close

Вход

Забыли?

вход по аккаунту

?

Miniaturization of Microwave Ablation Antennas

код для вставкиСкачать
Miniaturization of Microwave Ablation Antennas
By
Hung Luyen
A dissertation submitted in partial fulfillment of
the requirements for the degree of
Doctor of Philosophy
(Electrical Engineering)
at the
UNIVERSITY OF WISCONSIN-MADISON
2017
Date of final oral examination: 05/05/2017
The dissertation is approved by the following members of the Final Oral Committee:
Susan C. Hagness, Professor, Electrical and Computer Engineering
Nader Behdad, Professor, Electrical and Computer Engineering
Hongrui Jiang, Professor, Electrical and Computer Engineering
Mikhail Kats, Assistant Professor, Electrical and Computer Engineering
Joshua E. Medow, Associate Professor, Neurological Surgery
ProQuest Number: 10282661
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
ProQuest 10282661
Published by ProQuest LLC (2017 ). Copyright of the Dissertation is held by the Author.
All rights reserved.
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
i
Acknowledgements
I would like to express my deepest gratitude to Profs. Susan C. Hagness and Nader
Behdad for their continued support and invaluable guidance throughout this work.
I am also grateful to Dr. Fuqiang Gao, Dr. Jacob D. Shea, and my fellow doctoral
students, Yahya Mohtashami and James Sawicki, for their fruitful collaboration, valuable
assistance, and helpful discussions.
My thanks also go to Robert Weyker in the Meat Science Laboratory at the University
of Wisconsin-Madison for providing us with bovine livers for the experiments reported in
this work.
Last but not least, I would like to thank my parents and my wife for their constant
support and encouragement during the completion of the dissertation.
ii
Abstract
Microwave ablation (MWA) is a promising minimally invasive technique for the treatment of various types of cancers as well as non-oncological diseases. In microwave ablation,
an interstitial antenna is typically used to deliver microwave energy to the diseased tissue
and heat it up to lethal temperature levels. The desired characteristics of the interstitial
antenna include a narrow diameter to minimize invasiveness of the treatment, a low input
reflection coefficient at the operating frequency, and most importantly, a compact and localized heating zone. Most interstitial MWA antennas are fed by coaxial cables and designed for
operation at either 915 MHz or 2.45 GHz. Coax-fed MWA antennas are commonly equipped
with coaxial baluns to achieve localized heating. However, the conventional implementation
of coaxial baluns increases the overall diameters of the antennas and therefore make them
more invasive. It is highly desirable to develop new antenna designs that provide localized
heating and good impedance matching without the added thickness of coaxial baluns. In this
work, I demonstrate the feasibility of using higher frequency microwaves for tissue ablation
and present several techniques for decreasing antenna diameters for MWA applications. Operating at higher frequencies enables interstitial antennas with shorter active lengths. This
can be combined with smaller-diameter antenna designs to create less invasive applicators or
allow integration of multiple radiating elements on a single applicator to have better control
and customization of the heating patterns.
Chapter 1 provides an overview and background of MWA technology. Dielectric proper-
iii
ties, dielectric heating, and heat-induced injuries of biological tissues and designs of interstitial antennas are discussed.
Chapter 2 covers an investigation of the feasibility of using higher frequency microwaves
for tissue ablation by comparing ablation performance at 10 GHz and 1.9 GHz. Simulational
and experimental studies are presented.
Chapters 3, 4, and 5 present three coax-fed MWA antenna designs that have smallerdiameter configurations than conventional coax-fed balun-equipped antennas. The design,
operating principle, simulations and ex vivo experiments of each antenna are presented and
discussed in each chapter.
Chapter 6 presents a non-coaxial-based balanced antenna for MWA. Unlike coax-fed
antenna designs, this antenna uses a balanced parallel-wire transmission line to feed a balanced dipole to achieve localized SAR patterns and good impedance matching over a wide
frequency range. The design procedure, operating principle, simulations and experimental
characterization of the antenna are included.
iv
Contents
1 Introduction and Background
1
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
1.2
Dielectric properties of biological tissues . . . . . . . . . . . . . . . . . . . .
3
1.2.1
Dependence of dielectric properties on frequency . . . . . . . . . . . .
3
1.2.2
Dependence of dielectric properties on temperature . . . . . . . . . .
5
1.3
Microwave heating of tissue . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
1.4
Heat-tissue interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
1.5
Interstitial antennas for microwave ablation . . . . . . . . . . . . . . . . . .
8
2 Microwave Abalation at 10.0 GHz Achieves Comparable Ablation Zones
to 1.9 GHz in Ex Vivo Bovine Liver
12
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
2.2
Antenna Design and Simulations . . . . . . . . . . . . . . . . . . . . . . . .
15
2.3
Ablation Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
21
2.4
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
3 A Balun-Free Helical Antenna for Minimally Invasive Microwave Ablation 33
3.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
3.2
Antenna Design and Simulations . . . . . . . . . . . . . . . . . . . . . . . .
36
3.2.1
39
Impedance matching using a quarter-wave transformer . . . . . . . .
v
3.2.2
Impedance matching using a transmission line implementation of a pi
network of reactive elements
3.2.3
. . . . . . . . . . . . . . . . . . . . . .
39
Simulation results of the impedance-matched balun-free helical antennas 40
3.3
Ablation Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
3.4
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
4 Reduced-Diameter Designs of Coax-Fed Microwave Ablation Antennas
Equipped with Baluns
47
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
4.2
Antenna Designs and Simulations . . . . . . . . . . . . . . . . . . . . . . . .
49
4.3
Ablation experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
4.4
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
5 A Minimally Invasive Coax-Fed Microwave Ablation Antenna with a Tapered Balun
59
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
60
5.2
Antenna Designs and Simulations . . . . . . . . . . . . . . . . . . . . . . . .
62
5.2.1
Simulation Methods and Assumptions . . . . . . . . . . . . . . . . .
62
5.2.2
Interstitial Antenna Design with a Single-Slot Tapered Balun . . . . .
64
5.2.3
Interstitial Antenna Design with a Double-Slot Tapered Balun . . . .
67
5.3
Ablation Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
5.4
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
74
6 A Non-Coaxial-Based Balanced Antenna for Microwave Ablation
76
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
77
6.2
Antenna Design and Simulations . . . . . . . . . . . . . . . . . . . . . . . .
78
6.3
Ablation experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
6.4
Comparison with a Coax-fed Antenna . . . . . . . . . . . . . . . . . . . . . .
90
vi
6.5
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
93
vii
List of Figures
1.1
Basic topology of coax-fed monopole, dipole and slot antennas. . . . . . . . .
1.2
Topology of the choke monopole, floating sleeve dipole, triaxial, cap-choke
9
and double slot antennas. Black represents metal and light grey represents
dielectric. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1
10
Topology of the floating sleeve dipole (FSD) antenna used in the ablation
experiments. (a) Side view. (b) Cross-sectional view. The outer diameters
of the various layers are as follows: inner conductor = 0.515 mm, dielectric
= 1.676 mm, outer conductor = 2.2 mm, Teflon insulator = 2.5 mm, floating
sleeve = 3.2 mm, Teflon coating = 3.5 mm. . . . . . . . . . . . . . . . . . . .
2.2
15
Simulated SAR patterns for the (a) 10.0 GHz and (b) 1.9 GHz FSD antennas
examined in this study. (c) Direct comparison of the -25 dB contours of the
10 GHz (grey curve) and 1.9 GHz antennas (black curve). . . . . . . . . . . .
2.3
18
Transient thermal simulations show the expanding heating zone of the 10 GHz
antenna at (a) 10 seconds, (b) 1 min., (c) 2 min., (d) 4 min., (e) 5 min., and
(f) 10 min. The black contour represents the 60◦ C boundary from the ex
vivo simulation (no metabolic heat generation or blood perfusion). The red
contour represents the 60◦ C boundary from the in vivo simulation (including
the effects of metabolic heat generation and blood perfusion). . . . . . . . .
21
viii
2.4
Transient thermal simulations show the expanding heating zone of the 1.9
GHz antenna at (a) 10 seconds, (b) 1 min., (c) 2 min., (d) 4 min., (e) 5 min.,
and (f) 10 min. The black contour represents the 60◦ C boundary from the ex
vivo simulation. The red contour represents the 60◦ C boundary from the in
vivo simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5
22
VSWR measurement results of (a) 10 GHz FSD antenna in 5-min.-ablation
experiment, (b) 10 GHz FSD antenna in 10-min.-ablation experiment, (c)
1.9 GHz antenna in 5-min-ablation experiment, and (d) 1.9 GHz antenna in
10-min.-ablation experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6
24
Positions of the four temperature probes. (a) Cross-sectional view (horizontal
plane) of the 10.0 GHz antenna. (b) Cross-sectional view (horizontal plane)
of the 1.9 GHz FSD antenna. (c) 3D view of the plastic guide used to position
the probes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7
25
Temperature recorded by four temperature probes during the ablation experiments. (a) 10 GHz antenna, 5 min. ablation, (b) 1.9 GHz antenna, 5 min.
ablation, (c) 10 GHz antenna, 10 min. ablation, and (d) 1.9 GHz antenna, 10
min. ablation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8
26
Average temperature values recorded at each channel during the first 5 minutes of all of the (a) 10 GHz and (b) 1.9 GHz ablation experiments. The bars
show standard deviations. When a temperature level recorded by any of the
sensors reached 120◦ C, that sensor was removed from the ablation zone to
avoid damaging it. Therefore, the data plotted for Ch. 1 is terminated before
5 minutes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.9
29
Photographs of the ablation zone obtained using 42 W of microwave power
at (a) 10.0 GHz for 5 min., (b) 1.9 GHz for 5 min., (c) 10.0 GHz for 10 min.,
and (d) 1.9 GHz for 10 min. . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
ix
3.1
(a) Topology of a balun-free helical antenna. Dark grey represents copper,
light grey represents Teflon, and white represents air. (b) Simulated |S11 |
of the antenna with the reference plane placed at the base of the helix; the
first resonant frequency, f1 = 0.9 GHz, and the second resonant frequency,
f2 = 1.9 GHz, are marked. Normalized SAR pattern of the antenna at (c) f1
and (d) f2 . The antenna is inserted 85 mm deep into the liver tissue. . . . .
3.2
37
Topology of the proposed balun-free helical antenna matched by (a) a quarterwave transformer and (b) a pi network composed of two parallel capacitors and
one series inductor. Dark grey represents copper, light grey represents Teflon,
and white represents air. Dotted green rectangles highlight the boundaries of
the matching sections. (c) Equivalent circuit model of the pi network. . . . .
3.3
38
Simulated |S11 | of the 1.9 GHz balun-free helical antenna using the quarterwavelength transformer (dotted line) and the pi matching network (solid line).
The antenna design parameters are as follows: n = 10 turns, hh = 20 mm,
Dh = 1.63 mm, g = 2 mm. The optimized dimensions for the quarterwavelength transformer are as follows: a1 = 0.25 mm, l = 37 mm. The
optimized dimensions for the pi matching network are as follows: l1 = 6 mm,
l2 = 22 mm, l = 18 mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4
41
Normalized SAR pattern of the balun-free helical antenna matched by (a)
the quarter-wavelength transformer and (b) the pi network. The antenna is
operating in liver tissue and the insertion depth is 85 mm. The boundaries of
the matching sections are highlighted by the dotted green rectangles. . . . .
3.5
Simulation and measurement of |S11 | for the balun-free helical antenna, matched
by the pi network, operating in bovine liver tissue. . . . . . . . . . . . . . . .
3.6
42
44
Photographs of ablation zones created in ex vivo bovine liver by applying 42
W to the input of the balun-free helical antenna for (a) 5 min., (b) 10 min. .
45
x
4.1
(a) Impedance-matched transition between a Teflon-filled and an air-filled
coaxial cable. Topology of (b) a modified choke dipole and (c) a modified
floating sleeve dipole implemented on the air-filled coax sections. Black represents metal, gray represents Teflon, and white represents air. . . . . . . . .
4.2
Dielectric properties of pork loin measured using a dielectric probe kit (Agilent
85070E) connected to a vector network analyzer (Agilent E8364A). . . . . .
4.3
50
53
Simulation and measurement results for the input reflection coefficient of the
(a) conventional and modified CD, and (b) conventional and modified FSD
antennas. Dimensions (in mm) of the conventional CD antenna: la = 6.5,
lb = 7.5, g = 1.0, lc = 9.5. Dimensions (in mm) of the modified CD antenna:
la = 6.3, lb = 7.0, g = 1.0, lc = 9.3. Dimensions (in mm) of the conventional
FSD antenna: la = 5.0, lb = 5.0, g = 1.0, ls = 7.0. Dimensions (in mm) of the
modified FSD antenna: la = 5.0, lb = 5.0, g = 1.0, ls = 8.0. . . . . . . . . . .
4.4
54
Simulated 60◦ C contours (black) and -10 dB normalized SAR contours (gray)
at 7 GHz of the (a) conventional CD, (b) modified CD, (c) conventional FSD,
and (d) modified FSD antennas. . . . . . . . . . . . . . . . . . . . . . . . . .
4.5
55
Photographs of ablation zones generated by the fabricated (a) conventional
CD, (b) modified CD, (c) conventional FSD, and (d) modified FSD antennas
using an input power of 30 W at 7 GHz for 5 minutes. The measured values
for the maximum long-axis diameters and maximum short-axis diameters of
the ablation zones are displayed in cm. . . . . . . . . . . . . . . . . . . . . .
5.1
57
Topology of the interstitial antenna design with a single-slot tapered balun
and two active segments. (a) Views in the x-z and y-z planes. (b) Drawing
of the tapered outer conductor and the active segment connecting to it when
they are unrolled and placed on a flat surface. . . . . . . . . . . . . . . . . .
65
xi
5.2
Simulation results for the antenna with the single-slot tapered balun and two
active segments shown in Fig. 5.1. (a) Input VSWR. (b) Normalized SAR
pattern in the x-z plane. (c) Normalized SAR pattern in the y-z plane. (d)
Thermal simulation results for 50◦ C contours in the x-z plane (red contour)
and y-z plane (black contour) after 5 minutes of ablation using 20 W input
power. r1 = 10.75 mm and r2 = 13.25 mm. Yellow rectangles represent
the position of the single-slot tapered balun. The dimensions for the active
segments and the tapered balun of the antenna are as follows: la = 7 mm,
lb = 6.2 mm, lt = 16 mm, w1 = w2 = w3 = 0.5 mm. . . . . . . . . . . . . . .
5.3
66
Topology of the interstitial antenna design with a double-slot tapered balun
and three active segments. (a) Views in the x-z and y-z planes. (b) Drawing
of the tapered outer conductor and the two active segments connecting to it
when they are unrolled and placed on a flat surface. (c) Photographs of the
fabricated outer conductor for the double-slot tapered balun. . . . . . . . . .
5.4
68
Results for the antenna with the double-slot tapered balun and three active
segments shown in Fig. 5.3. (a) Simulation and measurement results for the
input VSWR. (b) Simulated, normalized SAR pattern in the x-z plane. (c)
Simulated, normalized SAR pattern in the y-z plane. (d) Thermal simulation
results for 50◦ C contours in the x-z plane (red contour) and y-z plane (black
contour) after 5 minutes of ablation with 20 W input power. Yellow rectangles
represent the position of the double-slot tapered balun. The dimensions for
the active segments and the tapered balun of the antenna are as follows: la = 7
mm, lb = 8 mm, lt = 18 mm, w1 = w3 = 0.5 mm, w2 = 0.7 mm. . . . . . . .
70
xii
5.5
Photographs of ablation zones produced by the fabricated antenna with the
double-slot tapered balun by using (a), (e) 20 W for 5 min., (b), (f) 20 W for
10 min., (c), (g) 30 W for 5 min., (d), (h) 30 W for 10 min. Ablation zones in
the x-z plane are shown in (a)-(d). Ablation zones in the y-z plane are shown
in (e)-(h). Locations of blood vessels seen in the ablation zones are marked
in (b)-(d) and (g). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1
73
Topology of a balanced antenna comprised of a dipole fed by a two-wire
shielded balanced transmission line. (a) Cross-sectional view of the two-wire
line. (b) Side view of the feed line and the antenna. Black represents copper, light gray represents Teflon, and white represents air. Dimensions for the
shielded two-wire line are as follows: d1 = 0.72 mm, d2 = 2.04 mm, d3 = 2.50
mm, and s = 0.20 mm. The floating shield is embedded in a Teflon catheter
with an overall diameter of 3.0 mm. Dimensions for the dipole are as follows:
ld = 3.5 mm and α = 60◦ . The lateral tip-to-tip distance between the two
dipole arms, dt , is 5.1 mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2
79
Measurement results for the relative permittivity and effective conductivity
of pork loin. Each curve represents the average of nine measurements (three
measurement sites on each of three samples). . . . . . . . . . . . . . . . . . .
