Biomedical Engineering Reference
In-Depth Information
the tumor, relevant organs, and the skin surface, are identified.
The 3D model created by this step enables viewing of anatomi-
cal structures from different directions, which facilitates visual-
ization of the target, sensitive normal tissues, and the acoustic
window. After the 3D models of these anatomical structures are
established, ultrasound parameter values, including the speed
of sound, density, attenuation, and absorption, are assigned to
each tissue type. The optimization step then determines ultra-
sonic, geometric, and electromechanical treatment parameters.
Ultrasonic parameters, as defined for these ultrasound hyper-
thermia applicators, include the excitation frequency and lens
focal length, which are specified by the tumor depth, locations
of sensitive normal tissues, and the transducer clearance above
the acoustic window. Next, the geometric parameters specify
the orientation and the scan path of the ultrasound transducer.
These parameters, which are selected in an effort to avoid exces-
sive intervening tissue heating, are specifically determined
by the geometric description of overlapping pressure beams.
Similarly, the electromechanical parameters describe the limita-
tions of the scanning machinery when interfaced to a human
patient. The electromechanical parameters are restricted by the
coupling bolus or coupling water bath, the acoustic window, the
scan path, and the patient position. Thus, the geometric param-
eters, as defined here, primarily consider ultrasonic energy
deposition within the tumor and in normal tissues, whereas
the electromechanical parameters are essentially concerned
with obstacles that might otherwise impede the motion of the
mechanical scanning apparatus. The thermal modeling step
estimates the temperature distribution within the patient, and
then the results are combined with the patient anatomical model
(Figure 6.8). The initial applicator settings are thus determined
by these steps, and these settings are updated based on feedback
collected during the treatment. Temperatures are also moni-
tored and stored for retrospective comparison and evaluation.
200
180
160
140
120
100
80
(a)
10%
90%
60
40
20
0
100
80
60
40
20
0
20
40
60
80
100
200
180
160
140
120
100
80
(b)
10%
90%
60
40
20
0
100
80
60
40
20
0
20
40
60
80
100
Distance from scan center (mm)
FIGURE 6.7
Simulated heating patterns produced by four mechani-
cally scanned ultrasound transducers. (a) All four transducers are
excited throughout the scan, which causes intervening tissue heating.
(b) The transducer nearest the central axis is turned off, and the inter-
vening tissue heating is significantly reduced. (After E. G. Moros, R. B.
Roemer, and K. Hynynen,
Int. J. Hypertherm.
, 6, 2, 1990.)
6.3.3.2 Ultrasound Hyperthermia with
phased array applicators
Numerical modeling of ultrasound phased arrays involves many
of the same calculations described for mechanically scanned
hypthermia applicators. The main difference is that phased
arrays steer and focus the pressure field electronically, whereas
mechanically scanned fixed phase applicators are tilted and
translated. Electronic steering and focusing enables phased
array beamforming, which optimizes the phase and amplitude
settings of the individual phased array elements. Beamforming
calculations are often performed in conjunction with tempera-
ture simulations, and the outcome of these computations is also
integrated with other aspects of treatment planning.
By adjusting the phases and amplitudes of the driving signals
applied to each of the transducer elements, ultrasound phased
arrays generate single focus patterns, annular patterns, and
multiple focus patterns. For a single-focus beam pattern, the
optimal driving signal phases are computed with field conjuga-
tion (Ibbini and Cain 1989). This approach excites each element
with a sinusoidal signal and calculates the phase of the complex
pressures generated by the SURLAS superpose the contributions
from the individual array elements multiplied by a plane wave
approximation for the attenuation, and then temperatures are
computed with a bioheat transfer model. Contributions from
individual arrays are evaluated at 1 MHz, 3 MHz, and 5 MHz,
and combined power depositions are computed for simultane-
ous excitations at 1 MHz and 5 MHz. The simulation results
show that the penetration depth is controlled by adjusting the
ratio of powers applied to the parallel opposed linear arrays.
When several of these numerical computations are combined
with other related calculations, the resulting numerical models
define a treatment planning approach for hyperthermia with a
single mechanically scanned ultrasound transducer (Lele and
Goddard 1987). This approach defines several steps, and the first
three, namely visualization, optimization, and thermal model-
ing, are specifically for prospective patient treatment planning.
In the visualization step, a 3D patient anatomical model is con-
structed from 2D CT scans, and various tissue types, including
