Information Technology Reference
In-Depth Information
cross-sections could compensate for the apparent asymmetry of 3D anatomical
structures, caused by improper patient positioning or patient motion during image
acquisition, and that reformatting can be applied to different anatomical structures.
In the case of spine images, reformatted images can be used to improve the visu-
alization of the spinal canal and intervertebral foraminae. In order to avoid mea-
surement errors, Birchall et al. [
8
] and Adam and Askin [
2
] computed the rotation
of vertebrae from the position of landmarks that were manually placed in each
oblique axial cross-section, de
ned in sagittal and coronal MR cross-sections
through the superior and inferior vertebral endplates, or parallel to the endplates
through the centers of each vertebral body. Liljenqvist et al. [
45
] focused their study
on vertebral morphology related to pedicle screw placement for the treatment of
scoliosis. The pedicle width, length and angle were measured in manually deter-
mined oblique MR cross-sections that were orthogonal to vertebral bodies. In a
study of automated survey of MR spine images [
95
], it was reported that automated
reformation of 3D spine images along the true sagittal, coronal or axial vertebral
body axes may potentially facilitate image interpretation. Vrtovec et al. [
86
] gen-
erated curved cross-section from MR spine images by extracting the 3D spine curve
and axial vertebral rotation, and representing them as polynomial functions that
guided the reformation procedure. The same method for the extraction of the spine
curve and axial vertebral rotation was used by Neubert et al. [
49
] to initialize
statistical shape models for the purpose of segmentation and analysis of high-
resolution MR spine images.
Although 3D image reformation is often used for observing and analysing a
variety of anatomical structures and related pathologies, it can be concluded that 3D
spine images can be in general reformatted according to the following two
principles:
multi-planar reformation (MPR) is de
ned and performed in the image-based
coordinate system (Sect.
3.1
),
curved-planar reformation (CPR) is de
ned in the spine-based coordinate
system and performed in the image-based coordinate system (Sect.
3.2
).
ned in the corresponding
coordinate system, which is then represented as a plane or a curved surface in the
image-based coordinate system, where image intensities are sampled. By
In both MPR and CPR, the plane of reformation is de
flattening
the extracted cross-sections onto a plane, visualization of the spine in both the
image-based and the spine-based coordinate system is enabled (Sect.
3.3
).
fl
3.1 Multi-planar Reformation
Multi-planar reformation is the most straightforward 3D image reformation. The
volume of the 3D image is cut by a plane, and image intensities are sampled on that
plane. According to the orientation of the sampling plane in the image-based