6.3
80
Simulation results for the balanced antenna for the situation where the twowire shielded transmission line feeding the antenna is differentially fed using a
lumped port in CST Microwave Studio. (a) Input VSWRs for various values of
the flare angle α. The VSWR values at 10 GHz are 1.67, 1.35, 1.25, 1.10, 1.08,
and 1.08 for the flare angle of 20◦ , 30◦ , 40◦ , 50◦ , 60◦ , and 70◦ , respectively.
(b)-(c) Normalized SAR patterns at 7 GHz, 10 GHz, and 13 GHz in the (b)
x-z plane and (c) y-z plane for the case where α = 60◦ . . . . . . . . . . . . .
82
xiii
6.4
Topology of an external transformer used to connect the two-wire transmission
line at the input of the antenna to the coaxial output of the power amplifier.
The transformer is implemented on the double-sided printed circuit board. For
the transformer, yellow represents copper trace on one side of the substrate,
and black represents copper trace on the other side. For the balanced line,
black represents copper, and light gray represents Teflon. . . . . . . . . . . .
6.5
Simulated reflection and transmission coefficients of the external transformer
shown in Fig. 6.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.6
83
84
Simulated and measured input VSWRs of the balanced antenna when fed
from the input of the coax to two-wire line transformer. The curves represent
the total input VSWRs as seen from the input of the transformer. . . . . . .
6.7
85
Normalized SAR patterns of the balanced antenna when the antenna is fed
with the coax to two-wire balanced line adapter shown in Fig. 6.4. (a)-(b)
SAR patterns at 9 GHz in the (a) x-z plane and (b) y-z plane. (c)-(d) SAR
patterns at 10 GHz in the (c) x-z plane and (d) y-z plane. (e)-(f) SAR patterns
at 11 GHz in the (e) x-z plane and (f) y-z plane. . . . . . . . . . . . . . . .
6.8
86
Photographs of ablation zones created in ex vivo pork loin in the (a)-(d) x-z
plane and (e)-(h) y-z plane by applying 18 W to the input of the balanced
antenna for 10 minutes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.9
88
Photographs of ablation zones created in ex vivo pork loin by applying 18 W
to the input of the triaxial antenna for 10 minutes. The photographs in (a),
(b), (c) and (d) are from four separate ablations. The white dashed ellipses
show the fitted boundaries of the ablation zones and are used for measuring
the dimensions of the ablation zones. . . . . . . . . . . . . . . . . . . . . . .
93
xiv
List of Tables
2.1
Quadratic Coefficients for the Temperature-Dependent Cole-Cole Parameters
in Equations (2.2)-(2.5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2
16
Dimensions of Antenna Segments Shown in Fig. 2.1 for the Two FSD Antenna
Prototypes Examined in this Study. . . . . . . . . . . . . . . . . . . . . . . .
17
2.3
Thermal Properties of Materials Used in Thermal Simulations . . . . . . . .
19
2.4
Ablation Zone Generated with 1.9 GHz and 10 GHz Antennas for 5-minute
Ablation Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5
Ablation Zone Generated with 1.9 GHz and 10 GHz Antennas for 10-minute
Ablation Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1
30
31
Dimensions of The Ablation Zones Produced by The Fabricated Antenna, Cut
in The x-z and y-z Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . .
74
6.1
Dimensions of Eight Ablation Zones Generated with The Balanced Antenna
90
6.2
Dimensions of The Ablation Zones Generated with The Triaxial Antenna . .
93
1
Chapter 1
Introduction and Background
2
1.1
Introduction
While surgical resection is considered the gold standard for the treatment of cancers in most
cases, only a small fraction of patients meets the criteria for surgery due to the incredible
invasiveness and high risk of complications associated with it [1]- [2]. With advances in
medical imaging technologies, minimally invasive and noninvasive thermal therapies have
become promising alternatives for cancer treatment. Compared to surgical resection, thermal
therapies offer advantages of less invasiveness, reduced cost, shorter hospitalization time and
faster recovery for patients [3]. The common issues of thermal therapies in the early days of
their development are associated with high recurrence rate and incomplete ablation of large
tumors (with diameters greater than 3 cm). However, recent advances in thermal ablation
techniques have improved the sizes of ablation zones (diameters of 5 cm) and efficiency of
the treatment [4]- [5]. With their promising features, thermal ablation therapies have been
increasingly adopted for the treatment of various types of tumors in liver, lung, kidney and
bone [6]- [17].
Thermal therapies refer to the use of heat or cold to destroy tumor. The heat-based
techniques utilize different types of energy to increase temperature of tissue to lethal levels,
including radiofrequency (RF) currents, microwave, laser and ultrasound. Among these
energy sources, RF currents and microwaves are the most widely used for ablation. Compared
to RF ablation (RFA), microwave ablation (MWA) offers several advantages, including being
less dependent on changes of tissue properties and producing more direct heating [18]- [19].
Microwaves can penetrate better in several tissues that present high impedances to RF
currents such as bone and lung tissues [20]. Additionally, deposition of microwave energy
is also less impacted by changes in physical properties of tissue during the heating process,
while RF current flow is severely hampered by charred or desiccated tissues [21]. Moreover,
microwave ablation has been shown to provide faster rates of temperature rise in both normal
and ablated tissues than RFA [22]. These higher direct heating rates help MWA overcome the
3
heat sink effect more easily in areas with high blood perfusion [18]. Due to these advantages,
MWA has a potential to perform better than RFA in various clinical scenarios and different
tissue types. Indeed, numerous studies compared MWA and RFA in controlled experiments
and reported larger ablation zones achieved with MWA in liver, kidney, lung and breast [23][27].
Microwave ablation can be performed as minimally invasive therapy where interstitial
antennas are deployed through biopsy needles to the ablation sites under ultrasound, computer tomography or magnetic resonance guidance. Once in position, the antennas radiate
electromagnetic (EM) waves to heat up the tumors in order to induce cell death. In a clinical
setup, one or up to three antennas can be used simultaneously depending on the sizes and
locations of the tumors. Interstitial antennas in commercially available MWA systems are
operated at either 915 MHz or 2.45 GHz as the Federal Communications Commission (FCC)
allocate these bands for industrial, scientific, and medical (ISM) use. In those commercial
systems, the antennas have diameters in the range of 12 gauge to 17 gauge, with the claimed
maximum operating powers varying from 40 W up to 150 W [28].
Microwave ablation technique relies on absorption of electromagnetic energy of biological
tissues to produce heat. It is very important to understand the radiation characteristics of
interstitial antennas, how electromagnetic waves interact with tissues and how heat induces
tissue injuries. In this chapter, dielectric properties of biological tissues, principle of microwave heating, interaction of tissues with heat, and designs of interstitial antennas will be
discussed.
1.2
1.2.1
Dielectric properties of biological tissues
Dependence of dielectric properties on frequency
Tissue consists of water, ions, dissolved molecules and insoluble matters. The dielectric properties of biological tissues reflect the interaction between tissues and electromagnetic waves
4
at the cellular and molecular levels. Different mechanisms that associate with dielectric
properties of tissues include ionic diffusion, interfacial polarization, and dipolar orientation.
These mechanisms respectively are the major principles under three main dispersion regions
in dielectric spectrum of tissue: α, β and γ dispersion [29]. In the low frequency region, the
α dispersion is associated with ionic diffusion at the site of cellular membrane. In the intermediate frequency region up to hundreds of kilohertz, the β dispersion associating with the
interfacial polarization of cellular membranes, proteins, and other organic macromolecules
takes the major role. At higher frequency in the microwave region (above a few hundreds
MHz), the γ dispersion due to dipolar polarization of water molecules is the dominant mechanism characterizing the dielectric properties of tissues. The complex permittivity for this
dispersion can be characterized by the following Cole-Cole equation:
ˆ(ω) = ∞ +
∆
1 + (jωτ )1−α
(1.1)
where ∞ is the permittivity at ωτ >> 1, ∆ = ∞ − s with s being the permittivity at
ωτ << 1, τ is the time constant, and α is the distribution parameter. The permittivity of
tissues over the wide spectrum can be modelled by multiple Cole-Cole dispersion terms in
addition to a conductivity term which accounts for the contribution of the electrolytes:
ˆ(ω) = ∞ +
X
n
σi
∆
+
1−α
n
1 + (jωτn )
jω0
(1.2)
The parameters of the 4-pole Cole-Cole dispersion for characterization of the dielectric properties of 17 types of tissues in the frequency range from 10 Hz to 20 GHz at the temperature
of about 37◦ C are reported in [30].
5
1.2.2
Dependence of dielectric properties on temperature
During thermal therapy, tissue temperatures are elevated causing changes in dielectric properties of tissues. A number of publications have reported measured dielectric properties of
several biological tissues as functions of temperatures, including brain (e.g. [31]), prostate
(e.g. [32]), and liver tissues (e.g. [33]- [38]). In [33], dielectric properties of bovine liver tissue
were characterized over a wide frequency range of 0.5-20 GHz in the temperature range of
20-60◦ C. Measured dielectric properties of bovine liver tissue at higher temperature than
60◦ C but in narrower frequency ranges are also reported in several other studies [34]- [38].
Specifically, these studies present measurement data at 915 MHz and 2.45 GHz, the two most
widely used frequencies for MWA. At 915 MHz, the relative permittivity decreases with a
rate of less than 0.2%/◦ C and the effective conductivity increases with a rate of less than
2%/◦ C as temperature increases up to 80-90◦ C [34]- [35]. At 2.45 GHz, both the relative
permittivity and effective conductivity decreases significantly when temperature increases
to over 60◦ C, with the most dramatic drops occurring between 90◦ C and 100◦ C [36]- [37].
At close to 100◦ C, the relative permittivity and effective conductivity are about 61% and
70% of their initial values at 20◦ C, respectively [37]. Additionally, the changes in dielectric properties of liver tissue at high temperature levels (e.g. > 60◦ C) were found to be
irreversible [37]- [38].
1.3
Microwave heating of tissue
The use of MW energy to heat up tissue is based on the mechanism of dielectric heating. Most
of biological tissues contain water (e.g. water content > 95% for vitreous humour, < 20% for
cortical bone, [30]) which is a polar molecule and contributes essentially to heat generation
when the tissues are exposed to EM fields. A polar molecule forms a permanent electric
dipole under natural condition due to lacking of center of symmetry in the molecule structure.
Occurrence of external electric field causes electric polarization inside dielectric materials,
6
resulting in rearrangement and intensification of these dipoles in a specific direction. Under
the applied alternating electric field, the polar molecule tends to continuously align with the
direction of the electric field. However, molecules cannot move freely due to intermolecular
binding force and fail to keep up with the rapid alternating electric fields, therefore exhibiting
a phase delay with respect to the applied electric fields. This phase lag between electric
polarization and electric fields is expressed in term of the imaginary part of permittivity of
the material and is accounted for power loss inside the material:
W0 = ω0 00 E 2 = 2πf 0 00 E 2 = σef f E 2
(1.3)
where the power loss per unit volume W0 [W/m3 ] is determined by the electric field strength
E [V /m], frequency f [Hz], and imaginary part of relative permittivity 00 of the material.
The absorbed energy is then converted to heat in term of molecule thermal vibration due to
collision and friction between molecules.
1.4
Heat-tissue interaction
The goal of MWA, or hyperthermic ablation techniques in general, is to elevate temperature
of diseased tissues to cytotoxic levels in order to induce cell death. Destruction of tumor
cells under applied heat results from direct injuries followed by a progression of indirect
injuries. These injuries manifest at different levels, from the tissue level to the subcellular
level. The relation between temperature and tissue damage has been thoroughly investigated
in literature. At relatively low hyperthermia temperature of 42-45◦ C, exposure for 30-60
minutes causes irreversible damage to tissue, mainly due to inactivation of vital enzymes [39][41]. The exposure time required for irreversible cellular damage decreases exponentially as
temperature is further increased to 60◦ C (i.e. 4-6 minutes for temperature of 50-52◦ C [42]).
7
At temperatures above 60◦ C, protein coagulation of cytosolic and mitochondrial enzymes and
nucleic-acid-histone complexes results in immediate cellular damage, inducing coagulation
necrosis of exposed cells over the course of several days [6], [42]. Temperature higher than
100◦ C leads to tissue water vaporization followed by tissue charring and carbonization and
smoke generation [41].
The direct injury mechanisms at the subcellular level include disruption in nucleic acid,
cell membrane, cytoskeleton and mitochondrial function [41], [43], [44]. Disruption in cell
membrane was initially considered the major cause of cell death. This was supported by
the observation that cell membrane fluidity and permeability changes as temperature increases [44]. The altered fluidity and permeability trigger intracellular metabolite accumulation and fluid influx that eventually lead to cell death [41]. However, it has been shown
that membrane damage is not directly correlated with applied heat, but rather an end result
of other subcellular processes [3], [41]. Instead, mitochondrial dysfunction may be a more
important factor in determining heat-induced injury. This is due to the fact that ultrastructure changes in mitochondria are well correlated with viability and metabolic functions
and can be observed within 15 minutes of heat injury [41], [44]. Other possible mechanisms
contributing to direct injury include damage to structural chromosomal proteins, release of
lysosomal enzymes, damage to RNA and inhibition of DNA replication [3], [41], [44].
Indirect injuries that cause progressive tissue damage have been observed in histological
examination after thermal application. Depending on ablation modalities, tumor biology
and microenvironments, progression of cell death can occur for hours to several days after
cessation of heat stimulant [45]- [47]. A well known cause for the progression of cell death is
apoptosis which is a regulated cell suicide program. Apoptosis may be induced as cells are
exposed to temperature levels in the range of 40◦ -45◦ C. Under the heat stress, apoptosis is
triggered through the mitochondrial pathway and the surface death receptor pathway [48][49]. The progression of apoptosis continues for many hours following the treatment [47].
Moreover, exposure to hyperthermia temperature also induces activation of Kupffer cell
8
and production of cytokines (e.g. IL1 and TNF-α) that cause cytotoxic effects on tumor
cells [50]- [51]. Other possible mechanisms for indirect injuries have been reported in several
studies, including vascular damage [44], increased lysosomal activity [52], and stimulation of
the immune response [53].
1.5
Interstitial antennas for microwave ablation
Interstitial antennas play a critical role in the performance of a MWA system. Generally,
there are three most important requirements for an interstitial antenna: a small diameter,
low input reflection coefficient, and localized heating pattern. Firstly, the antenna should
have small diameter in order to minimize its invasiveness during the treatment. Less invasive antennas potentially help decrease the risk of complications for patients treated with
MWA [54]. Secondly, the antenna should provide a good impedance match with the source
at its input in order to reduce its input reflection coefficient. An antenna with a lower input reflection coefficient not only deliver energy to tissue more effectively but also reduce
unwanted heating of the feed line. Moreover, the antenna should produce localized specific
absorption rate (SAR) pattern in order to focus energy deposition in diseased tissues while
lessening the problematic heating of healthy tissues along the antenna shaft.
Most interstitial antennas are constructed based on and fed by coaxial cables. Coaxbased interstitial antennas usually can be categorized in three groups: monopole, dipole,
and slot [55]. Fig. 1.1 shows the most basic configurations of these three antenna classes.
A monopole antenna can be simply created by extending the inner conductor and dielectric
layer out of the outer conductor. A dipole antenna consists of two arms and a gap between
these arms. One arm is created by an extension of the inner conductor, and the other arm
is formed by a section of the outer surface of the outer conductor that is inserted in tissue.
A slot antenna is realized by cutting a slot on the outer conductor while short-circuiting the
outer and inner conductors at the distal end of the cable.
9
Monopole
Dipole
Slot
Figure 1.1: Basic topology of coax-fed monopole, dipole and slot antennas.
Coaxial cables are unbalanced transmission lines and they allow unbalanced currents
to flow on the outer surface of the outer conductors. These unbalanced currents cause
two common issues that negatively impact the ablation performance of coax-fed interstitial
antennas. First, these currents cause elongation in the SAR pattern that may translate into
detrimental heating of healthy tissue along the feed cable. Moreover, the distribution of these
currents depends on the insertion depth of the antenna, resulting in changes in the input
reflection coefficient and SAR pattern as the insertion depth varies [56]. Therefore, a majority
of interstitial antennas are equipped with coaxial baluns in order to suppress these undesired
outer-surface currents. A coaxial balun is most commonly implemented by encompassing
the outer conductor of the feed line with a hollow circular conductor. This conductor and
the outer surface of the outer conductor of the main coaxial line form a new transmission
line. The length and termination (in the direction towards the source) of this line are chosen
so that the balun can provide a high impedance for the currents flowing on the outer surface
of the outer conductor of the main coaxial line. Examples of coax-fed antennas equipped
with coaxial baluns include the choke monopole [57], cap-choke antenna [58]- [59], coaxial
slot dipole [60], triaxial choked dipole [61], expanded tip wire antenna [62], floating sleeve
dipole [63], and sleeved slot antenna [64]. However, a few coax-fed interstitial antenna designs
that do not use coaxial baluns have also been reported, including the triaxial antenna [65]
and double-slot antennas [66]- [67]. In the triaxial antenna design, a biopsy needle, used
to introduce a monopole antenna to an ablation site, is also utilized to adjust the insertion
10
depth of the monopole antenna to minimize the input reflection coefficient. The triaxial
antenna, however, does not produce the same level of outer-surface current suppression as
many balun-equipped antennas do. In [67], the double-slot antenna uses the destructive
interference caused by the two slots to effectively suppress the undesired currents flowing on
the outer surface of the coaxial line. As a result, the double-slot antenna was demonstrated
to create quite localized SAR and heating patterns [67]. Fig. 1.2 illustrates the topology of
some of these prominent coax-fed antenna designs.
