going all the way through the interpolated volumes, as detected by the co-localized probes, need be considered. The processor 16 executes a cubic B-Spline curve algorithm 36 to model the obtained contours. The processor additionally executes one or more additional sample aggregation algorithms 26 to stack the modeled contours to provide a 3D volumetric surface representation, e.g., as a closed triangulated mesh or the like.

In another embodiment, R-snakes or deformable surface models are used instead of or in addition to B-spline curves, for automatic detection of blood vessels (co- localized stains).

The foregoing techniques can be employed for detection and visualization of blood vessels in general and are not limited to angiogenic vasa vasorum, but rather can be applied to any fluorescently labeled vessels that cannot be imaged with conventional 3D structural imaging modalities.

Figures 2-9 show examples of graphical representations of micro vasculature, such as can be generated by the system 10 of Figure 1. Lectin appears as a yellow stain indicating blood vessel tissue, since lectin binds to the lumen of the vessel wall. Since lumen only occurs in vessels large enough to pass blood, staining and imaging lectin facilitates identifying microvascular structures.

FIGURE 2 is a screenshot 50 of a single z-stack showing segmented microvasculature 52 such as is generated via execution by the processor 16 of the various algorithms stored in the memory 18, and displayed on the display 30.

FIGURE 3 is a screenshot 54 of a plurality of aggregated z-stacks showing aggregated microvasculature 52 in a 3D representation. The volume and density of the micro vessels is calculated (e.g., using the system 10 of Fig. 1) via quantification of their surface in 3D. FIGURE 4 is a screenshot 60 of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in the adventitia. Microvessels 52 are highlighted by white circles.

FIGURE 5 is a screenshot 62 of detected and visualized microvessels 64 in the smooth muscle after processing by the processor 16 and execution of the computer- executable instructions stored in the memory 18 of the system 10.

FIGURE 6 is a screenshot 70 of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in a blood vessel wall. Microvessels 52 are highlighted by white circles.

FIGURE 7 is a screenshot 72 of detected and visualized microvessels 64 in the vessel wall after processing by the processor 16 and execution of the computer- executable instructions stored in the memory 18 of the system 10.

FIGURE 8 is a screenshot 80 of a slice from a z-stack generated using confocal microscopy, showing stains of co-localized lectin and smooth muscle actin in an atherosclerotic plaque. Microvessels 52 are highlighted by white circles. FIGURE 9 is a screenshot 82 of detected and visualized microvessels 64 in the plaque after processing by the processor 16 and execution of the computer- executable instructions stored in the memory 18 of the system 10. It will be appreciated that in some embodiments, more than one stain or probe may be employed. For instance, lectin may be used to identify microvascular lumen in a sample, and a second stain or probe can be administered to identify or mark another feature associated with the vasculature, such as a growth factor or the like. It will further be appreciated that the subject techniques can be employed for any microvasculature, such as for imaging of lung tissue, kidney tissue, other highly vascularized tissue, embryonic vascular development, ischemic tissue, etc. Additionally the described techniques may be employed to evaluate vascular regression, such as may occur during or after a therapy regimen or medication dosing regimen. The innovation has been described with reference to several embodiments.

Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. It is intended that the innovation be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

CLAIMSHaving thus described the preferred embodiments, the invention is now claimed to be:

1. A micro vasculature detection and visualization system (10), including: a processor (16) that receives image data corresponding to a plurality of image slices of a tissue sample, generates one or more z-stacks from the plurality of image slices, and executes a thresholding process on the one or more z-stacks to segment out micro vessels (64) from other tissue; and a memory (18) that stores computer-executable instructions for generating the one or more z-stacks, and for segmenting the one or more z-stacks to identify the microvessels (64).

2. The system according to claim 1 , further including a high-resolution volume imager such as a confocal microscope (12), or a micro-CT imaging device, that generates the image slices.

3. The system according to claim 2, wherein the micro-CT imaging device is a 1 -micron resolution CT imager.

4. The system according to claim 1, wherein the processor (16) calculates microvessel density in the tissue sample.

5. The system according to claim 4, wherein the processor (16) calculates the microvessel density in the tissue sample by dividing a total microvessel volume, Vsv, by a total volume of the one or more z-stacks, Vz.

6. The system according to claim 1, wherein the processor (16) calculates a surface area value for the microvessels (64).

7. The system according to claim 1, wherein the tissue sample is stained with a first contrast agent.

8. The system according to claim 7, wherein the first contrast agent, such as a lectin-based stain binds to the lumen of the microvessels (64).

9. The system according to claim 8, wherein the tissue sample is stained with a second contrast agent that binds to a tissue different from the lumen.

10. The system according to claim 1, further comprising: a display (30) on which a 3D volume representation of the tissue sample with quantified microvessels is displayed.

11. A method of detecting and quantifying microvessels in a tissue sample, including: generating image slices of the tissue sample; aggregating the image slices to generate one or more 3D z-stacks; segmenting the one or more z-stacks to identify the microvessels (64); and rendering a graphical representation of the microvessels (64) on a display (30).

12. The method according to claim 11, further including: applying at least one of a thresholding technique and a ray casting technique to the one or more z-stacks to segment the microvessels (64).

13. The method according to claim 11, further including: generating the image slices using a confocal microscope (12) or a micro-CT imaging device.

14. The method according to claim 11, further including: calculating a density of the microvessels (64) in the tissue sample by dividing a total microvessel volume, Vsv, by a total volume of the one or more z-stacks, Vz.

15. The method according to claim 11, further including: calculating an aggregate microvessel surface area value for the microvessels (64).

16. The method according to claim 11, wherein the tissue sample is stained with at least one of: a first stain that binds to the lumen of the micro vessels (64); and a second stain that binds to non-lumen tissue.

17. A computer-readable medium (18) having stored thereon software for controlling one or more computers to perform the method according to claim 11.

18. The method according to claim 11, further including: using quantified microvessel information as input for one or more of therapy planning, evaluating efficacy of a therapy regimen, and calculating a physiological parameter; wherein the quantified microvessel information includes one or more of microvessel volume and microvessel surface area.

19. An apparatus for detecting and quantifying microvessels in a lumen-stained tissue sample, including: means (12) for generating image slices of the tissue sample; means (16) for stacking the image slices into one or more z-stacks, and for segmenting the one or more z-stacks to identify microvessels; and means (30) for rendering a 3D representation of the microvessels (64); wherein the image slices are generated at sub-micron resolution.

20. A method of determining a volume or surface area of microvessels, comprising: aggregating tissue image slices to generate one or more z-stacks; segmenting the one or more z-stacks to identify microvessels; and comparing the segmented images to the tissue image slices to obtain the volume or surface area of the microvessels.