Elasticity Microscopy Imaging

Investigators:
M. O'Donnell
S.Y. Emelianov
Graduate Students:
N.A. Cohn
SUPPORT:
Grant from National Institutes of Health & Whitaker Biomedical Engineering Research Graduate Fellowship


Abstract

An elasticity microscope images tissue stiffness at fine resolution. Possible applications include dermatology, ophthalmology, pathology and tissue engineering, especially if the resolution of these images can approach cellular dimensions. Elasticity images are reconstructed from displacements and strains measured throughout the specimen during controlled external loading. We use high frequency ultrasound to obtain these images by tracking coherent speckle motion during deformation. We also present methods that deal with speckle decorrelation caused by limited depth of field from a single element, focused transducer. Our 50 MHz elasticity microscope tests the methods by imaging axial displacement and normal axial strain fields in a tissue mimicking phantom. These results demonstrate the potential of elasticity microscopy. Ultimately, we would like to move to even higher frequencies, and improved resolution.


Experimental Results

This is the experimental setup. This is a B-Scan illustrating sample deformation, total image dimensions are less than 1 millimeter
An experiment on a tissue mimicking phantom is illustrated below. The phantom consists of gelatin and fine graphite particles used as acoustical backscatter. The thickness of the phantom is approximately 2mm undeformed. The phantom is deformed a total of 29 microns. The original and final deformed B-scan images are shown below. The transducer is located above the images.

Original Deformed
B-scan images, 40dB dynamic range, image dimensions less than 1mm

The displacement is calculated between stepwise deformations using a correlation search. The correlation magnitude function is interpolated to obtain values to sub-pixel accuracy. The axial values are further improved by utilizing the phase of the complex baseband signal. The images of axial and lateral displacement are then accumulated over all the stepwise deformations, taking the motion between frames into consideration. The accumulated displacement images are shown below, in the geometry of the original image.

Accumulated displacement images, image dimensions less than 1mm
Axial Displacement Lateral Displacement
black=no motion black=46 microns left
white=105 microns upward white=46 microns right

The normal axial strain is computed using the derivative of axial displacement. The elasticity distribution can be reconstructed from displacements, strains, and boundary conditions. These methods will be analyzed in future work. Once the elasticity distribution is obtained, information regarding internal tissue stiffness will be available at very high resolution. We anticipate many possible applications for an Elasticity Microscope.

For the present time, results are displayed in the form of normal axial strain images. Such images can be used to obtain a qualitative idea of internal stiffness distribution. Several examples are given below, both for controlled cases with phantoms containing regions of differing stiffness, and in a actual study involving tissue engineered smooth muscle. The axial resolution of normal axial strain images with the current system is better than 90 microns.


Circular cross-section of hard cylindrical inclusion (265 micron diameter measured optically before phantom creation) in soft homogeneous background (normal axial strain image and elasticity reconstruction)




Two-Layered Phantom (softer bottom layer)




Three-Layered Phantom (softer middle layer)


It is interesting to note a low strain region (black) in the middle of the soft layer (white band across the image). This corresponds very closely to what a hard inclusion looks like (see hard inclusion experiment above). Most likely, there exists a small hard particle near the soft middle layer in this phantom -- probably an undissolved gelatin crystal.


Hard Spherical Collagen Microcarrier Beads Embedded in Soft Gel (normal axial strain image and elasticity reconstruction)

Note the high strain regions above and below the hard collagen beads (depicted by the black low strain regions throughout the image); this represents positions where the soft gel is deformed against the beads.


Tissue Engineered Smooth Muscle
Histology Section Elasticity Micrograph
Smooth muscle layer is visible in both the histology (optical micrograph) and elasticity micrograph, to be a couple hundred microns thick at the surface. In tissue engineering, the development of the smooth muscle layer is limited by nutrient delivery, especially oxygen. The inner layer is comprised of lower density of cells and scar tissue, and is expected to be stiffer than the smooth muscle layer. The elasticity micrograph demonstrates significantly greater strain (softer region) in the top layer.


Histology Section Elasticity Micrograph
In this case, a polymer matrix was cultured in exactly the same manner except cells were not first seeded. There is no smooth muscle layer visible on the surface, either from histology or from elasticity micrograph. There are sparse cells visible in the histology section, due to some ingrowth from surrounding tissue.

Conclusion: Elasticity Microscopy can provide a noninvasive measure of internal tissue stiffness with resolution better than 90 microns. Such a system will be applicable in many different areas, especially as future systems use the methodology developed here to create a unit more specifically designed to work in a clinical environment, and one that will have even better resolution (by using a higher frequency single element ultrasound transducer in a similar system).


Possible Applications

Imaging tissue stiffness at high resolution can provide valuable information that may complement alternative modalities such as light microscopy and staining techniques, in addition to methods for bulk elasticity measurement using applied force. There are many applications that might benefit from imaging a high resolution elasticity distribution.

Some of these items are listed here.

Dermatology

Ophthalmology

Pathology

Tissue Engineering



Publications

  • Ultrasonic Imaging and Tissue Characterization Conference, Oral Presentation, June 5, 1996, Rosslyn, VA.
  • Whitaker Foundation Biomedical Engineering Research Conference, Poster, August 1996, Snowbird, UT.
  • Biomedical Engineering Symposium, Oral Presentation, September 1996, Ann Arbor, MI.
  • IEEE Ultrasonics, Ferroelectrics & Frequency Control Symposium, Oral Presentation, November, 1996, San Antonio, TX.
  • Ultrasonic Imaging and Tissue Characterization Conference, Oral Presentation, June 4, 1997, Rosslyn, VA.



    • Correspondence:

    Link to: http://bul.eecs.umich.edu/research/elast_micro/
    Last modified by: N.A. Cohn