Choke monople
Floating sleeve dipole
Triaxial antenna
Cap-choke antenna
Double slot antenna
Figure 1.2: Topology of the choke monopole, floating sleeve dipole, triaxial, cap-choke and
double slot antennas. Black represents metal and light grey represents dielectric.
Another goal in designing interstitial antennas is to improve power handling capability
of the antennas while reducing unwanted tail heating along the antenna shafts. Due to
ohmic losses, the outer conductor of the feeding cable can become excessively hot and unintentionally raise the temperature of healthy tissues surrounding the shaft of the antenna
to dangerous levels [20]. Hence, in order to avoid damaging the healthy tissues along the
antenna shaft, the input power of the antenna must be limited to less than a certain level.
However, the limit on the input power may prevent the antenna from creating sufficiently
large ablation zones that are required in many clinical scenarios. A prominent solution for
this problem is to use dedicated cooling mechanisms to take heat away from the shaft of the
11
antenna. Incorporation of active cooling mechanisms has been shown to significantly reduce
the detrimental heating along the antenna shaft and allow for higher input power levels and
longer ablation duration, leading to an effective extension of ablation zone sizes [68], [69].
Cooling effects can be achieved by circulating low temperature saline solution, water, or
compressed air along the outer surface of the feeding cables of the interstitial antennas. It
has become so widely adopted that eight out of nine commercial MWA systems reported
in [28] employ cooled shaft antennas.
12
Chapter 2
Microwave Abalation at 10.0 GHz
Achieves Comparable Ablation Zones
to 1.9 GHz in Ex Vivo Bovine Liver
The full manuscript was published as:
H. Luyen, F. Gao, S. C. Hagness, and N. Behdad, “Microwave ablation at 10.0 GHz
achieves comparable ablation zones to 1.9 GHz in ex vivo bovine liver,” IEEE Trans. Biomed.
Eng., vol. 61, no. 6, pp. 1702-1710, 2014.
13
2.1
Introduction
Numerous studies have examined tissue ablation at microwave frequencies using different
types of interstitial antennas (e.g. [56]- [72]). In most of these studies, the antenna is fed
with a coaxial cable and the topology of the antenna and its feed structure are optimized to
achieve the desired heating pattern while choking the currents flowing on the outer surface of
the outer conductor of the feeding coaxial cable. Specific examples of these designs include
the cap-choke [58]- [59], floating sleeve dipole [63], triaxial [65], and choke [70] antennas.
While many different antenna types and designs have been examined for microwave ablation
(MWA), the vast majority of these studies share a common trait. Namely, they have examined MWA at frequencies below 2.5 GHz (e.g. [56]- [63], [65]- [59]). In fact, most of the studies
reported in this area to date have used frequencies around 915 MHz (e.g. [56], [59], [73], [74])
and 2.45 GHz (e.g. [58], [60], [65]- [72], [75]- [76]) for ablation purposes.
A careful review of the literature in this area reveals only a few studies that have examined
using higher frequencies for MWA applications. In [77], a 9.2 GHz endometrial ablation
system was developed for treatment of abnormal menstrual bleeding; the frequency was
chosen so that the total depth of heating matched the endometrial thickness [78]. In [79][80], a multi-functional antenna operating at 14.5 GHz was proposed for both tissue ablation
and characterization and examined for treatment of liver cancer. The frequency 14.5 GHz
was reported to be optimal for both high power ablation and low power characterization [79].
In [81], 18 GHz was identified as the optimum frequency in the 0.9-30 GHz range for MWA
of xenografted mice tumors with high tissue specificity (low collateral damage) and high
efficiency (low input power). These studies, however, are exceptions rather than the norm.
The widespread availability of high-power generators at 915 MHz and 2.45 GHz and
the FCC allocation of those bands for industrial, scientific, and medical (ISM) use are two
practical factors that have motivated the use of low frequencies in MWA. However, these are
not limiting factors, as higher-frequency sources as well as higher-frequency ISM bands are
14
readily available. In terms of technical considerations, the choice of low frequencies has been
driven by concerns that smaller penetration depths at higher electromagnetic frequencies
would preclude the creation of sufficiently large ablation volumes. In fact, factors other
than frequency may have a greater influence on the effective penetration depth of microwave
radiation from interstitial applicators [82]. However, assumptions about penetration-depth
constraints have not been extensively tested.
In this chapter, we examine the use of higher frequency microwaves for tissue ablation
and compare with lower frequency MWA performance. Two interstitial floating sleeve dipole
(FSD) antennas operating at 10.0 GHz and 1.9 GHz are used in ex vivo ablation experiments
conducted in fresh bovine livers. The antennas are based on the design reported in [63]. In
both the 10 GHz and 1.9 GHz experiments, the same ablation time (5 or 10 min.) and input
power level (42 W) are used. The dimensions of the ablation zones achieved at 10.0 GHz
are found to be comparable to those achieved at 1.9 GHz. Additionally, the heating rates
in the vicinity of the antenna are higher in the 10 GHz experiments than in the 1.9 GHz
experiments.
These experimental results suggest that the penetration depth of propagating electromagnetic waves in lossy biological tissues is not the most appropriate metric for determining
the suitability (or lack thereof) of a given frequency for tissue ablation. Moreover, these
results suggest that using a relatively low microwave frequency (such as 1.9 GHz) does not
offer any advantages over using higher frequencies (on the order of 10.0 GHz) in terms of
the ablation sizes that may be achieved. Using high-frequency microwaves, however, offers
a number of practical advantages. Specifically, at higher frequencies, smaller and potentially less intrusive antennas may be used for tissue ablation. Furthermore, smaller antennas
permit the design of compact multi-element arrays that may potentially generate heating
patterns which are not easily achieved with a single low-frequency interstitial antenna.
15
2.2
Antenna Design and Simulations
Fig. 2.1 shows the longitudinal side view and the cross-sectional view of the FSD antennas
used in these experiments. Each antenna is composed of a main feeding coaxial cable with
a dipole at the end. The dipole has arm lengths of ha and hb and a gap length of g. A
floating sleeve with a length of hs is used to choke the RF currents excited on the outer
surface of the outer conductor of the feeding coaxial cable. This helps prevent the heating of
tissue along the insertion path of the antennas and also makes the response of the antenna
independent of the insertion depth into the tissue [63]. The entire structure is embedded in
a Teflon catheter as shown in Fig. 2.1(a) and 2.1(b). Two sets of FSD antennas operating
at 1.9 GHz and 10.0 GHz are designed for ablation experiments in bovine liver.
hs
hb
g
Teflon coating
Floating sleeve
Teflon insulation
ha
Outer conductor
Dielectric
Inner conductor
(a)
(b)
Figure 2.1: Topology of the floating sleeve dipole (FSD) antenna used in the ablation experiments. (a) Side view. (b) Cross-sectional view. The outer diameters of the various layers
are as follows: inner conductor = 0.515 mm, dielectric = 1.676 mm, outer conductor = 2.2
mm, Teflon insulator = 2.5 mm, floating sleeve = 3.2 mm, Teflon coating = 3.5 mm.
For simulation purposes, the permittivity of bovine liver is modeled with the following
one-pole Cole-Cole dispersion model:
ˆ(ω) = ∞ +
σi
∆
+
1−α
1 + (jωτ )
jω0
(2.1)
The Cole-Cole parameters (∞ , ∆, τ and σi ) are determined from the following secondorder polynomials that characterize the temperature dependence over the range from room
16
temperature to 60◦ C [33]:
∞ (T ) = A1 T 2 + B1 T + C1
(2.2)
∆(T ) = A2 T 2 + B2 T + C2
(2.3)
τ (T ) = A3 T 2 + B3 T + C3
(2.4)
σi (T ) = A4 T 2 + B4 T + C4
(2.5)
The values of the quadratic coefficients are given in Table 2.1.
Table 2.1: Quadratic Coefficients for the Temperature-Dependent ColeCole Parameters in Equations (2.2)-(2.5)
n
An
∞
1
−0.0127
0.8610
∆
2
0.0115
−0.8933
58.3598
τ [ps]
3
−0.0014
−0.0640
13.3749
σi [mS/m]
4
0.185
Bn
0.349
Cn
−5.4119
569.673
The specific absorption rate (SAR), defined as the power dissipated per unit volume
[W/m3 ] normalized by the tissue mass density [kg/m3 ], provides a measure of the amount
of microwave energy absorbed in the tissue. The spatial variation in SAR in the region of
tissue surrounding the antenna is calculated for each of the two FSD antennas using fullwave computational electromagnetic simulations in CST Microwave Studio. The dimensions
(ha , hb , hs , and g) of each antenna are tuned to ensure that it is impedance matched and
provides a localized SAR pattern at the desired frequency of operation. The final optimized
dimensions of both antennas are provided in the caption of Fig. 2.1 and Table 2.2. For
additional details about the principles of operation of an FSD antenna, the reader is referred
to [63].
Figs. 2.2(a) and 2.2(b) show the normalized SAR in the vicinity of the 10.0 GHz and the
1.9 GHz FSD antennas, respectively. In both antennas, the SAR levels are reduced by at
17
Table 2.2: Dimensions of Antenna Segments Shown in Fig. 2.1 for the Two
FSD Antenna Prototypes Examined in this Study.
hs [mm]
hb [mm]
ha [mm]
g [mm]
1.9 GHz
23.5
10.4
10.4
2
10 GHz
5.3
4
4
1
least 20 dB at the edge of the floating sleeve compared to the peak SAR value. This indicates
that the floating sleeves effectively suppress the currents excited on the outer surface of the
outer conductor of the feeding coaxial cable as indicated in [63]. Fig. 2.2(c) shows the -25
dB contours of both the 1.9 GHz and 10 GHz antennas plotted in the same graph, along
with the outline of the catheter, to allow for direct comparison. We observe that the -25 dB
contour for the SAR pattern of the 1.9 GHz antenna encompasses a larger volume compared
to that of the 10.0 GHz antenna. This is indeed due to the larger penetration depth of
electromagnetic waves at 1.9 GHz compared to that at 10.0 GHz. However, as will be shown
later, this does not adversely affect the size of the ablation zones that may be achieved using
the 10 GHz FSD antenna.
We also examined the expected heating patterns using combined electromagnetic/thermal
simulations in CST Multiphysics Studio. The assumed thermal properties of copper, Teflon,
liver, and air are listed in Table 2.3. The thermal properties and mass density of liver
are taken from [83]. First, we simulated the ex vivo case where there is no metabolic heat
generation or blood perfusion. In these ex vivo simulations, the liver is assumed to be at room
temperature (20◦ C) before ablation. Subsequently, in a second round of thermal simulations,
we included the effects of metabolic heat generation and blood perfusion to simulate the in
vivo performance of MWA at the two frequencies. We assumed a heat generation rate of
10.41 W/kg and blood perfusion rate of 1000 ml/kg/min. For each simulation, the calculated
microwave power dissipation per unit volume within the bovine liver, scaled for an input
power of 42 W, is used as the external heat source.
In all of these simulations, the liver is modeled as a homogeneous tissue. Moreover,
18
y [mm]
10
−25 dB
8
−20 dB
6
−15 dB
−10 dB
4
−5 dB
2
0
−2
−4
−6
−8 (a)
−10
−35 −30 −25 −20 −15 −10 −5
z [mm]
40
-25 dB
-20 dB
-15 dB
-10 dB
-5 dB
20
y [mm]
0
0
−20
(b)
−40
−70 −60 −50 −40 −30 −20 −10
z [mm]
40
0
1. 9 GHz
y [mm]
20
10 GHz
0
−20
(c)
−40
−70 −60 −50 −40 −30 −20 −10
z [mm]
0
Figure 2.2: Simulated SAR patterns for the (a) 10.0 GHz and (b) 1.9 GHz FSD antennas
examined in this study. (c) Direct comparison of the -25 dB contours of the 10 GHz (grey
curve) and 1.9 GHz antennas (black curve).
19
changes in the physical properties of tissue that occur during ablation (e.g. temperaturedependent dielectric properties) and thermodynamic mechanisms such as the latent heat
of water vaporization are not taken into account. These will inevitably introduce some
discrepancies between the results obtained in these simulations and those obtained from the
ex vivo ablation experiments. Nonetheless, these simplified simulations do provide useful
information about the relative performance of the 1.9 GHz and 10 GHz systems in both ex
vivo and in vivo ablation scenarios.
Table 2.3: Thermal Properties of Materials Used in Thermal Simulations
Thermal
Specific heat
Heat
conductivity
capacity
diffusivity
[W/m.K]
[kJ/kg.K]
[m2 /s] ×10−7
Liver
0.488
3.37
1.38
Copper
401
0.39
1151
Teflon
0.2
1
0.91
Air
0.026
1.005
214.8
Fig. 2.3 shows the results of the transient thermal simulation for the 10.0 GHz FSD
antenna at different observation times after the start of the ablation procedure. Specifically,
the graph shows the boundary of the 60◦ C temperature contours of the ex vivo (black curves)
and the in vivo (red curves) simulation at different observation times. We chose 60◦ C, the
median temperature of the congestive zone [84], to define the ablation zone boundary for
the purpose of estimating the dimensions of the ablation zone. The temperature inside the
volume bounded by each contour is higher than 60◦ C. Fig. 2.4 shows the results of a similar
transient thermal simulation conducted for the 1.9 GHz FSD antenna. For each antenna,
the high temperature regions first appear around the slot and arms and then expand as time
increases. Upon comparing the results shown in Fig. 2.3(b) and Fig. 2.4(b) for both ex vivo
and in vivo cases, we see that for shorter ablation times (e.g. 1 min.), the ablation zone
provided by the 10 GHz antenna is larger than the one provided by the 1.9 GHz antenna.
20
This larger ablation zone may be explained by examining the SAR patterns shown in Fig.
2.2. As shown in Fig. 2.2(c), the energy provided to the 10.0 GHz antenna is deposited
in a smaller volume compared to the volume at 1.9 GHz. Therefore, the thermal energy
density in the vicinity of the 10 GHz antenna is higher, which results in a faster temperature
rise compared to the 1.9 GHz case. As the ablation time increases, however, the sizes of
the ablation zones in the ex vivo simulations of the two FSD antennas become comparable.
This is particularly evident from the 5-min.-ablation results shown in Figs. 2.3(e) and 2.4(e)
and the 10-min.-ablation results shown in Figs. 2.3(f) and 2.4(f). Thus, the large ablation
volume of the 1.9 GHz antenna may be primarily attributed to the large volume in which this
antenna deposits the input microwave energy (see Fig. 2.2(b)). For the 10 GHz antenna,
however, the smaller energy deposition volume does not appear to adversely impact the
size of the ablation zone, as experimentally confirmed in Section 2.3. This is due to the
thermal diffusion process that transfers heat from the direct-microwave-absorption areas in
the immediate vicinity of the 10.0 GHz FSD antenna to the surrounding areas. Therefore,
even though these distant regions do not absorb a significant part of the input microwave
energy (see Fig. 2.2(a)), they nonetheless get hot enough to cause ablation on time scales
that are comparable to (or even faster than) those associated with low-frequency MWA.
Comparison between the ex vivo and in vivo simulations of each antenna in Figs. 2.3 and
2.4 shows that the effect of blood perfusion has significant impact on the size of the achievable
ablation zones at both frequencies. For both antennas, the in vivo ablation zones reach their
steady-state dimensions after a short ablation time (e.g. 2 min.) and their largest short-axis
diameters (measured on the y axis) achieved at the end of 10 min. are approximately half
the size of those achieved in the ex vivo case. The cooling effects of the blood perfusion
result in the reduction of the ablation zone dimensions for both the 1.9 GHz and the 10 GHz
systems. However, the 10 GHz antenna, which delivers higher-density energy in the central
heating zone, creates a slightly larger ablation zone compared to the 1.9 GHz antenna as
shown in Figs. 2.3 and 2.4.
20
20
10
10
y [mm]
y [mm]
21
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
(b)
20
20
10
10
y [mm]
y [mm]
(a)
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
0 10
(d)
20
20
10
10
y [mm]
y [mm]
(c)
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
0 10
(e)
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
(f)
Figure 2.3: Transient thermal simulations show the expanding heating zone of the 10 GHz
antenna at (a) 10 seconds, (b) 1 min., (c) 2 min., (d) 4 min., (e) 5 min., and (f) 10 min. The
black contour represents the 60◦ C boundary from the ex vivo simulation (no metabolic heat
generation or blood perfusion). The red contour represents the 60◦ C boundary from the in
vivo simulation (including the effects of metabolic heat generation and blood perfusion).
