Information Technology Reference
In-Depth Information
Fig. 4 Sagittal reformation
with a slice thickness of 3 mm
of a routine thoracic and
abdominal MDCT. Note the
osteoporotic vertebral fracture
of L2
fracture. These active appearance models can also be used in DXA-based VFA
images to detect osteoporotic vertebral fractures [
44
]. Baum et al. [
43
] developed a
prototype algorithm for automatic spine segmentation in routine thoracic and
abdominal MDCT images and used this algorithm to automatically detect osteo-
porotic vertebral fractures. The algorithm automatically localized and identi
ed the
vertebrae. Then, each vertebra was automatically segmented by using corre-
sponding vertebrae surface shape models that were adapted to the original images
(Fig.
5
). Finally, anterior, middle, and posterior height of each segmented vertebra
was automatically determined, and anterior-posterior-ratio and middle-posterior-
ratio were computed to diagnose osteoporotic vertebral fractures. The prototype
algorithm demonstrated a good performance for the automatic detection of pre-
valent and incident osteoporotic vertebral fractures cross-sectionally and longitu-
dinally, respectively.
4 BMD Measurements
As outlined in the Background section, measurements of BMD are performed to
diagnose osteoporosis. Subjects at high risk for osteoporotic fractures can be
identi
ed, so that pharmacological treatment can be initiated [
1
]. Furthermore,
BMD examinations are used to monitor
treatment
response [
34
]. Fractured
