Collaborator: Dr. Derek R. Boughner (LHSC Division of Cardiology, Robarts Imaging Laboratories, and UWO Department of Medical Biophysics)
Background: Heart valves possess two or three cusps (a.k.a. leaflets) consisting of layered structures that enable the cusps to flex and extend as the valves open and close. The thickness and microstructural organization of the tissue layers determine the mechanical properties of the cusps and affect diffusion of oxygen through the cusps, which may influence the ability of the tissue to repair itself. Approximately 40% of the 300 thousand patients undergoing heart valve replacement each year receive bioprosthetic valves fabricated from porcine or bovine tissue because bioprosthetic valves are less likely than mechanical valves to produce blood clots. Tissue fixation processes are employed to prevent immune responses to bioprosthetic valves, but fixation also dehydrates the tissue and alters its mechanical properties. Those structural alterations are considered a principal cause of the limited 10- to 15-year durability of bioprosthetic valves.
Objectives: We are developing ultrasound micro-imaging techniques for application to the preparation of heart valve biomaterials in a manner analogous to the use of ultrasonic nondestructive evaluation in manufacturing industries. Ultrasound can be used to map the thickness of each tissue layer and to analyze the microstructural organization of the layers without requiring a specimen to be fixed and sliced as in conventional histological analysis. This technology is expected to support the development of improved preparation methods for bioprosthetic valves and provide provide a means to monitor tissue growth and remodeling during fabrication of tissue-engineered replacement valves.
Approach: Fresh cusp specimens are examined ultrasonically while submerged in a physiologic saline solution. A series of two-dimensional images are acquired in parallel planes and reconstructed into a three-dimensional image of the cusp. The tissue layers can be distinguished in the three-dimensional images, as illustrated by the porcine aortic valve image in the Image Gallery. A semi-automated image processing algorithm has been implemented to detect the boundaries of the tissue layers and enable measurements of layer thickness. The algorithm consists of speckle reduction by median filtering, contrast enhancement by morphological filtering, and layer segmentation using a deformable contour method. The definition of thickness can be ambiguous for objects with complex surface topology such as heart valve tissues, so we are comparing different methods of thickness measurement to identify the most appropriate definition for heart valve characterization. Another short-term objective is to develop a method to analyze fibre orientation within the tissue layers, which should be particularly valuable for monitoring collagen formation in engineered biomaterials to ensure the mechanical properties of natural heart valves are reproduced as accurately as possible.
Recent Journal Papers:
Q. Qiu, J. Dunmore-Buyze, D.R. Boughner, and J.C. Lacefield, "Evaluation of an algorithm for semiautomated segmentation of thin tissue layers in high-frequency ultrasound images," IEEE Trans. Ultrason. Ferroelect. Freq. Contr., vol. 53, pp. 324-334, 2006.
J.C. Lacefield, J. Weaver, J.R. Spence, J. Dunmore-Buyze, and D.R. Boughner, "Three-dimensional visualization and thickness estimation of aortic valve cusps using high-frequency ultrasound," Physiol. Meas., vol. 25, pp. 27-36, 2004.