2.3
Ablation Experiments
The 1.9 GHz and 10.0 GHz FSD antennas were fabricated out of 50 Ω UT-085C-LL semirigid coaxial cables from Micro-Coax. The floating sleeves were created from hollow copper
tubes (the physical dimensions are provided in Fig. 2.1 and Table 2.2). SMA connectors
20
20
10
10
y [mm]
y [mm]
22
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
(b)
20
20
10
10
y [mm]
y [mm]
(a)
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
20
20
10
10
0
−10
(e)
0 10
(d)
y [mm]
y [mm]
(c)
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
0 10
0
−10
−20
−60 −50 −40 −30 −20 −10
z [mm]
0 10
(f)
Figure 2.4: Transient thermal simulations show the expanding heating zone of the 1.9 GHz
antenna at (a) 10 seconds, (b) 1 min., (c) 2 min., (d) 4 min., (e) 5 min., and (f) 10 min. The
black contour represents the 60◦ C boundary from the ex vivo simulation. The red contour
represents the 60◦ C boundary from the in vivo simulation.
were attached to the feeding ends of the semi-rigid coaxial cables. The entire assembly was
wrapped in Teflon tape according to the topology shown in Fig. 2.1. The overall length of
the assembly, from the tip of the antenna to the end of the SMA connector, is 38 cm for
each antenna.
Prior to beginning the ablation experiments, we measured the dielectric properties of the
23
livers at room temperature over the frequency range of 500 MHz to 20 GHz using an Agilent
vector network analyzer (E8364A) and an Agilent dielectric probe kit (85070E). In all cases,
the dielectric properties were found to be similar to those used in our simulations, with the
maximum difference less than 15%. We also measured the input VSWR of each antenna at
different insertion depths into the liver using a vector network analyzer (Agilent E5071C).
In all cases, the responses of the antennas at their desired frequency of operation did not
change when the insertion depth was changed. This indicates that the floating sleeve baluns
used in these antennas effectively suppress the currents excited on the outer surfaces of the
outer conductors of the feeding coaxial cables. Moreover, we found the input VSWR to be
consistently lower than 1.6, indicating a good impedance match in all cases. Immediately
after each ablation experiment, we also measured the VSWR to determine whether the
impedance match degraded due to ablation-induced changes in tissue dielectric properties.
Fig. 2.5 shows the comparison between pre- and post-ablation VSWRs. While we do observe
slight degradation, the post-ablation VSWRs were consistently below 1.8 in all cases. This
demonstrates that a good impedance match was maintained for every antenna throughout
the ablation process, despite significant changes in the dielectric properties of the tissue.
A total of 16 high-power ablation experiments were conducted in ex vivo bovine livers
– four for each unique pairing of frequency (1.9 or 10.0 GHz) and ablation duration (5 or
10 min.). In all cases, the antennas were inserted at a depth of 13 cm into the bovine liver.
For the 1.9 GHz experiments, a signal generator (HP 8350B Sweep Oscillator) connected to
a high-power solid state amplifier (DMS 7066) was used as the source. The output of the
power amplifier was connected to the input of the FSD antenna using a flexible coaxial cable.
The input power level of the power amplifier was adjusted to achieve an output power level
of 42 W (at the input of the semi-rigid coaxial cable). For the 10.0 GHz experiments, the
same signal generator connected to a high-power traveling wave tube (TWT) amplifier (IFI
T186-40) was used as the source. The output of the TWT amplifier was connected to the
input of the antenna using a flexible coaxial cable. Similar to the previous case, the input
24
5
3
2
1
0
2
4
6
8
10
Frequency [GHz]
before ablation
after ablation
4
VSWR
4
VSWR
5
before ablation
after ablation
3
2
1
0
12
2
4
6
8
10
Frequency [GHz]
(a)
5
(b)
5
before ablation
after ablation
(a)
4
VSWR
VSWR
4
3
2
1
0
1
12
2
3
4
Frequency [GHz]
5
6
before ablation
after ablation
(b)
3
2
1
0
1
(c)
2
3
4
Frequency [GHz]
5
6
(d)
Figure 2.5: VSWR measurement results of (a) 10 GHz FSD antenna in 5-min.-ablation
experiment, (b) 10 GHz FSD antenna in 10-min.-ablation experiment, (c) 1.9 GHz antenna
in 5-min-ablation experiment, and (d) 1.9 GHz antenna in 10-min.-ablation experiment.
power level of the TWT was adjusted to achieve an output power level of 42 W (at the input
of the semi-rigid coaxial cable). During each ablation, the reflected power from the antenna
was monitored and no significant changes in reflected power was detected, which agrees with
the post-ablation VSWR measurements of each antenna.
During the ablation process, four fiber-optic temperature probes connected to a fluoroptic
thermometer (Luxtron 3100) were placed in the vicinity of the antennas to monitor the
temperature changes in the liver. The locations of the probes and their relative positions
with respect to the two antennas are shown in Figs. 2.6(a) and 2.6(b). Channels 1, 2, and
3 are placed along the radial direction at distances of 5, 10, and 15 mm from the antenna
(similar to [22]). They are placed at the z coordinate (referring to Figs. 2.2, 2.3, and 2.4)
that corresponds to the largest short-axis diameters (measured on the y axis) of both the
SAR contours and the 60◦ C contours of each antenna. Channel 4 is placed at a distance
25
(a)
5 mm
10 mm
15 mm
10 mm
Ch. 4
Ch. 1
10 mm
Ch. 2
Ch. 3
(b)
5 mm
10 mm
15 mm
10 mm
Ch. 1
Ch. 4
20 mm
Ch. 2
Ch. 3
(c)
Ch. 4
Ch. 1
Ch. 2
Ch. 3
Antenna
5 mm
10 mm
Plastic guide
Liver
Figure 2.6: Positions of the four temperature probes. (a) Cross-sectional view (horizontal
plane) of the 10.0 GHz antenna. (b) Cross-sectional view (horizontal plane) of the 1.9 GHz
FSD antenna. (c) 3D view of the plastic guide used to position the probes.
of 10 mm beyond the tip of the dipole. A 3D-printed plastic guide was placed on top of
the liver to hold the probes precisely at the desired positions. The plastic guide consists of
an array of holes through which the temperature probes are routed, and parallel ridges to
provide vertical support to the temperature probes as shown in Fig. 2.6(c). Each hole has a
diameter of 1.8 mm to allow an 18-gauge needle to pass through. The needles were used to
guide the temperature probes to the desired insertion depth from the top of the livers. Once
each of the probes was guided to the desired location, the needle was removed from the hole
leaving only the temperature probe inside the liver.
26
120
100
80
60
40
20
0
60
120
180
Time [s]
Ch. 1
Ch. 2
Ch. 3
Ch. 4
240
300
Temperature [°C]
Temperature [°C]
120
100
80
Ch. 1
Ch. 2
Ch. 3
Ch. 4
60
40
20
0
60
(a)
80
60
0
120
240
360
Time [s]
(c)
Ch. 1
Ch. 2
Ch. 3
Ch. 4
480
600
Temperature [°C]
Temperature [°C]
120
100
20
240
300
240
360
Time [s]
480
600
(b)
120
40
120
180
Time [s]
100
80
Ch. 1
Ch. 2
Ch. 3
Ch. 4
60
40
20 0
120
(d)
Figure 2.7: Temperature recorded by four temperature probes during the ablation experiments. (a) 10 GHz antenna, 5 min. ablation, (b) 1.9 GHz antenna, 5 min. ablation, (c) 10
GHz antenna, 10 min. ablation, and (d) 1.9 GHz antenna, 10 min. ablation.
Fig. 2.7 shows representative temperature measurements for the 5- and 10-min.-ablation
experiments at each frequency. Whenever a probe recorded a temperature level above 120◦ C
(the maximum rated temperature for the probes), that probe was removed from the liver
to protect it from being damaged. This was the case for Channel 1 temperature probes in
all 10 GHz ablation experiments conducted (see Figs. 2.7(a) and 2.7(c)). A comparison
of Channel 1 curves in Figs. 2.7(a) and 2.7(b) (or, similarly, in Figs. 2.7(c) and 2.7(d))
shows that the area immediately surrounding the antenna heats up much more rapidly for
the 10 GHz system than the 1.9 GHz system. Furthermore, in most of the 10 GHz ablation
experiments, the temperature readings of the more distant temperature probes (Channels 2,
3, and 4) remained flat for approximately the first 15-20 seconds and then rapidly increased
27
(e.g., compare Channels 2-4 of Fig. 2.7(a) and 2.7(b)). This is in contrast to the results
we obtained for the 1.9 GHz system where in every experiment we conducted, all probes
registered a rise in temperature as soon as the power was turned on. This observed delay
in heating at locations away from the 10 GHz antenna suggests that ablation at 10 GHz is
caused by intense localized heating in conjunction with indirect heating of the regions away
from the central heating zone. This is further supported by noting that the locations of
the Channel 2, 3, and 4 probes fall outside of the -25 dB contour lines of the SAR pattern
of the 10 GHz antenna (see Fig. 2.2(a)). This indicates that the amount of EM energy
deposited in these areas is not significant and hence, heating at these locations is not direct.
Possible mechanisms for this indirect heating include thermal diffusion and the movement of
the generated water vapor from the intensely heated region to the surrounding areas. Since
our simulations do not include the thermal effects of the generation and movement of steam
and these simulations do predict similar behaviors as the ones shown in Fig. 2.7, we believe
that the primary cause of this indirect heating is thermal diffusion effects.
It is also important to note a few exceptions to these general trends that were observed
during our ablation experiments. For example, in Fig. 2.7(c) that shows the temperature
levels for one of the 10-min ablation experiments conducted at 10 GHz, the temperature
levels registered by Ch. 4 start to rise very rapidly and do not show the aforementioned 1520 seconds delay which can be observed in Ch. 2-3 temperature levels. We explain this by
pointing out that the liver is not a homogeneous environment (e.g., small veins are present at
many locations within the liver). Specifically, in the case shown in Fig. 2.7(c), the locations
of Ch. 1 and 4 sensors were near a small vein which connected them together. This vein acts
as a conduit to transfer heat from the areas nearby the center of the antenna to locations close
to Ch. 4 and explains why the temperature levels registered by Ch. 4 rise almost as fast as
those registered by Ch. 1. Nonetheless, in the majority of the experiments, the temperature
levels Ch. 2-4 showed the aforementioned delay that justifies our hypothesis that ablation
in these locations is caused by heat diffusion. Irrespective of the heating mechanism (direct
28
heating with EM waves or thermal diffusion), however, the temperature levels at all four
locations reach the ablation threshold level (60◦ C) faster in the 10 GHz ablation system
than that in the 1.9 GHz one. Finally, comparison of Figs. 2.7(a) and 2.7(c) with Figs.
2.7(b) and 2.7(d) shows that the 10 GHz ablation system provides significant heating in the
areas beyond the tip of the antenna whereas the 1.9 GHz ablation system does not (e.g., see
Ch. 4 curves in all of these figures).
Aside from these exceptions, the results and the trends observed in these measurements
were found to be repeatable across experiments. This is demonstrated in Fig. 2.8. Figs.
2.8(a) and 2.8(b) show the average temperature values recorded at the four different observation locations for the 10 and 1.9 GHz ablation experiments, respectively. These results
include data obtained from both the 5- and 10-minute ablation experiments. Therefore, the
observation time for the average data is limited to 5 minutes. The bars show the standard
deviations. The relatively large standard deviations are primarily attributed to the small
total number of ablation experiments used to generate these results. Nonetheless, the same
trends demonstrated in Fig. 2.7 and discussed earlier in this section are observable in these
figures.
At the conclusion of each ablation experiment, the liver was cut along a plane through the
insertion path of the antenna and the ablation zone was visually examined. The measured
dimensions of each ablation zone, given by the maximum long- and short- axis diameters
of the boundaries of the congestion zone, are reported in Table 2.4 and Table 2.5. The
photographs of the ablation zones corresponding to the four experiments whose temperature
measurements are shown in Fig. 2.7 are shown in Fig. 2.9. Figs. 2.9(a) and 2.9(b) show the
ablation zones obtained after 5 minutes of ablation at respectively 10.0 GHz and 1.9 GHz.
Figs. 2.9(c) and 2.9(d) show the ablation zones obtained after 10 minutes of ablation at
respectively 10.0 GHz and 1.9 GHz. The results show that the dimensions of the ablation
zones obtained at 10.0 GHz are comparable to those obtained at 1.9 GHz. Additionally,
the photographs shown in Fig. 2.9 show that the ablation zones in the 10 GHz ablation
29
120
Temperature [ ° C]
100
80
60
60
120
180
Time [s]
(a)
240
Ch. 1
Ch. 2
Ch. 3
Ch. 4
300
60
120
180
Time [s]
(b)
240
300
40
20
0
120
Temperature [ ° C]
100
Ch. 1
Ch. 2
Ch. 3
Ch. 4
80
60
40
20
0
Figure 2.8: Average temperature values recorded at each channel during the first 5 minutes
of all of the (a) 10 GHz and (b) 1.9 GHz ablation experiments. The bars show standard
deviations. When a temperature level recorded by any of the sensors reached 120◦ C, that
sensor was removed from the ablation zone to avoid damaging it. Therefore, the data plotted
for Ch. 1 is terminated before 5 minutes.
experiments extend farther beyond the tips of the antennas, as confirmed by the temperature
measurements of Ch. 4 in Fig. 2.7. Moreover, the size of the charred region in the vicinity
of the 10 GHz antennas is slightly larger than that of the 1.9 GHz antenna. This agrees with
30
(b)
(a)
(c)
(d)
Figure 2.9: Photographs of the ablation zone obtained using 42 W of microwave power at
(a) 10.0 GHz for 5 min., (b) 1.9 GHz for 5 min., (c) 10.0 GHz for 10 min., and (d) 1.9 GHz
for 10 min.
Table 2.4: Ablation Zone Generated with 1.9 GHz and 10 GHz Antennas for
5-minute Ablation Times
Frequency
[GHz]
Size
Photograph
Maximum long-axis
Maximum short-axis
diameter [cm]
diameter [cm]
10
7.0
3.2
10
5.5
3.3
10
5.5
3.0
10
5.0
3.0
1.9
5.7
3.5
1.9
5.5
3.0
1.9
5.0
3.0
1.9
5.1
2.8
Fig. 2.9(a)
Fig. 2.9(b)
the higher temperatures measured at Ch. 1 in the 10 GHz experiments in Fig. 2.7.
31
Table 2.5: Ablation Zone Generated with 1.9 GHz and 10 GHz Antennas for
10-minute Ablation Times
Frequency
[GHz]
2.4
Size
Photograph
Maximum long-axis
Maximum short-axis
diameter [cm]
diameter [cm]
10
6.4
4.0
10
7.2
4.2
10
7.0
3.8
10
6.0
4.0
1.9
7.3
3.8
1.9
6.5
3.5
1.9
7.0
3.7
1.9
6.7
3.5
Fig. 2.9(c)
Fig. 2.9(d)
Conclusions
Our numerical and experimental studies presented in this chapter demonstrate the feasibility of using higher-frequency microwaves for tissue ablation. In particular, we demonstrated
that, despite an increase in frequency by a factor of five, the sizes of the ablation zones obtained in 10 GHz MWA experiments are comparable to those achieved at 1.9 GHz conducted
at the same power level and for the same duration. This is contrary to the widely accepted
argument that lower-frequency microwaves provide larger ablation sizes due to higher penetration depths of electromagnetic waves into the body and the lower losses in biological tissues
at such frequencies. Our experiments show that, as far as ablation size is concerned, MWA
using a relatively low frequency of 1.9 GHz does not have any significant advantages over
MWA using a relatively high frequency of 10.0 GHz. However, higher frequency microwaves
offer the advantage of smaller antenna lengths. Additionally, new antenna topologies are
currently under investigation that allow for eliminating coaxial baluns in high-frequency,
coaxial-cable-fed MWA antennas thereby reducing the antenna diameter in addition to the
32
antenna length. Therefore, MWA may potentially be performed using less invasive antennas
at higher frequencies. Furthermore, using higher microwave frequencies, it may be possible
to develop compact multi-element arrays that are capable of generating heating zones not
achievable from large single-element antennas used in conventional low frequency MWA.
Finally, in situations where a small ablation zone may be needed, using higher frequencies
allows for the desired ablation size to be created in a shorter period of time. Due to these
possible advantages, further investigation into using high-frequency microwaves for ablation
experiments and studying new antenna designs and antenna arrays that take advantage of
the unique properties offered at such high frequencies is indeed warranted.
33
Chapter 3
A Balun-Free Helical Antenna for
Minimally Invasive Microwave
Ablation
The full manuscript was published as:
H. Luyen, S. C. Hagness, and N. Behdad, “A balun-free helical antenna for minimally
invasive microwave ablation,” IEEE Trans. Antennas Propag., vol. 63, no. 3, pp. 959-965,
2015.
