TY - GEN
T1 - Age-related changes in vertebral morphometry by statistical shape analysis
AU - Whitmarsh, Tristan
AU - Del Río Barquero, Luis M.
AU - Di Gregorio, Silvana
AU - Sierra, Jorge Malouf
AU - Humbert, Ludovic
AU - Frangi, Alejandro F.
PY - 2012
Y1 - 2012
N2 - The morphological changes of the vertebrae associated with normal aging are still subject of debate, whereas this knowledge is important in detecting vertebral fractures and degenerative shape changes. The aim of this study is to present a method to statistically analyze the vertebral shape and determining the morphometric changes related to normal aging. The analysis is performed on the L2 lumbar vertebrae from a large dataset of Computed Tomography scans. The surface meshes of all vertebrae, with a groupwise vertex correspondence between them, are first acquired by an intensity based registration process onto a segmented reference. Principal component analysis then reduces the dimensionality to the main modes of variation which were subsequently analyzed by multiple linear regression to acquire the global shape variations with respect to the age of the subjects. In addition, the correlation with age of the deformation at each mesh vertex is analyzed, giving a significance map of the age related changes. This analysis shows several shape changes which are in agreement with previous studies while also giving a more detailed global shape analysis. Understanding the normal shape changes allows for a better diagnosis of vertebral fractures and spinal pathologies.
AB - The morphological changes of the vertebrae associated with normal aging are still subject of debate, whereas this knowledge is important in detecting vertebral fractures and degenerative shape changes. The aim of this study is to present a method to statistically analyze the vertebral shape and determining the morphometric changes related to normal aging. The analysis is performed on the L2 lumbar vertebrae from a large dataset of Computed Tomography scans. The surface meshes of all vertebrae, with a groupwise vertex correspondence between them, are first acquired by an intensity based registration process onto a segmented reference. Principal component analysis then reduces the dimensionality to the main modes of variation which were subsequently analyzed by multiple linear regression to acquire the global shape variations with respect to the age of the subjects. In addition, the correlation with age of the deformation at each mesh vertex is analyzed, giving a significance map of the age related changes. This analysis shows several shape changes which are in agreement with previous studies while also giving a more detailed global shape analysis. Understanding the normal shape changes allows for a better diagnosis of vertebral fractures and spinal pathologies.
UR - http://www.scopus.com/inward/record.url?scp=84871464671&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-33463-4_4
DO - 10.1007/978-3-642-33463-4_4
M3 - Conference contribution
AN - SCOPUS:84871464671
SN - 9783642334627
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 30
EP - 39
BT - Mesh Processing in Medical Image Analysis - MICCAI 2012 International Workshop, MeshMed 2012, Proceedings
T2 - MICCAI 2012 International Workshop on Mesh Processing in Medical Image Analysis, MeshMed 2012
Y2 - 1 October 2012 through 1 October 2012
ER -