Biomedical Engineering Reference
In-Depth Information
(a)
(b)
70
60
50
40
30
20
10
0
30
80
20
60
10
40
0
20
-10
0
-20
-40
-30
-40
0
20
40
60
80 0
Time (
x
5 s)
20
40
60
80
(c)
(d)
FIGURE 3.7
Motion errors in PRF MRTI of liver. Respiratory motion can result in shifts of the background magnetic field, which decreases the
conspicuity of the edges of the temperature changes in the lesion (a). Corrections that estimate the background magnetic field outside the treatment
area as well as extrapolating into the treatment area can minimize the errors and often produce usable temperature images (b). Suppression of the
background errors outside the treatment area can aid in visualization of the heating (c), while corrections inside the treatment area can minimize
the large errors associated with large movements (d).
As mentioned before, macroscopic susceptibility changes
in tissue can also affect the accuracy of temperature estimates
with the CPD technique (De Poorter 1995). Susceptibility does
change with temperature, but when compared to the PRF as a
function of temperature, this effect is four to five times smaller
in magnitude. It has been approximated as being linear with
0.0026 ppm/°C for pure water and 0.0016 ppm/°C in muscle
(De Poorter 1995). It has been found though that the tem-
perature dependence of the susceptibility is tissue dependent
(Young et al. 1996). There is also a temperature dependence on
the orientation and geometry of the heating source (Peters et al.
1999). One possible method to correct for susceptibility effects
is to measure the PRF response to temperature in a chemical
specimen that is not sensitive to temperature changes. Lipid is
such a specimen. It is expected that lipid will experience sus-
ceptibility effects in the voxel. Therefore, using lipid, or any
covalently bonded proton, as an internal reference to correct for
susceptibility is a possible means to improve temperature mea-
surements (Kuroda 2005).
System magnetic field drift during the therapy is yet another
effect that must be considered when the PRF is used for MRTI.
Temporal drift of the local magnetic field due to increased
gradient duty cycles and eddy currents can change the PRF
(El-Sharkawy et al. 2006). These changes are most prominent on
older systems, or when high-speed imaging sequences are used.
Simple corrections are often possible by using a reference phase
in unheated nearby tissue or a phantom at a fixed temperature
(De Poorter 1994). More elaborate corrections are also possible
using multiple reference phantoms to estimate the linear shifts
across the image (De Poorter et al. 1995). A method that cal-
culates the apparent diffusion coefficient (ADC) and PRF has
also been proposed to correct for field drift (Das et al. 2005). As
with susceptibility, lipid or another internal covalently bonded
proton reference can be used to aid in correcting for field drifts
(Kuroda 2005).
An alternative to the indirect CPD method is a direct method
of measuring PRF shift via chemical shift imaging (CSI) tech-
niques (Kuroda 2005). CSI has several advantages over CPD
techniques. One is that the water proton resonant frequency
is measured separately from other resonances whereas CPD
measures the mixture of frequency components in each voxel,
resulting in a variable response and relieving the problem of
intravoxel lipid contamination seen in CPD. Also, the lipid
signal can be used as an internal reference to account for field
drifts, susceptibility, and motion since it is relatively insensitive
to temperature (Kuroda 2005). Therefore, CSI would be par-
ticularly useful in areas with high lipid content, such as bone
marrow, breast, and head and neck lesions (McDannold et al.
2001). In addition, it may be useful in patients with fatty-liver
for minimally-invasive treatments of liver lesions (Hussain
et al. 2009). It is also important to note that areas with little or
no lipid can also benefit from CSI-based temperature imaging