34
3.1
Introduction
Most interstitial antennas designed for MWA are implemented using coaxial cables equipped
with baluns to choke the electric currents excited on the outer conductors of their coaxial
feeds [18]. If not properly suppressed, these currents may cause unwanted heating of healthy
tissue along the shaft of the antenna and degrade the impedance match between the antenna
and the source. An interstitial antenna equipped with a properly designed balun exhibits
good impedance matching, produces a highly localized specific absorption rate (SAR) pattern, and provides stable performance with respect to changes in insertion depth [55]. A
coaxial balun is typically implemented by encompassing the feeding coaxial cable with a
hollow circular conductor which may be either electrically connected to the outer conductor
of the coaxial cable (e.g. cap-choked antenna [58], coaxial slot dipole [60], triaxial choked
dipole [61], expanded tip wire antenna [62]) or electrically isolated from it (e.g. floating sleeve
dipole [63], sleeved slot antenna [64]). The implementation of a balun, however, increases the
overall diameter of the antenna and therefore its invasiveness. Thus, new antenna designs
which provide localized heating patterns and achieve a good impedance match without the
use of baluns are highly desirable for minimally invasive MWA.
There are several existing solutions for reducing the invasiveness of coax-fed antennas.
In [66], a double-slot antenna that does not use a balun was evaluated for MWA. By introducing an additional slot, this antenna provides a better localized SAR pattern compared
to the single-slot coaxial antenna and its performance is independent of insertion depth.
In [70], [85], [86], biopsy needles used to introduce coax-fed antennas to tissue served as
adjustable chokes for the antennas. This practical implementation offers a less invasive solution than conventional choked antenna designs for improving impedance matching and
SAR localization. The triaxial antenna presented in [65] is another design that integrates
a biopsy needle into the antenna structure. However, as opposed to the adjustable choke
design of [70], [85]- [86], the biopsy needle in [65] is electrically isolated from the outer con-
35
ductor of the coaxial cable and is used to adjust the insertion depth of the monopole antenna.
The triaxial antenna, while providing excellent impedance matching, does not produce such
highly localized SAR patterns as coaxial balun antenna designs do.
In this chapter, we propose and demonstrate a novel balun-free coax-fed helical antenna
that offers the desired performance characteristics for MWA, particularly a localized SAR
pattern. The concept of the balun-free design is described in Section 3.2. While this concept
is applicable to monopole antennas of any arbitrary geometry, the helical shape is chosen here
to reduce the length of the antenna. The proposed design exploits the fact that at a frequency
at which the electrical length of the helix (or any monopole antenna of arbitrary shape) is
approximately half a wavelength (i.e., the second resonant frequency of the antenna), the
electric current at the feed point achieves a minimum while the voltage is maximized. The
minimum feed point current (and hence the high feed point impedance) creates a natural
choke point for the currents that tend to flow on the outer surface of the outer conductor of
the coaxial cable, hence eliminating the need to use a balun. The high feed-point impedance
can be matched to the impedance of the main coax line using a compact impedance matching
section such as a quarter-wavelength transformer or a pi network. Therefore, operating the
antenna at this second resonant mode does not create any practical problems in terms of
feed design.
In Section 3.3, we report experimental results for a prototype of a 1.9 GHz balun-free
helical antenna matched by a pi network. The pi network effectively matches the input
impedance of the antenna to that of the main coax line as confirmed by the measured input
reflection coefficient of the antenna embedded in liver tissue. Two ablation experiments were
conducted in ex vivo bovine liver using the fabricated prototype with an input power of 42 W
for durations of 5 and 10 minutes. The achieved ablation zones showed localized heating and
their dimensions were comparable to those produced by the 1.9 GHz floating sleeve dipole
(FSD) antenna reported in Chapter 2 ( [87]) under the same experimental conditions (input
power of 42 W, ablation time of 5 or 10 minutes). The results demonstrate the effectiveness
36
of the proposed balun-free helical antenna for minimally invasive MWA.
3.2
Antenna Design and Simulations
The topology and operating characteristics of a balun-free helical antenna is shown in Fig.
3.1. The antenna operating in ex vivo bovine liver tissue is simulated in CST Microwave
Studio. The assumed dielectric properties for liver are taken from the one-pole Cole-Cole
model reported in [33] for room temperature. Fig. 3.1(b) shows the simulated input reflection
coefficient of the antenna at the reference plane defined at the helix base. Figs. 3.1(c) and
3.1(d) show the normalized SAR patterns of the antenna at the first and second resonant
frequencies, respectively.
At the first resonant frequency, f1 , where the electrical length of the helix is approximately
a quarter of the wavelength, the balun-free antenna is matched to 50 Ω as can be seen from
Fig. 3.1(b). However, the SAR pattern at this frequency is not localized; specifically, Fig.
3.1(c) shows that this SAR pattern exhibits long tails along the shaft of the antenna toward
the input of the feeding coaxial line and the air-liver interface. As shown in Fig. 3.1(c), the
tails of the -10 dB, -15 dB and -20 dB contours of the SAR pattern extend to the boundary
of the liver at the insertion plane. This would cause unwanted heating of healthy tissue along
the antenna shaft and hence, the balun-free antenna is unusable at this frequency for MWA.
In contrast, the SAR pattern of the balun-free antenna is highly localized at the second
resonant frequency, f2 , where the electrical length of the helix is approximately half a wavelength. This compact SAR pattern is a direct result of the current minimum occurring at the
feed point (base) of the helix. This current minimum results in a high feed point impedance
and provides a natural choke point for the currents that tend to get excited on the outer
surface of the outer conductor of the coaxial cable. The high input impedance also causes
a large impedance mismatch as illustrated by the high value of |S11 | at f2 in Fig. 3.1(b).
However, this problem can be easily resolved by introducing an impedance matching section
37
hh
Dh
High Zin
at f2
−10
f2
−20
|S
Reference
plane
| [dB]
0
11
g
−30
0
f1
1
40
40
f1
20
5
f2
20
0
y [mm]
y [mm]
4
(b)
(a)
−5 dB
−10 dB
−15 dB
−20
−40
2
3
Frequency [GHz]
Liver-air interface
−80
−60
(c)
−5 dB
−10 dB
−15 dB
−20
− 20 dB
−40
−20
z [mm]
0
0
20
−40
− 20 dB
Liver-air interface
−80
−60
−40
(d) −20
z [mm]
0
20
(d)
Figure 3.1: (a) Topology of a balun-free helical antenna. Dark grey represents copper, light
grey represents Teflon, and white represents air. (b) Simulated |S11 | of the antenna with the
reference plane placed at the base of the helix; the first resonant frequency, f1 = 0.9 GHz,
and the second resonant frequency, f2 = 1.9 GHz, are marked. Normalized SAR pattern of
the antenna at (c) f1 and (d) f2 . The antenna is inserted 85 mm deep into the liver tissue.
between the antenna and the main coaxial line. Once the impedance matching circuit is implemented, the balun-free antenna can be used efficiently at the second resonant frequency
for MWA.
While the most widely used operating frequencies for MWA antennas are 915 MHz and
2.45 GHz, we chose to test the proposed balun-free antenna at 1.9 GHz due to the fact
that the available microwave power amplifier in our laboratory operates in a narrow band
from 1.8 GHz to 2.0 GHz. Therefore, the balun-free helical antenna investigated in this
study is designed to operate at f2 = 1.9 GHz in bovine liver. The dimensions of the main
coaxial feed line in the simulations are based on 50 Ω UT-085C-LL semi-rigid coaxial cable
38
from Micro Coax. The outer diameters of the various layers of this cable are as follows:
0.574 mm for the center conductor, 1.676 mm for the Teflon insulation, and 2.197 mm for
the outer conductor. The diameter, Dh , the total length, hh , and the number of turns, n,
of the helical antenna were optimized in CST Microwave Studio to yield a localized SAR
pattern at f2 = 1.9 GHz. The CST-computed input impedance at the base of the helix was
used to design the aforementioned matching section. We present two solutions to achieve
this impedance matching between the balun-free helical antenna and the main coaxial line.
The first solution is to use a simple quarter-wavelength transformer as shown in Fig. 3.2(a)
and the second one is to use a transmission line implementation of a pi network of reactive
elements as shown in Figs. 3.2(b) and 3.2(c).
g
Quarter-wave
transformer
hh
l
Dh
(a)
Pi network
g
l2
l
l1
hh
(b)
Dh
(c)
C2
L
C1
Helical
Antenna
Figure 3.2: Topology of the proposed balun-free helical antenna matched by (a) a quarterwave transformer and (b) a pi network composed of two parallel capacitors and one series
inductor. Dark grey represents copper, light grey represents Teflon, and white represents air.
Dotted green rectangles highlight the boundaries of the matching sections. (c) Equivalent
circuit model of the pi network.
39
3.2.1
Impedance matching using a quarter-wave transformer
A quarter-wavelength transformer with a characteristic impedance of Z01 =
√
Z0 Zin is used to
match the feed-point impedance of the helical antenna (Zin ) to the characteristic impedance
of the feeding coaxial line (Z0 = 50 Ω) as shown in Fig. 3.2(a). A section of 50 Ω coaxial
line was modified to achieve a quarter-wavelength transformer with a higher characteristic
q
impedance than 50 Ω. To increase the characteristic impedance of a coaxial cable, Z01 = C̄L̄ ,
we may decrease its capacitance per unit length and increase the inductance per unit length
of the line:
C̄ =
2π
b
µ
ln( )
, L̄ =
b
2π a1
ln( a1 )
(3.1)
where a1 and b are the outer diameter of the inner conductor and the inner diameter of
the outer conductor of the modified coaxial cable section, respectively and and µ are the
permittivity and permeability of the insulating dielectric. The capacitance per unit length of
the coaxial cable can be decreased by removing the Teflon insulation layer between the outer
and center conductors. Further reduction of C̄ is also possible by decreasing the diameter
of the center conductor. The desired diameter of the modified center conductor, a1 , may
be derived from the value of Z01 calculated previously. The diameter of the modified inner
conductor and the length of the quarter-wave transformer are optimized in CST Microwave
Studio to achieve good impedance matching between the antenna and the main coaxial line.
3.2.2
Impedance matching using a transmission line implementation of a pi network of reactive elements
The physical implementation and the equivalent circuit model of a pi network of reactive
elements are shown in Figs. 3.2(b) and 3.2(c), respectively. The capacitors C1 and C2 are
realized using short sections of a low impedance coaxial cable. These are implemented by
40
inserting a hollow copper tube in the region between the inner and outer conductors of the
main coaxial cable. The tube is electrically connected to the inner surface of the outer
conductor to form a new outer conductor with a reduced inner diameter, b2 . This increases
the capacitance per unit length of the coaxial line. Since the lengths of these transmission
line sections are short, they can be treated as effective parallel capacitors as shown in Fig.
3.2(c). The series inductor, L, is realized using a short section of the main coaxial cable
in which the Teflon insulator is removed to decrease the capacitance per unit length of the
line. Since this new section has a relatively short length and a high impedance, it acts as a
series inductor as shown in Fig. 3.2(c). The values of C1 , C2 and L are chosen to provide an
impedance match between the high-input-impedance antenna and the main feed line. The
length of each segment of the matching section is estimated assuming a short transmission
line approximation for the reactive elements as follows:
C1 = l1
2π1
µ
b
2π1
, C2 = l2 b2 , L = l ln( )
b2
2π a
ln( a )
ln( a )
(3.2)
where a and b are respectively the outer diameter of the center conductor and inner diameter
of the outer conductor of the main coaxial cable, b2 = 0.876 mm is the inner diameter of the
copper tube used to realize the highly capacitive segments, 1 is the permittivity of Teflon,
and µ is the permeability of air. The dimensions for the pi network section are then finetuned using CST Microwave Studio to yield a low input reflection coefficient of the antenna
at the desired frequency.
3.2.3
Simulation results of the impedance-matched balun-free helical antennas
Fig. 3.3 shows the simulated input reflection coefficients of the 1.9 GHz balun-free helical antenna matched by a quarter-wave transformer or by a pi network. An excellent
impedance match (S11 < −20 dB) is achieved in both cases at the operating frequency
41
0
|S 11 | [dB]
−10
−20
−30
−40
−50
0
1
Quarter−wave transformer
Pi network
2
3
4
Frequency [GHz]
5
Figure 3.3: Simulated |S11 | of the 1.9 GHz balun-free helical antenna using the quarterwavelength transformer (dotted line) and the pi matching network (solid line). The antenna
design parameters are as follows: n = 10 turns, hh = 20 mm, Dh = 1.63 mm, g = 2 mm.
The optimized dimensions for the quarter-wavelength transformer are as follows: a1 = 0.25
mm, l = 37 mm. The optimized dimensions for the pi matching network are as follows:
l1 = 6 mm, l2 = 22 mm, l = 18 mm.
of the antenna. The quarter-wave transformer provides a slightly better impedance match
and broader bandwidth than the pi network implementation. The dimensions for the helix,
quarter-wave transformer and pi network section are provided in the caption of Fig. 3.3.
Fig. 3.4 shows the simulated normalized SAR patterns of the balun-free helical antenna
using the two methods of impedance matching. In both cases, the antenna is inserted into
the liver at the insertion depth of 85 mm. The SAR levels are reduced by more than 20 dB
compared to the peak SAR value at the longitudinal distance of 60 mm from the tip of the
antenna. Overall, the SAR patterns of the impedance-matched antennas are similar to that
of the unmatched helical antenna shown in Fig. 3.1(d). The localization of the SAR pattern
of the impedance-matched helical antenna indicates that, even with the absence of a balun,
no current is excited on the outer surface of the outer conductor of the feeding cable.
42
30
y [mm]
20
10
0
−5 dB
−10 dB
−15 dB
−20 dB
−10
−20
−30
−80
−60
−40
−20
z [mm]
0
20
0
20
(a)
30
y [mm]
20
10
0
−5 dB
−10 dB
−15 dB
−20 dB
−10
−20
−30
−80
−60
−40
−20
z [mm]
(b)
Figure 3.4: Normalized SAR pattern of the balun-free helical antenna matched by (a) the
quarter-wavelength transformer and (b) the pi network. The antenna is operating in liver
tissue and the insertion depth is 85 mm. The boundaries of the matching sections are
highlighted by the dotted green rectangles.
3.3
Ablation Experiments
The fact that the quarter-wavelength transformer requires a narrower-diameter center conductor to operate properly makes the fabrication process more difficult as precise machining
techniques would need to be used to fabricate the structure shown in Fig. 3.2(a). The pi
network implementation, however, is easy to fabricate in a laboratory setting. Hence, we fabricated a prototype of the proposed 1.9 GHz helical antenna matched by the pi network using
43
a 50Ω UT-085C-LL semi-rigid coaxial cable. The antenna was placed in a Teflon catheter
with an outer diameter of 3.2 mm. The relatively large dimensions for the coaxial cable
and catheter were chosen simply to ease the fabrication process during the proof-of-concept
demonstration phase. The overall diameter of this antenna can be significantly reduced with
the proper choice of smaller coaxial cables and a correspondingly thinner catheter.
We tested the fabricated prototype by measuring the S11 of the antenna while it was
inserted in a 45:55 mixture of methanol and deionized water. At 1.9 GHz, the relative
permittivity (≈ 48.8) and effective conductivity (≈ 1.26 S/m) of this mixture match that of
liver. Initially, we fabricated the prototype with the same dimensions as those assumed in the
simulation. However, the measured operating frequency of the antenna was slightly shifted
from the desired frequency of 1.9 GHz to 2.05 GHz. We attribute this frequency shift to the
non-idealities that exists in the fabrication process such as slight deviation of the fabricated
dimensions, and air gap in the Teflon insulation layer of the highly capacitive segments of
the impedance matching network that may reduce the capacitance per unit length of these
sections. This frequency shift was eliminated in the second prototype wherein we slightly
increased the length of the highly capacitive segments. Specifically, l1 was increased from 6
mm to 7 mm and l2 was increased from 22 mm to 24 mm. The second prototype was used
to perform the ablation experiments in bovine livers.
We acquired fresh bovine livers from the University of Wisconsin Meat Science Laboratory
for use in the ablation experiments. Before performing the ablation, we verified that the
antenna is impedance matched when inserted in the liver. This was done by measuring the
input reflection coefficient of the antenna using a vector network analyzer (Agilent E5071C).
Fig. 3.5 shows the measured S11 of the fabricated prototype plotted in the same graph with
the simulated S11 of the helical antenna using pi matching network. The S11 measurement
results show excellent impedance matching at 1.9 GHz. Additionally, the measured S11
results were observed to be stable as the insertion depth of the antenna in the tissue was
changed. This indicates that the antenna is balanced and no noticeable current is excited
44
on the outer surface of the outer conductor of its feeding coaxial cable.
|S 11 | [dB]
0
−10
−20
Measurement
Simulation
−30
0
1
2
3
Frequency [GHz]
4
5
Figure 3.5: Simulation and measurement of |S11 | for the balun-free helical antenna, matched
by the pi network, operating in bovine liver tissue.
