Biomedical Engineering Reference
In-Depth Information
to the temperature in PD-weighted images. PD temperature sen-
sitivity goes inversely with sample temperature and is relatively
low in sensitivity with ranges around -0.30 ± 0.01%/ °C reported
from 37°C to 80°C (Johnson 1974) so that very high SNR is
required to limit uncertainty in the temperature estimates. To
reduce the impact of the temperature-dependent
T
1
, very long
repetition times are required, which means spatiotemporal reso-
lution is often sacrificed (Chen et al. 2006). When lower TR val-
ues are used, it may be difficult to isolate changes in PD from
other relaxation times (Gultekin et al. 2005). Because of these
limitations, and the uncertainty in measurement, like diffusion,
PD is rarely used for clinical MRTI.
The temperature dependence of relaxation parameters was
predicted early on in NMR theory (Bloembergen et al. 1947). The
temperature-dependence of
T
1
has been well documented since
early NMR studies related it to the correlation time and hence
diffusion (Bloembergen et al. 1948).
T
1
was one of the earliest
and most aggressively pursued parameters considered for MRTI
measurement (Lewa et al. 1980; Parker et al. 1983; Matsumoto
et al. 1992; Matsumoto et al. 1994; Hynynen et al. 2000). In tis-
sues,
T
1
depends primarily on dipolar interactions during the
temperature-dependent rotational and translational motion of
molecules. As with diffusion,
T
1
can be described by the follow-
ing first-order Arrhenius rate process where
into account the small nonlinear temperature dependence of
M
0
,
the change in
S
as a function of temperature can be described as
dS
dT
dS
dT
S
T
=α −
(3.7)
1
which takes into account both the decrease in
M
0
and the
increase in temperature or vice versa (α is the temperature
sensitivity). Note that
TT
1,
= α−
(
TT
)
(3.8)
1
ref
ref
where
ref
is the baseline for
T
1
and temperature,
T
. Therefore, by
taking Equations 3.7 and 3.9 and by neglecting the small
T
2
(or
T
2
* for gradient echo acquisitions) temperature dependence and
weighting in
T
1
-
W
images, the change in signal in
T
1
-
W
images
as a function of temperature can be described as
=− =−
−
SS
S
1
ref
TTT
(3.9)
ref
α−θ
−−θ
TR E
EE T
(1
cos)
+
1
ref
T
2
(1
)(1 os
)
ref
1,
ref
=
− +α−
.
For many of these techniques, the signal changes remain
approximately linear up to about 54°C, where irreversible
changes in tissue tend to lead to an altered temperature sensi-
tivity coefficient. While this may seem unsatisfactory for moni-
toring tissue temperature for safety, it may be a useful adjuvant
measurement for determining tissue damage.
The spin-spin relaxation time (
T
2
) has a similar relationship
with temperature as
T
1
, but with a different activation energy
and break points (Lewa et al. 1990). In observing the signal from
a
T
2
-
W
image with increasing temperature, there are large sig-
moidal decreases that can even remain during cooling of the tis-
sue (Graham et al. 1998). This is an important characteristic of
irreversible tissue damage that is useful to define the outcome of
a thermal treatment. Therefore, although
T
2
may not be a good
quantitative means to measure temperature changes, the irre-
versible changes with thermal damage means it can play a role
in treatment verification of high-temperature thermal ablations.
Magnetization transfer contrast (MTC) techniques use spec-
trally selective radiofrequency (RF) pulses to saturate protons
bound to macromolecules, which are normally not visible in
MR due to very short
T
2
or
T
2
* values. During the RF pulse,
the bound, saturated protons either enter the primary pool of
water protons or transfer the magnetization to the primary pool
(Wolff et al. 1989). These exchanges are temperature dependent.
For example, MTC signals were determined to be tissue depen-
dent, with some tissues showing relatively no change in signal,
such as in adipose and brain tissues. In tissues where a change in
signal was seen with temperature, effects on MT-W images were
nonlinear with either increasing signal (as in the muscle, heart,
prostate, liver) or decreasing signal (blood) (Young et al. 1994;
where
Ee
TR
/(
T
(
T
T
))
1,
ref
ref
ET
()
a
1
∝
−
(3.5)
TT
()
e
kT
1
and
E
a
(
T
1
) is the activation energy of the relaxation. This acti-
vation energy is tied to dipole-dipole interaction and therefore
is extremely sensitive to the microenvironment. High tempera-
tures that can effect changes in the tissue state, such as dena-
turation and coagulation, can cause changes in the activation
energy that affect temperature measurements (Peller et al.
2002).
The heavy tissue dependence of the
T
1
temperature depen-
dence has been extensively documented. The sensitivity has
been measured in several tissues with different values including
0.97%/°C in adipose tissue (Hynynen et al. 2000) and 1-2%/°C
in soft tissue (Lewa et al. 1980; Parker et al. 1983; Matsumoto et
al. 1992; Cline et al. 1994; Matsumoto et al. 1994). It is important
to note that it is one of the only techniques conducive to measur-
ing temperature in adipose tissue.
Quantitative mapping of
T
1
requires a long imaging time,
so oten,
T
1
temperature dependence is related to the change
in signal from
T
1
T1-weighted spin echo or gradient echo images.
Generally, the signal,
S,
can be described as
−
TR
TT
TR
TT
1
1cos
−
−θ
e
()
1
ST
()
∝
MT
()sin
θ
(3.6)
0
−
()
e
1
where θ is the flip angle. The signal decreases with increasing
temperature due to increasing
T
1
and decreasing
M
0
. By taking
