26 - Landmark Based Morphometric Analysis of the Pulmonary Valve

Poster Board Number: 26
There are separate poster presentation times for odd and even posters.
Odd poster #s – first hour
Even poster #s – second hour

Co-authors:

Mitchell Katkic - Edward Via College of Osteopathic Medicine; Ann Rizkallah - Edward Via College of Osteopathic Medicine; Jonathan Millard - Discipline Chair - Anatomical Sciences, Biomedical Sciences, Edward Via College of Osteopathic Medicine

Abstract Body : The pulmonary semilunar valve is positioned in the cardiac outflow tract between the right ventricle and pulmonary trunk. Pulmonary valve dysfunction may be treated with open heart approaches or minimally invasive transcatheter techniques. Greater pressure gradients across prosthetic valves result in suboptimal cardiac performance, so prosthetic matching to the native valve is essential for success. Prosthetic valve selection generally centers outflow caliber, however, consideration of cusp dimensions may also affect outcomes. Semilunar valve cusps are a complex shape, consisting of a flat “sheet” of tissue molded into a three-dimensional cup attached to the interior wall of the outflow tract. Assessment of cusp morphology often includes volumetric measurement of its cavitation, although this simple volume does not adequately account for the shape of the structure. The primary aim of this research is to evaluate cusp shape using geometric morphometric techniques. Eighteen pulmonary valves were harvested from formalin-fixed body donors. Valves were opened between the anterior and left cusps and standardized photographs were taken of lunula attachments to the wall of the outflow tract. Images were imported into tpsDig2 (v.2.32) where 7 discrete landmarks along valve apices and 90 semi-landmarks along lunule attachment were collected on each image. Associated Cartesian X and Y coordinates were imported into tpsRelw (v.1.11) for semi-landmark sliding to minimize Procrustes distances between participating points. Boas-aligned coordinates were imported into MorphoJ (v.1.07) for further analysis. Generalized Procrustes superimposition was used to remove the effect of shape and size while centering the coordinates to a common centroid. Principal component analysis (PCA) resulted in 16 Eigenvalues, with principal component (PC)1 contributing to 54.1% of the shape variation and PC2 contributing to 17.0% of the shape variation. PC1 is characterized by concomitant deepening of the anterior and left cusp, with shallowing of the right cusp. Image recollection and analysis were performed to ensure accuracy and confirmed similar findings along PC1. PC2 scores had a weak positive correlation with centroid size r(16) = 0.289, p = 0.10. Early findings from this pilot data suggest no association between size and shape. Findings corroborated between the original and replicate studies suggest the landmarking protocol is robust. These preliminary results suggest relative cusp morphology could be predicted, which may inform prosthesis selection or engineering. Future directions should include increased sample size and reassessing cusps in an alternate order as a method of strengthening the current results.