Two ablation experiments were conducted with the balun-free helical antenna for durations of 5 and 10 minutes in ex vivo bovine livers. In both experiments, a signal generator
(HP 8350B Sweep Oscillator) connected to a high-power solid state amplifier (DMS 7066)
was used to provide 42 W at 1.9 GHz at the input of the SMA connector of the antenna.
Upon the completion of each ablation experiment, the liver was cut along the insertion plane
of the antenna to reveal the ablation zone. Figs. 3.6(a) and 3.6(b) show the photographs
of the ablation zones resulting from the 5 and 10-minute experiments, respectively. Both
photographs show localized ablation zones with relatively short tails of heating along the
antenna shaft. The dimensions of the ablation zones, given by the maximum long-axis diameter and the maximum short-axis diameter of the boundary of the congestive region (pink
tissue in the photographs), are 5.5 cm × 3 cm for the 5-minute experiment and 7.5 cm × 4.3
cm for the 10-minute experiment. These dimensions are similar to those achieved using an
FSD antenna reported in Chapter 2 ( [87]). The ablation experiments reported in Chapter
2 were conducted under the same exact conditions as the ones reported here (42 W input
power at 1.9 GHz for 5 and 10 minutes). Specifically, the dimensions of the largest ablation
zones created by the FSD antenna in our recent study were 5.7 cm × 3.4 cm for a 5-minute
45
(a)
(b)
Figure 3.6: Photographs of ablation zones created in ex vivo bovine liver by applying 42 W
to the input of the balun-free helical antenna for (a) 5 min., (b) 10 min.
ablation and 7.3 cm × 3.8 cm for a 10-minute ablation. Furthermore, the shape of the ablation zones obtained using balanced FSD antennas are very similar to the ones obtained using
the proposed balun-free helical antenna (compare Figs. 2.9(b) and 2.9(d) with Figs. 3.6(a)
and 3.6(b)). Given the similar performance (in term of achievable ablation zone sizes and
shapes) of the balun-free helical antenna compared to the MWA antennas that use coaxial
baluns, the proposed antenna shows promise as an alternative to existing MWA antennas as
it offers a practical method of reducing the overall diameter of the interstitial antenna, and
hence the invasiveness of MWA therapy.
3.4
Conclusions
We presented the design of a balun-free helical antenna using two different types of coaxial
impedance matching sections. The balun is eliminated by operating a base-fed helix at the
second resonant frequency of the antenna where its feed point current is very small and its
input impedance is very high. This effectively chokes the currents on the outer surface of
46
the outer conductor of the feeding cable and acts as a natural balun. As a result, the SAR
pattern of the antenna is localized at the operating frequency. Despite the high feed-point
impedance, perfect impedance matching between the antenna and the feeding coaxial line
can be achieved by using either of the two impedance matching sections demonstrated in
this chapter.
We fabricated a prototype of the proposed helical antenna using a pi network impedance
matching section to test the ablation performance of the proposed antenna. The input reflection measurement of the fabricated antenna shows a |S11 | value of -30 dB at the working
frequency, demonstrating the effectiveness of the matching section. We also conducted ablation experiments in bovine liver using 42 W applied at the input SMA connector of the
antenna for 5 and 10 minutes. The antenna produced localized ablation zones comparable
to which can be achieved by using an FSD antenna with the same input power and duration
as reported in Chapter 2 ( [87]). Therefore, the proposed balun-free helical antenna offers a
practical solution to decrease the overall diameter of coax-fed interstitial antennas. This is
expected to reduce the invasiveness of microwave ablation as a potential treatment of cancer
without compromising the performance compared to other coax-fed MWA antennas that
use coaxial baluns. In Chapter 2 ( [87]), we demonstrated that using higher frequency microwaves (e.g. at 10 GHz) can provides ablation zones comparable to those produced using
lower frequency microwaves (e.g. at 1.9 GHz). However, using higher frequencies for MWA
offers an advantage of reducing the length of interstitial antennas. This can be combined
with using the balun-free antenna design to create even more compact devices for tissue
ablation.
47
Chapter 4
Reduced-Diameter Designs of
Coax-Fed Microwave Ablation
Antennas Equipped with Baluns
The full manuscript was published as:
H. Luyen, S. C. Hagness, and N. Behdad, “Reduced-diameter designs of coax-fed microwave ablation antennas equipped with baluns,” IEEE Antenn. Wirel. Pr., 2016 (in
press).
48
4.1
Introduction
Most interstitial microwave abaltion (MWA) antennas are coax-fed and utilize coaxial baluns
to choke the undesired currents excited on the outer surface of the outer conductors of
the feed lines [55]. By suppressing these currents, the antenna achieves localized specific
absorption rate (SAR) and heating patterns and its input impedance becomes independent of
the insertion depth in the tissue. The only drawback of a conventionally implemented coaxial
balun is that it increases the diameter, and therefore the invasiveness, of the antenna. Scaling
down the diameters of the antennas by using smaller-diameter coaxial cables increases the
ohmic losses of the cable and can cause significant unwanted heating of the feed cable along
the insertion path of the antenna. Moreover, smaller coaxial cables have lower average power
handling capabilities that hinder the antenna from producing large ablation zones. Therefore,
solutions that can provide a reduction in overall diameter of the MWA antenna without
significantly increasing cable loss and decreasing its average power handling capability are
highly desired.
In this chapter, we present a technique to reduce the overall diameter of coax-fed antennas
equipped with coaxial baluns. The technique involves introducing an impedance-matched
air-filled coax section at the end of the Teflon-filled coax feed. Due to the lower relative
permittivity of air, the air-filled coax has a smaller diameter than the Teflon-filled coax for
the same inner conductor diameter and characteristic impedance. The active portion of the
antenna and the coaxial balun are constructed out of this air-filled coax section. The overall
diameter of the air-filled coax section with the balun does not exceed that of the Teflonfilled coax feed. The average power handling capability of this antenna is constrained by the
maximum allowable temperature of the Teflon as the cable heats during operation, and thus
is not impacted by the presence of the air-filled coax section [88]. Additionally, the length
of the air-filled coax section is relatively short; thus the increase in loss due to the use of a
smaller-diameter coax section is minimal.
49
We built prototypes of a choke dipole (CD), based on the design in [57], and a floating
sleeve dipole (FSD), based on the design in [63], to demonstrate the efficacy of the proposed
design concept. Moreover, we conducted simulations and ablation experiments to compare
the performances of these antennas with those of conventional CD and FSD antennas. The
designs, simulations, experiment results, and comparison of these antennas will be discussed
in Section 4.2 and 4.3.
4.2
Antenna Designs and Simulations
Fig. 4.1(a) shows the transition between a 50-Ω Teflon-filled coaxial feed to a 50-Ω air-filled
coax. For a fixed, common inner conductor diameter and characteristic impedance, the ratio
of the inner diameter of the outer conductor of the air-filled coax (da ) to that of the Teflonq
q
r,air
1
da
= 0.69, where r,air and r,Teflon are
filled coax (dT ) is expressed by dT = r,Teflon = 2.1
the relative permittivities of air and Teflon, respectively. This translates to a smaller outer
diameter for the air-filled coax compared to that of the Teflon-filled coax. The coaxial balun
can be implemented on this air-filled coax section to avoid the increase in overall MWA
antenna diameter which would otherwise occur with adding the balun on the Teflon-filled
coax.
Figs. 4.1(b) and 4.1(c) show the topology of the modified choke dipole and floating
sleeve dipole used to demonstrate the design concept. Each antenna consists of a dipole
comprising two arms with lengths la and lb , a feed gap with a length g, and a coaxial balun.
One dipole arm (la ) is created by extending the inner conductor of the air-filled coax and
connecting it to a metal tip. The gap between the truncated outer conductor and the metal
tip is considered the feed point of the dipole. The other dipole arm (lb ) is defined by the
outer-surface section of the outer conductor of the air-filled coax between the feed point and
the distal end of the coaxial balun. The choke for the CD works as a quarter-wavelengthlong transmission line with a short circuit at the proximal end to present an open circuit
50
point with high impedance at the distal end. The floating sleeve for the FSD works as a
half-wavelength-long transmission line with an open circuit at the proximal end to present
another open circuit point at the distal end. The high impedance at the distal end of each
coaxial balun creates a choke point to suppress the currents flowing on the outer surface of
the air-filled coaxial line.
Z01 = 50 Ω
Z02 = 50 Ω
Air
Teflon
da
dT
(a)
g
lc
lb
la
(b)
ls
lb
g
la
(c)
Figure 4.1: (a) Impedance-matched transition between a Teflon-filled and an air-filled coaxial
cable. Topology of (b) a modified choke dipole and (c) a modified floating sleeve dipole
implemented on the air-filled coax sections. Black represents metal, gray represents Teflon,
and white represents air.
Each antenna is fed by a 50-Ω UT-085C semirigid coaxial cable. The outer diameters of
the inner conductor, Teflon insulation layer, and outer conductor are 0.511 mm, 1.676 mm,
and 2.197 mm, respectively. We used a brass tube with an inner and outer diameter of 1.05
mm and 1.50 mm to make the outer conductor for the air-filled coax section. Hollow tubes
51
with the same inner and outer diameters as the outer conductor of the Teflon-filled coax
were used to encompass the air-filled coax to construct the choke for the CD and the floating
sleeve for the FSD. Both CD and FSD antennas are embedded in a Teflon catheter with an
outer diameter of 2.5 mm.
The topology of the conventional CD and FSD are similar to the topology of the modified
versions except that the dipoles and coaxial baluns are implemented on the same Teflon-filled
coax used for their feed lines. Therefore, we use the same notations for the lengths of the
feed gap (g), dipole arms (la and lb ), and coaxial baluns (lc or ls ) for both the conventional
and modified versions of each antenna type. Prototypes of the conventional CD and FSD
antennas are constructed out of the same semirigid coaxial cable type used for the modified
versions. Copper tubes with an inner and outer diameter of 2.5 mm and 3.2 mm were used
to construct the choke for the conventional CD and floating sleeve for the conventional FSD.
Teflon catheters with an outer diameter of 3.5 mm were used to cover the two antennas of
the conventional designs. These materials and dimensions are the same as the ones used
to fabricate the original FSD antenna presented in [63]. Compared to the conventional
versions of the CD and FSD antennas, the diameters of the modified antennas in these
implementations are 1 mm (30%) smaller.
In Chapter 2 ( [87]), we demonstrated that doing ablation at a relatively high frequency
(e.g. 10 GHz) allows for achieving comparable ablation zones and faster direct heating
rates compared to doing ablation at a relatively low frequency (e.g. 1.9 GHz). Using higher
frequencies for MWA enables antennas with shorter active lengths and more spherical heating
patterns at the expense of lower average power handling capabilities due to more severe cable
heating. For this study, we chose an operating frequency of 7 GHz for the antennas in order
to balance between the advantage of shorter antenna lengths and the disadvantage of higher
ohmic losses.
We designed the four antennas in CST Microwave Studio to operate at 7 GHz in porcine
muscle tissue. Fig. 4.2 shows measured data for the relative permittivity (r ) and effective
52
conductivity (σeff ) of pork loin at room temperature (20◦ C) in the frequency range from
0.5 GHz to 10 GHz. These dielectric properties were incorporated into the electromagnetic
(EM) simulations to model the tissue. At 7 GHz, the relative permittivity and effective
conductivity of the tissue are 41.9 and 7.2 S/m, respectively. For designing the modified CD
and FSD antennas, we first simulated the coaxial transition to verify that the air-filled coax
section was impedance-matched to the Teflon-filled coax section. The transmission coefficient
|S21 | was greater than −0.1 dB over the frequency range from 0 to 10 GHz, confirming the
good impedance match. Subsequently, we conducted EM simulations of the modified CD
and FSD antennas based on this coaxial transition. Similar simulations were conducted for
the conventional CD and FSD antennas. We optimized the dimensions of each antenna to
achieve a good impedance match and localized SAR pattern at the operating frequency.
These dimensions include the two dipole arm lengths, la and lb , and the gap length, g,
for each antenna, the choke length, lc , for each version of the CD and the floating sleeve
length, ls , for each version of the FSD. The optimized values for these dimensions of the
four antennas are shown in the caption of Fig. 4.3. Finally, transient thermal simulations
were conducted in CST Multiphysics Studio to predict temperature rises in the tissues for
a 5-minute ablation using an input power of 30 W for each antenna. Thermal conductivity
(0.566 Wm−1 K−1 ) and thermal diffusivity (0.168 mm2 s−1 ) of porcine muscle reported in [89]
were used in the thermal simulations.
Figs. 4.3(a) and 4.3(b) show the simulated |S11 | of the CD and FSD antennas, respectively. All antennas provide a good impedance match (|S11 | < −10 dB) to their feed lines at
the desired operating frequency of 7 GHz with |S11 | better than -20 dB. Fig. 4.4 shows the
simulated -10 dB normalized SAR contours (gray) of the four antennas. Also, the simulated
60◦ C contours (black), used to predict boundaries of ablation zones, are presented for the
four antennas. The shapes of the SAR contours and predicted ablation zones of the modified
CD and FSD antennas are similar to those of their conventional counterparts. SAR levels
are reduced by 10 dB near the distal end (toward the source) of the chokes along the CD
60
30
50
25
40
20
30
15
20
10
10
0
0
2
4
6
Frequency [GHz]
8
σeff [S/m]
εr
53
5
0
10
Figure 4.2: Dielectric properties of pork loin measured using a dielectric probe kit (Agilent
85070E) connected to a vector network analyzer (Agilent E8364A).
antennas and near the proximal end (toward the source) of the floating sleeves along the
FSD antennas. The simulated ablation zones produced by all antennas are compact and
have comparable sizes. While the maximum longitudinal dimensions (including the tails) of
the four ablation zones range from 3.5 cm to 4.0 cm, the maximum lateral dimensions are
the same and equal to 2.7 cm. Overall, all four antennas exhibit localized SAR and heating
patterns, indicating the effectiveness of the coaxial baluns in suppressing the currents flowing
on the outer surfaces of the feed lines.
4.3
Ablation experiments
We fabricated one prototype of each antenna to verify their performance in ex vivo ablation
experiments. Prior to conducting the ablations, we inserted each antenna into a porcine
muscle sample (i.e. pork loin) and measured its input reflection coefficient using a vector
network analyzer (Agilent E5071C). The |S11 | measurement results are plotted in Fig. 4.3.
The measured |S11 | for each antenna agrees quite well with the simulation results and shows
a good impedance match predicted by the simulations. At the design frequency of 7 GHz,
the measured |S11 | value was −24 dB for the conventional CD, −23 dB for the modified CD,
−27 dB for the conventional FSD, and −18 dB for the modified FSD. We also varied the
54
0
(a)
|S 11 | [dB]
−10
−20
−30
−40
0
Conventional CD: Simulation
Conventional CD: Measurement
Modified CD: Simulation
Modified CD: Measurement
2
4
Frequency [GHz]
6
0
8
(b)
|S 11 | [dB]
−10
−20
−30
−40
0
Conventional FSD: Simulation
Conventional FSD: Measurement
Modified FSD: Simulation
Modified FSD: Measurement
2
4
Frequency [GHz]
6
8
Figure 4.3: Simulation and measurement results for the input reflection coefficient of the
(a) conventional and modified CD, and (b) conventional and modified FSD antennas. Dimensions (in mm) of the conventional CD antenna: la = 6.5, lb = 7.5, g = 1.0, lc = 9.5.
Dimensions (in mm) of the modified CD antenna: la = 6.3, lb = 7.0, g = 1.0, lc = 9.3.
Dimensions (in mm) of the conventional FSD antenna: la = 5.0, lb = 5.0, g = 1.0, ls = 7.0.
Dimensions (in mm) of the modified FSD antenna: la = 5.0, lb = 5.0, g = 1.0, ls = 8.0.
insertion depth of each antenna in pork loin and observed that the |S11 | measurement was
unchanged. This indicates that the baluns effectively suppressed the currents on the outer
surfaces of the feed lines of the antennas.
Each ablation was conducted with a power level of 30 W, applied at the input SMA
connector of the feed line of the antenna, for 5 minutes. The microwave signal was generated
by a signal generator (HP 8350B Sweep Oscillator) connected to a high-power traveling wave
tube amplifier (IFI T186-40) and introduced to the input of the antenna via a flexible coaxial
55
(a)
10
y [mm]
y [mm]
10
0
−40
−20
z [mm]
−40
0
(c)
10
y [mm]
y [mm]
0
−10
−10
10
(b)
0
−10
−40
−20
z [mm]
0
−20
z [mm]
0
(d)
0
−10
−20
z [mm]
0
−40
Figure 4.4: Simulated 60◦ C contours (black) and -10 dB normalized SAR contours (gray) at
7 GHz of the (a) conventional CD, (b) modified CD, (c) conventional FSD, and (d) modified
FSD antennas.
cable. Reflected power, monitored at the output port of the power amplifier, never exceeded
1 W for all ablation experiments, confirming that all antennas maintained good impedance
matching during the whole ablation duration.
