Information Technology Reference
In-Depth Information
coronal pro
les. In such cross-sections, however, morphometric measurements
based on Euclidean metrics are not possible. On the other hand, curved-planar
cross-sections that display the anatomy of all vertebrae along the whole length of
the spine without anatomical deformations can be obtained, but in this case the
spine curve is no longer always displayed as a straight line. As a result, the spinal
curvature of the normal spine in the sagittal plane is visible in sagittal orthogonal
and oblique curved-planar cross-sections (Fig.
26
d, f), while the spinal curvature of
the scoliotic spine in the coronal plane is visible in coronal orthogonal and oblique
curved-planar cross-sections (Fig.
29
d, f). However, a signi
cant advantage of such
visualization is that morphometric measurements based on Euclidean metrics can be
performed directly from these cross-sections. The axial orthogonal curved-planar
cross-sections are always orthogonal to the spine curve and aligned with the cor-
responding axial vertebral rotation, both in the case of the normal spine (Fig.
30
c)
and scoliotic spine (Fig.
31
c).
4 Conclusion
Techniques for visualization and quantitative evaluation of medical images are in
general valuable for the development of image-assisted diagnosis, planning of
surgical interventions and assessment of medical treatment outcomes. In the
field of
spine image analysis, visualization and quantitative evaluation of spinal curvature
and axial vertebral rotation is important not only for planning of orthopaedic sur-
gical procedures and analysis of surgical results, but also for diagnosing and
monitoring of the progression of spinal deformities. Computer-assisted visualiza-
tion and quantitative evaluation of 3D spine images therefore remain challenging
tasks in the
field of medical image analysis.
In this chapter, automated determination of the spine-based coordinate system
for an ef
cient cross-sectional visualization of 3D spine images was presented. The
introduction of the spine-based coordinate system allows to determine curved-
planar cross-sections that follow the spine curve and axial vertebral rotation along
the whole length of the spine. The main purpose of the described image reformation
technique is to reduce the structural complexity in favor of an improved feature
perception of the spine, and to provide clinically relevant quantitative analysis of
the 3D spinal anatomy. Displaying the whole length of the spine within a single 2D
image makes the inspection of images quicker and more precise, while the prob-
ability of overlooking certain important features of the spine is reduced. The spine
curve and rotation of vertebrae about the spine curve can be obtained automatically
and used to transform 3D spine images from the image-based to the spine-based
coordinate system. As the spine curve and axial vertebral rotation are inherent
properties of the spine and therefore not affected by rigid body transformations, the
generated curved-planar cross-sections are independent of the position of the patient
in the scanner and of the orientation of the image acquisition plane. When visu-
alizing and inspecting 3D images
