Information Technology Reference
In-Depth Information
reference phantom integrated into the table mat. Short- and long-term reproduc-
ibility errors for BMD measurements in the sagittal reformations amounted 2.09
and 7.70 %, respectively [
53
]. Baum et al. [
53
,
54
] used only MDCT examinations
with a scan delay of 70 s after intravenous contrast medium injection. Acu et al.
[
55
] pointed out that scan protocols with different scan delay times after intravenous
contrast medium injection signi
cantly change the apparent BMD values. They
reported increasing apparent BMD values with longer scan delay times. Thus, the
scan delay time after intravenous contrast medium injection has to be taken into
account for the derivation of standard QCT equivalent BMD values from routine
contrast-enhanced MDCT examinations. Lastly, Summers et al. [
56
] and Pickhardt
et al. [
57
] demonstrated the feasibility to calculate BMD from computed tomo-
graphic colonography scans.
5 Measurements of Bone Microstructure
T-scores and BMD values of patients with versus without osteoporotic fractures
overlap as outlined in the Background section [
11
,
12
]. Bone strength re
ects the
integration of BMD and bone quality including bone microstructure [
1
]. Therefore,
substantial research efforts have been undertaken to assess bone microstructure
by using high-resolution imaging techniques to improve fracture risk prediction
[
58
,
59
]. Trabeculae have a diameter between 50 and 200
fl
m and the cortical
thickness varies between 0.2 to 5 mm. Thus, the spatial resolution of the imaging
techniques used for bone microstructure analysis is critical.
High-resolution peripheral quantitative computed tomography (hr-pQCT) allows
for an isotropic spatial resolution of 82
ʼ
m
3
in-vivo with a relatively low effective
μ
dose of approximately 4
Sv for a scan at the distal radius or tibia [
60
]. However,
hr-pQCT systems are limited to peripheral sites and cannot be applied to the spine.
Magnetic resonance imaging (MRI) lacks ionizing radiation. Bone tissue has low
MR signal and consequently appears dark in most clinically accessible pulse
sequences. Bone marrow has a relatively high MR signal, i.e. has positive contrast,
depending on the fat content [fatty (yellow) or hematopoietic (red) bone marrow]
and the applied pulse sequence [
58
,
61
]. High-resolution MRI has been performed
mostly at the peripheral skeleton such as radius, tibia, and calcaneus due to their
easy accessibility. Voxel sizes up to 137
ʼ
m
3
were reported for high-
resolution MRI at the distal radius [
62
,
63
]. Due to higher
×
137
×
410
ʼ
field strength and
sequence development, in-vivo MR imaging at the proximal femur as important
clinical fracture site has become feasible [
61
,
64
]. In contrast to peripheral sites, the
proximal femur contains not only fatty bone marrow, but also hematopoietic bone
marrow. The visualization of the trabeculae in the proximal femur is partly
obscured by the dark, hematopoietic bone marrow. An even higher percentage of
hematopoietic bone marrow is found in the vertebral bodies resulting in insuf
cient
signal-to-noise ratios to obtain high-resolution MR images of the spine.