Upon completion of each ablation, we dissected the pork loin to examine the ablation
zone in a cut plane through the insertion path of the antenna. Fig. 4.5 shows photographs of
the ablation zones produced by the four antennas. Each ablation zone was visually distinct
from normal pork loin tissue (pink) as it consisted of a small region of charred tissue (dark
brown/black) in the center and a coagulated zone (white). The charred regions in the
ablation zones confirm the regions of strongest energy deposition surrounding the dipole
arms as indicated by the simulated -10 dB SAR patterns of the antennas. We measured
the dimensions of each ablation zone in terms of the maximum long-axis diameter and
56
maximum short-axis diameter of the boundary of the coagulation zone. Those ablation-zone
dimensions were 4.1 cm × 2.7 cm for the conventional CD, 4.0 cm × 2.8 cm for the modified
CD, 4.3 cm × 2.8 cm for the conventional FSD, and 3.7 cm × 2.8 cm for the modified FSD
antenna. The maximum short-axis and long-axis diameters of the ablation zones agree well
with the predicted values from the thermal simulations for all antennas. The dimensions of
ablation zones of the regular and modified CD antennas are comparable to each other. The
ablation zone produced by the modified FSD antenna is closer to a spherical shape than
that produced by its original version. The variation in ablation zone dimensions produced
by these antennas is likely due to heterogeneity of different pork loin samples. Nevertheless,
all antennas achieved localized ablation zones with minimized tail heating along the shaft of
the antennas, indicating the the coaxial baluns work effectively on both the conventional and
modified versions of both antenna types. Moreover, the modified antennas performed as well
as the conventional versions in terms of ability to produce good impedance matching and
localized heating while having 30% smaller overall diameters. This proves the effectiveness of
the proposed design concept in reducing overall diameters of coax-fed balun-equipped MWA
antennas.
4.4
Conclusion
We presented a new strategy for reducing the overall diameter of a coax-fed interstitial
antenna equipped with a coaxial balun by using a coaxial transition between a Teflonfilled coax feed and an air-filled coax section. Due to the lower relative permittivity of air
compared to Teflon, the inner diameter of the outer conductor of the air-filled coax section
is 30% smaller for the same characteristic impedance and inner conductor diameter. By
implementing the coaxial baluns on this air-filled coax section, we achieve a reduced-diameter
design relative to conventional coaxial balun implementations. Using this design concept,
we developed modified choke and floating sleeve dipole antennas for operation in ex vivo
57
4.0
4.1
2.8
2.7
(b)
(a)
3.7
4.3
2.8
2.8
(c)
(d)
Figure 4.5: Photographs of ablation zones generated by the fabricated (a) conventional CD,
(b) modified CD, (c) conventional FSD, and (d) modified FSD antennas using an input
power of 30 W at 7 GHz for 5 minutes. The measured values for the maximum long-axis
diameters and maximum short-axis diameters of the ablation zones are displayed in cm.
porcine muscle tissue. The modified antennas were compared in simulations and ablation
experiments with conventional choke and floating sleeve dipole antennas constructed out of
the same type of semirigid coaxial cables. All four antennas were shown to be well matched
to the feed lines and produced localized ablation zones. However, the modified antennas
delivered similar ablation performance while having 30%-smaller diameters compared to
58
the conventional antennas. The coaxial transition is not expected to negatively impact
the average power handling capability of the antenna, since the average power handling
capability of the coax is limited by the maximum temperature level in the Teflon region.
The proposed designs provide a practical solution for reducing the invasiveness of interstitial
antennas equipped with baluns without sacrificing MWA performance.
59
Chapter 5
A Minimally Invasive Coax-Fed
Microwave Ablation Antenna with a
Tapered Balun
60
5.1
Introduction
Coaxial baluns represent the most widely used solution for choking the undesired outersurface currents for coax-fed microwave ablation (MWA) antennas [55]. A coaxial balun
is typically in the form of a third cylindrical conductor with a circular cross section that
encompasses the coaxial feed line (e.g. [55], [57], [63], [64], [70], [85], [86]). The third conductor and the outer-conductor section of the enclosed coax feed constitute a new coaxial
transmission line for the outer-surface currents of the feed line. The length and termination
at the proximal end of the new transmission line can be properly chosen to create a choke
point with high impedance at its distal end to effectively suppress the outer-surface currents. However, this balun implementation increases the overall diameter and invasiveness
of a typical coax-fed interstitial antenna.
Several studies have presented alternative solutions for reducing the invasiveness of coaxfed interstitial antennas. In [70], the metallic biopsy needle used to introduce the antenna
to an ablation site was also leveraged to serve as an adjustable choke. The short-circuited
end of the choke is realized by using a copper collar, soldered to the outer conductor of the
coaxial feed line at a fixed position, to make an electrical contact between the biopsy needle
and the outer conductor. This implementation eliminates an extra layer of conductor used to
make a choke in a conventional implementation, hence reducing the overall diameter of the
antenna. In the triaxial antenna design reported in [65], a metallic biopsy needle was used
as a sleeve to adjust the insertion depth of a monopole antenna in order to achieve good
impedance matching and reduce the currents flowing on the outer surface of the coaxial
feed line. The triaxial antenna provides a very low input reflection coefficient but its specific
absorption rate (SAR) pattern has more elongated tails than that of a typical balun-equipped
antenna. Other antennas that achieve localized SAR patterns without using coaxial baluns
include double-slot antennas [66]- [67] and a balun-free, base-fed monopole [90]. In [67], the
double-slot antenna, based on and modified from the original design reported in [66], was
61
optimized to produce a more spherical heating pattern and a low input reflection coefficient.
The optimized antenna geometry has two annular slots separated by approximately half a
wavelength along the outer conductor of the coax feed. These two slots create destructive
interference to suppress the currents flowing on the outer surface of the coax feed. However,
the outer conductor of the coax is in direct contact with the surrounding tissue, causing the
antenna’s performance (e.g. impedance matching and SAR pattern) to be more sensitive
to changes in the dielectric properties of tissue compared to other antenna designs that use
insulated sheaths. The balun-free, base-fed monopole antenna reported in Chapter 2 ( [90])
achieves a localized SAR pattern by operating the monopole at the second resonant frequency,
where the currents are minimized at its base, thereby effectively suppressing the currents
excited on the outer surface of the coax. The very high input impedance at the base of the
monopole at the second resonant frequency requires an internal impedance matching section
to match the monopole to a standard 50 Ω coax feed. However, this matching section only
provides a good impedance match over a narrow bandwidth (e.g. about 200 MHz) around
the operating frequency.
In [91], we reported a coax-fed MWA antenna that used a non-invasive tapered balun to
achieve a localized ablation zone and a wideband impedance match for operation in egg white.
While the preliminary results presented in [91] demonstrate the potential of the proposed
antenna design, greater details of the design process and operating principle along with
more thorough simulations and experimental characterization of the antenna are warranted.
Therefore, in this chapter, we present a more comprehensive study of this antenna design.
The balun equipped for this antenna uses a coaxial line with a tapered outer conductor.
Such tapered coaxial cables have been used as wideband baluns to transform unbalanced
coaxial lines to balanced parallel-wire lines [92]- [93]. In these tapered balun designs, the
outer conductor is gradually tapered to make a smooth transition to a strip that has the
same width as that of the inner conductor at the distal end. The optimum taper profile
that provides the lowest input reflection coefficient, given a certain length of the taper line,
62
was presented by Klopfeinstein in [94]. However, to simplify the design and fabrication
process of the tapered coaxial balun demonstrated in this study, we used a linear tapering
profile for the width of the outer conductor. We investigated two coax-fed dipole antennas
using two different topologies for the tapered slot balun: a single-slot and a double-slot.
The dipole connected to the single-slot balun has two active segments. One segment is an
extension of the inner conductor, which forms one arm of the dipole. The other segment is
connected to the distal end of the tapered outer conductor and located in the slot created
by the tapered outer conductor, forming the second arm of the dipole. The dipole using
the double-slot balun is similar to the single-slot design with the exception that it uses a
symmetric implementation of conductor and slot patterns to implement the balun and the
second arm of the antenna. The design and simulation results for these two antennas are
presented in Section 5.2. Simulations show that the symmetric implementation provides
a better impedance match and SAR pattern compared to the asymmetric configuration.
Therefore, we fabricated the antenna with the double-slot balun and used it to perform ex
vivo ablation experiments in bovine livers. The experimental setup and results are presented
and discussed in Section 5.3.
5.2
Antenna Designs and Simulations
5.2.1
Simulation Methods and Assumptions
Both electromagnetic (EM) and thermal simulations were performed for the two antennas
considered in this work. The antennas were modeled based on the dimensions of a 50-Ω UT085C semirigid coaxial cable from Micro-Coax. The outer diameters of the inner conductor,
Teflon insulation layer, and outer conductor of this coaxial cable are 0.511 mm, 1.676 mm,
and 2.192 mm, respectively. While the most commonly used frequencies for MWA are below
2.5 GHz, the use of higher frequencies (e.g. 10 GHz) for MWA has been demonstrated with
several advantages over the use of lower frequencies (e.g. 1.9 GHz), including higher direct
63
heating rates, more spherical heating zones, and shorter active lengths of the antennas [87].
However, using higher frequencies also comes with some disadvantages that are attributed
to higher ohmic losses of the conductors, such as more unwanted heating and lower power
handling capabilities of the feed cables. In this study, an operating frequency of 6 GHz
was chosen for the antennas for a balance between the advantages and the disadvantages of
high-frequency MWA.
Full-wave EM simulations of the antennas were conducted using CST Microwave Studio
to design them to operate at 6 GHz in liver tissue. Dielectric properties of liver at room temperature were modeled using the one-pole Cole Cole model presented in [33] for the frequency
range from 0.5 to 8 GHz. Absorption of electromagnetic fields in tissue, calculated from the
EM simulations, was scaled for an input power of 20 W to be the heat source in transient
thermal simulations in CST Multiphysics Suite. Thermal properties, including specific heat
capacity and thermal conductivity, of liver were modeled as temperature-dependent. The
specific heat capacity of liver tissue was modeled using the method described in [95] to account for water evaporation as a function of tissue temperature in the range from 20◦ C to
over 100◦ C. The thermal conductivity of liver was modeled using the measurement data presented in [96] for tissue temperature in the range of 20◦ -90◦ C. The thermal conductivity for
temperature levels above 90◦ C, due to the lack of such data in the literature, was assumed
to be constant and equal to that at 90◦ C. Simulated 50◦ C contours were used to predict
the boundaries of the ablation zones since this temperature level is widely considered as the
lower threshold for inducing tissue coagulation [6].
The simplified EM and thermal simulation models we used omitted several physical
phenomena such as the dependence of dielectric properties of liver on temperature, changes
of energy deposition in tissue caused by changes of dielectric properties with temperature,
and changes in thermal conductivity of liver for temperature above 90◦ C. While data for
temperature-dependent dielectric properties of liver tissue are available in literature (e.g. [33],
[37]), CST lacks a coupled EM-thermal simulation scheme that automatically and frequently
64
switches back and forth between EM and thermal solvers to account for the effect of the
changes in dielectric properties of tissue on the heating process. The deficiency of the
simulation models leads to higher predicted temperature values in the vicinity of the antennas
and along the antenna shafts compared to experiments. We noticed that the simulated 50◦ C
contour exhibited a longer tail along the antenna shaft than the ablation zones achieved
in ablation experiments, with the same setting of input power level (20 W) and duration
(5 minutes), did. However, if we disregard this incorrectly predicted elongated tail in the
simulated 50◦ C contour, the general dimensions and shape of the predicted ablation zone
are very close to the experimentally achieved ones.
5.2.2
Interstitial Antenna Design with a Single-Slot Tapered Balun
Fig. 5.1 shows the topology of the dipole antenna with a single-slot tapered balun. The
outer conductor is linearly tapered into a single strip which along with the inner conductor
form a balanced parallel-wire line to feed a dipole antenna. One arm of the dipole is created
by extending the inner conductor. The other arm is placed in the slot of the tapered outer
conductor and connected to the strip at its distal end. At its operating frequency, the balun
provides balanced currents for two arms of the dipole. As a result, unbalanced currents
flowing back on the outer surface of the outer conductor are minimized. This is expected to
help the antenna provide a localized SAR pattern at the operating frequency.
We optimized the parameters of the dipole and those of the tapered balun to achieve
localized SAR and heating patterns and a good impedance match between the antenna and
the feed line at 6 GHz. These parameters include the lengths of the dipole arms (la and
lb ), the width of the second arm of the dipole (w2 ), the width of the strip at the end of the
tapered outer conductor (w1 ) and the width of the conducting ring that connects the second
dipole arm to the tapered outer conductor (w3 ). All of these parameters are marked in the
different views of the antenna shown in Fig. 5.1. The optimized values for these parameters
are reported in the caption of Fig. 5.2.
65
la
x
z
Tapered
outer conductor
y
Active segments
z
(a)
lb
w3
w2
w1
lt
(b)
Figure 5.1: Topology of the interstitial antenna design with a single-slot tapered balun and
two active segments. (a) Views in the x-z and y-z planes. (b) Drawing of the tapered outer
conductor and the active segment connecting to it when they are unrolled and placed on a
flat surface.
Fig. 5.2(a) shows the simulation result for the input VSWR, seen at the input of the
coaxial feed line, of the antenna. The VSWR value is approximately 1.38 corresponding to
a |S11 | value of -16 dB at 6 GHz, demonstrating that the single-slot tapered balun provides
a good impedance match between the dipole and the feed line. Indeed, the antenna exhibits
a wide band of impedance matching with a VSWR lower than 2 from 4 GHz to over 8 GHz.
Figs. 5.2(b) and 5.2(c) show the normalized SAR pattern of the antenna at 6 GHz, plotted
66
5
20
10
x [mm]
VSWR
4
3
2
SAR pattern
0
−10
1
2
4
6
Frequency [GHz]
8
−20
−60
−40
20
SAR pattern
x, y [mm]
y [mm]
10
0
-10 dB
-20 dB
-30 dB
−10
−20
−60
−40
−20
z [mm]
(c)
0
(b)
(a)
20
−20
z [mm]
-10 dB
-20 dB
-30 dB
10
x−z plane
y−z plane
r1
0
r2
−10
0
−20
−60
−40
−20
z [mm]
0
(d)
Figure 5.2: Simulation results for the antenna with the single-slot tapered balun and two
active segments shown in Fig. 5.1. (a) Input VSWR. (b) Normalized SAR pattern in the
x-z plane. (c) Normalized SAR pattern in the y-z plane. (d) Thermal simulation results
for 50◦ C contours in the x-z plane (red contour) and y-z plane (black contour) after 5
minutes of ablation using 20 W input power. r1 = 10.75 mm and r2 = 13.25 mm. Yellow
rectangles represent the position of the single-slot tapered balun. The dimensions for the
active segments and the tapered balun of the antenna are as follows: la = 7 mm, lb = 6.2
mm, lt = 16 mm, w1 = w2 = w3 = 0.5 mm.
in two orthogonal cut planes. SAR levels are reduced by 20 dB near the proximal end of
the second dipole arm connected to the tapered outer conductor and by 30 dB near the
proximal end of the tapered balun. This shows that outer-surface currents are suppressed
effectively with the use of the single-slot tapered balun, resulting in a compact SAR pattern
with minimal tails along the shaft of the feeding cable. However, Fig. 5.2(c) also reveals
an asymmetric SAR pattern in the y-z plane. This is due to the lack of symmetry in the
placement of the second dipole arm that is connected to the tapered outer conductor in
the y-z plane. As the result, EM radiation is stronger on the lower side, where the second
67
dipole arm is located, compared to the upper side of the antenna axis. Fig. 5.2(d) shows
the simulated 50◦ C contour, used to predict the boundary of the ablation zone, in the xz and y-z planes at the end of a 5-minute ablation using an input power of 20 W. The
ablation zone is symmetrical in the x-z plane and slightly asymmetrical in the y-z plane, as
a result of the corresponding SAR pattern in these two cut planes. However, the maximum
lateral dimensions of the ablation zone in the two cut planes are the same and equal to
24 mm. In the y-z plane, the maximum lateral radius of the ablation zone is 13.25 mm
for the lower side and 10.75 mm for the upper side. While this degree of asymmetry is
not significant, further modification of the antenna design to create a more asymmetric
ablation zone can be implemented such as tilting the two dipole arms further toward the
lower half of the y-z plane (in Fig. 5.1(a)) or deploying a hemicylindrical metallic reflector
similar to the one presented in [97] to prevent EM radiation on the upper half of the y-z
plane. Such asymmetrically enhanced heating is desirable in certain clinical scenarios where
tumors are highly asymmetric or where an antenna has to heat a tumor from a peripheral
position because the central region of the tumor is inaccessible (e.g. blockage by other vital
organs) [98].
5.2.3
Interstitial Antenna Design with a Double-Slot Tapered Balun
Fig. 5.3 shows the topology of the dipole antenna with a double-slot tapered balun. The
outer conductor is gradually tapered into two parallel strips instead of just one as for the
single-slot tapered balun. This taper profile leaves two opposite slots on the outer conductor
for placing two active segments of the dipole that are connected to the the distal end of
the tapered outer conductor. These two active segments are connected to each other and
constitute one arm of the dipole. The other arm of the dipole is defined by an extension
of the inner conductor. At the operating frequency of the balun, the total currents flowing
on the two active segments connected to the tapered outer conductor are balanced by the
currents flowing on the segment connected to the inner conductor.
68
la
x
z
Tapered
outer conductor
y
Active segments
z
(a)
lb
w3
w2
w1
lt
(b)
x
z
y
z
(c)
Figure 5.3: Topology of the interstitial antenna design with a double-slot tapered balun and
three active segments. (a) Views in the x-z and y-z planes. (b) Drawing of the tapered outer
conductor and the two active segments connecting to it when they are unrolled and placed on
a flat surface. (c) Photographs of the fabricated outer conductor for the double-slot tapered
balun.
69
The dimensions of the double-slot balun and the three active segments of the dipole,
noted in Figs. 5.3(a)-(b), were optimized in the simulations in order for the antenna to
achieve a localized SAR pattern and good impedance match at 6 GHz. These optimized
dimensions are reported in the caption of Fig. 5.4. Fig. 5.4(a) shows the simulated input
VSWR of the antenna. The antenna exhibits good impedance matching with the feed line
over a wide frequency band from 5 GHz to over 8 GHz. The three-segment dipole connected
to the double-slot balun achieves a slightly better impedance matching (VSWR = 1.21,
|S11 | = −20 dB) than the two-segment dipole connected to the single-slot balun (VSWR =
1.38, |S11 | = −16 dB) at the operating frequency of 6 GHz. While increasing the length of
the tapered section helps reduce the reflection coefficient, the frequency of best impedance
match is most sensitive to the dimensions of the three active segments of the dipole and the
thickness of the outermost Teflon coating of the antenna.
Figs. 5.4(b) and 5.4(c) show the normalized SAR pattern at 6 GHz produced by the
dipole connected to the double-slot balun, viewed in the x-z and y-z planes. SAR values are
reduced by 30 dB near the proximal ends of the two active segments connected to the tapered
outer conductor. The antenna using the double-slot balun produces a highly compact volume
enclosed by the -30 dB contours, of which the lateral diameter (about 26 mm) is even larger
than the axial diameter (about 20 mm). Due to the lack of an axial symmetry of the antenna,
the SAR pattern in the x-z plane is slightly different from that in the y-z plane. Fig. 5.4(d)
shows the simulated ablation zone, bounded by the 50◦ C contours, viewed in the x-z and
y-z planes, in the tissue after 5 minutes of ablation using an input power of 20 W. Despite
the slight difference of the SAR pattern in the x-z and y-z planes, the ablation zone appears
to be identical in these two cut planes. This demonstrates that the antenna is capable of
producing a rotationally symmetric ablation zone.
Compared to the antenna using the single-slot balun, SAR values less than -30 dB fall
off faster along the shaft of the antenna using the double-slot balun. This is evident in the
slightly longer tails of the -30 dB contours shown in Figs. 5.2(b)-(c) in comparison with
70
5
x [mm]
4
VSWR
20
Simulation
Measurement
3
SAR pattern
10
0
-10 dB
-20 dB
-30 dB
−10
2
1
2
4
6
Frequency [GHz]
8
−20
−60
−40
20
SAR pattern
0
-10 dB
-20 dB
-30 dB
−10
−20
−60
−40
−20
z [mm]
(c)
0
x, y [mm]
y [mm]
10
0
−20
z [mm]
0
(b)
(a)
20
−20
z [mm]
10
x−z plane
y−z plane
0
−10
−20
−60
−40
(d)
Figure 5.4: Results for the antenna with the double-slot tapered balun and three active
segments shown in Fig. 5.3. (a) Simulation and measurement results for the input VSWR.
(b) Simulated, normalized SAR pattern in the x-z plane. (c) Simulated, normalized SAR
pattern in the y-z plane. (d) Thermal simulation results for 50◦ C contours in the x-z plane
(red contour) and y-z plane (black contour) after 5 minutes of ablation with 20 W input
power. Yellow rectangles represent the position of the double-slot tapered balun. The
dimensions for the active segments and the tapered balun of the antenna are as follows:
la = 7 mm, lb = 8 mm, lt = 18 mm, w1 = w3 = 0.5 mm, w2 = 0.7 mm.
Figs. 5.4(b)-(c). Moreover, the antenna using the double-slot balun produces a symmetric
SAR pattern in the y-z plane, as opposed to the asymmetric one in this cut plane of the antenna using the single-slot balun. While thermal simulation results show that both antennas
provide ablation zones with similar dimensions, the one provided by the antenna using the
double-slot balun is more rotationally symmetric. Overall, the better impedance matching
and more symmetric heating pattern make the antenna using the double-slot balun a more
desirable design for ablation applications where directional heating is not needed.
71
5.3
Ablation Experiments
We fabricated a prototype of the antenna with the double-slot tapered balun to conduct
ablation experiments in ex vivo bovine livers. The antenna was constructed out of a 50Ω UT-085C semirigid coaxial cable with an outer diameter of 2.2 mm. The two slots in
the outer conductor were created using laser cutting technology to form the tapered outer
conductor and the two active segments of the dipole connecting to it. Fig. 5.3(c) shows the
fabricated tapered outer conductor of the balun. The whole antenna assembly was embedded
in a Teflon coating with a diameter of 2.6 mm.
Prior to each ablation experiment, we inserted the antenna into a fresh ex vivo bovine
liver at room temperature and measured its input reflection coefficient using a vector network
analyzer (Agilent E5071C). Fig. 5.4(a) shows the input VSWR measurement result of the
antenna plotted against the simulation result. The measurement result is in agreement
with the simulation result and shows a good impedance matching between the antenna and
the feed line over a wide frequency range (from 5 GHz to over 8 GHz). At the operating
frequency of 6 GHz, the measured VSWR is 1.19 (|S11 | of −22 dB), which is slightly better
than the simulated values (VSWR = 1.21, |S11 | = −20 dB). Additionally, the measured input
impedance of the antenna was unchanged as we varied the insertion depth of the antenna,
as long as the whole tapered-slot balun section was immersed in liver. This confirmed that
the outer-surface currents were effectively suppressed along the feed line up to the starting
point of the tapered balun.
Microwave power at 6 GHz was generated by a signal generator (HP 8350B Sweep Oscillator) connected to a high-power amplifier (IFI T186-40). The output of the power amplifier
was connected to the input SMA connector of the antenna through a semirigid coaxial cable.
The loss of the semi-rigid coaxial cable was taken into account to adjust the output power
level of the amplifier for a desired input power level of the antenna. Given the rated output
power level of the amplifier (40 W) and the cable loss factor, the power levels applied at
72
the input of the antenna were limited to be equal to or less than 30 W for the experiments
reported in this study. Nevertheless, these input power levels were sufficient to create relatively large ablation zones and to demonstrate the ability of producing localized heating of
the fabricated antenna.
We conducted eight ablation experiments with four different combinations of applied
power level and the ablation duration to investigate the sizes and shapes of ablation zones
that can be achieved using the antenna across these different settings. The four combinations
are: 20 W for 5 minutes, 20 W for 10 minutes, 30 W for 5 minutes, and 30 W for 10 minutes.
Two experiments were conducted using each combination of the applied power level and the
ablation duration. Reflected power from the antenna was monitored and calculated through
the measured reflected power at the output port of the amplifier during each ablation. The
reflected power at the input of the antenna never exceeded 2 W for the experiments with 20
W input power, and 3 W for the experiments with 30 W input power, indicating that the
antenna maintained the good impedance matching through out the ablations.
Upon the completion of each ablation, we dissected the liver to reveal the ablation zone
in either of the two orthogonal cut planes (the x-z and y-z planes shown in Fig. 5.3). Fig.
5.5 shows photographs of the ablation zones created in 8 experiments. Figs. 5.5(a)-(d) show
the ablation zones in the x-z cut plane and Figs. 5.5(e)-(f) show the ablation zones in the y-z
cut plane. Each ablation zone consists of three concentric regions: a small charred region in
the center (black), a coagulated region (brown) in the middle, and an outer congested region
(pink). The ablation zones are relatively symmetric with respect to the antenna axis in most
of the photographs. Slight asymmetry observed in some cases (e.g. Figs. 5.5(b)-(d)) is due
to inhomogeneous structure of liver, specifically with the presence of blood vessels. Figs.
5.5(b)-(d) show evidence of blood vessels near the boundary of the ablation zones. These
vessels created channels for conducting hot fluid or water vapor from locations closer to the
central heating region in the vicinity of the antenna. The heat conducted through the vessel
walls ablated surrounding tissues and therefore, expanding the boundaries of the ablation
73
Figure 5.5: Photographs of ablation zones produced by the fabricated antenna with the
double-slot tapered balun by using (a), (e) 20 W for 5 min., (b), (f) 20 W for 10 min., (c),
(g) 30 W for 5 min., (d), (h) 30 W for 10 min. Ablation zones in the x-z plane are shown
in (a)-(d). Ablation zones in the y-z plane are shown in (e)-(h). Locations of blood vessels
seen in the ablation zones are marked in (b)-(d) and (g).
zones towards the locations of these vessels. Nevertheless, the ablation zones created by the
antenna in all experiments were confined and no excessive tail heating occurred.
The dimension of each ablation zone was recorded as the maximum long-axis diameter
and the maximum short-axis diameter of the boundary of the congestion zone. An aspect
ratio, defined as the ratio of the maximum long-axis diameter and the maximum short-axis
diameter, was calculated to quantify the shape of each ablation zone. An aspect ratio of 1
represents a spherical shape while a larger aspect ratio represents a more elongated ablation
zone. Table 5.1 reports the dimensions and aspect ratios of the ablation zones shown in the xz and y-z planes. The short-axis diameters of the ablation zones in the 5-minute experiments
using 20-W input power are 2.4 cm and 2.5 cm, agreeing well with the predicted value of
2.4 cm from the thermal simulations (shown in Fig. 5.4(d)). For the same combination
of the input power and duration, the sizes of the ablation zones in the two cut planes are
comparable. This also agrees with the thermal simulation results and suggests that the
74
Table 5.1: Dimensions of The Ablation Zones Produced by The Fabricated
Antenna, Cut in The x-z and y-z Planes
Cut
Experiment
Size
Plane
Conditions
[cm × cm]
20 W, 5 min.
x-z
y-z
AR
Photograph
3.2 × 2.5
1.28
Fig. 5.5(a)
20 W, 10 min.
3.8 × 3.0
1.27
Fig. 5.5(b)
30 W, 5 min.
3.7 × 2.8
1.32
Fig. 5.5(c)
30 W, 10 min.
4.6 × 3.5
1.31
Fig. 5.5(d)
20 W, 5 min.
3.0 × 2.4
1.25
Fig. 5.5(e)
20 W, 10 min.
4.2 × 3.2
1.31
Fig. 5.5(f)
30 W, 5 min.
3.4 × 2.6
1.31
Fig. 5.5(g)
30 W, 10 min.
4.3 × 3.5
1.23
Fig. 5.5(h)
AR: aspect ratio
antenna is capable of producing relatively rotationally symmetric ablation zones despite the
lack of an axial symmetry in the active segments of the dipole. The aspect ratios of the
ablation zones range from 1.23 to 1.32, indicating that the ablation zones are compact and
close to a spherical shape. The variation in the aspect ratio is small with respect to changes
in input power and ablation duration.
5.4
Conclusion
We presented a coax-fed interstitial antenna design for MWA using a tapered slot balun. The
balun is created by tapering the outer conductor of the coax to make a smooth transition
from a coaxial line to a parallel-wire line. This balun design is non-invasive compared
to the conventional coaxial baluns (e.g. short-circuited choke and floating sleeve), which
add an extra layer of conductor over the feed line of MWA antennas. Two tapered-slot
balun designs, a single-slot and a double-slot balun, were investigated for their use with
corresponding dipole antennas. Simulation results demonstrate that both antennas achieve
good impedance matching over a wide bandwidth and localized SAR patterns at the designed
75
frequency of 6 GHz for operation in ex vivo bovine liver. The antenna using the single-slot
balun provides a slight asymmetric heating in the plane containing the axes of the dipole
arms. The degree of asymmetry can be further enhanced (e.g. by tilting the two dipole arms
or using a back semicylindrical metallic reflector) to make the antenna suitable for clinical
applications where directional heating is needed. For ablating tumors with symmetric shapes,
the antenna using the double-slot balun is a more preferable design since it provides more
symmetric heating and a slightly lower input reflection coefficient than the other antenna.
A prototype of the antenna using the double-slot tapered coaxial balun was fabricated
and tested in ablation experiments in ex vivo bovine livers. Four different combinations of
input power and duration were used for the experiments: 20 W for 5 minutes, 20 W for 10
minutes, 30 W for 5 minutes, and 30 W for 10 minutes. Measurement results show that
the antenna provided good impedance matching and localized, symmetric ablation zones, as
predicted by the simulation results. The antenna produced ablation zones with dimensions
of up to 4.6 × 3.5 × 3.5 cm3 for the case of a 10-minute ablation using 30-W input power.
The dimensions of the ablation zones achieved in the experiments with 20 W power for 5
minutes are very similar to those predicted by the thermal simulation results if the elongated
tails are excluded from the simulated ablation zones. The aspect ratios of the ablation zones
were relatively consistent across different settings of power and duration, varying from 1.23
to 1.32. The presented antenna design exhibited good ablation performance and offers a less
invasive solution compared to coax-fed MWA antennas equipped with conventional coaxial
baluns.
76
Chapter 6
A Non-Coaxial-Based Balanced
Antenna for Microwave Ablation
77
6.1
Introduction
Coaxial cables are by far the most widely used types of transmission lines for feeding interstitial antennas primarily because they are compact, well shielded, and compatible with
a variety of microwave instruments. Coaxial cables, however, are unbalanced transmission
lines. If a coaxial cable is used to directly feed a balanced antenna (e.g. a dipole), electric
currents can get excited on the outer surface of the outer conductor of the coax. These
undesired currents flow towards the source and can ablate the healthy tissue surrounding
the shaft of the antenna along its insertion path [55]. Additionally, these currents cause
problematic changes in the input impedance of the antenna as a function of the insertion
depth. These undesired features are typically eliminated by equipping coax-fed interstitial
antennas with coaxial baluns that suppress these currents and produce localized specific
absorption rate (SAR) patterns. However, the use of a conventional coaxial balun often
increase the overall diameter of a coax-fed antenna. Alternative approaches to suppress the
undesired outer-surface currents of the coaxial feed lines while avoiding the added bulkiness
of a conventional coaxial balun were presented in Chapters 3, 4, and 5. In this chapter, we
propose the use of a balanced transmission line to feed a balanced antenna in order to avoid
the inherent problem associating with the unbalanced nature of coaxial cables [99].
The non-coaxial-based balanced antenna presented in this chapter consists of a dipole
fed by a shielded two-wire transmission line. Because of the inherently balanced nature
of the antenna and its feed line, no current is excited on the outer surface of the floating
shield. This allows the antenna to generate localized SAR patterns over a wide bandwidth
without using a coaxial balun. The proposed antenna is designed to operate at 10 GHz in
porcine muscle. A prototype of the antenna is fabricated and experimentally characterized.
Simulation and measurement results of this prototype confirm that the antenna exhibits
a good impedance match and compact SAR patterns over a wide frequency band. The
fabricated prototype was used to perform ex vivo ablations in porcine muscle using 18 W
78
of power applied for 10 minutes. In each experiment, confined ablation zones were obtained
and no ablation occurred along the antenna shaft. These desirable features stem from the
balanced nature of the antenna and its feed network and illustrate the suitability of the
proposed antenna for MWA applications. Furthermore, the performance of this antenna is
compared in simulations and in experiments with that of a coaxial-fed MWA antenna [65]
of the type used in an FDA-approved commercially-available MWA system (Certus 140,
Neuwave Medical Inc.). We demonstrate that the proposed balanced antenna can produce
significantly more spherical ablation zones compared to the triaxial antenna. In what follows,
the principles of operation, details of design process, and the experimental results of this
study are presented and discussed.
6.2
Antenna Design and Simulations
Fig. 6.1 shows the topology of the proposed dipole antenna fed by a balanced transmission
line. The line consists of two parallel wires that are insulated from each other. The current
flowing in one wire is balanced by that flowing in the other one. The two-wire line is placed
inside a floating shield. The shield isolates the fields of the two-wire line from the lossy
biological tissue surrounding the feed line. This shielded two-wire transmission line provides
a similar shielding effectiveness as that of a conventional coaxial cable. This ensures that the
fields of the transmission line are entirely confined to the region within the floating shield and
do not penetrate the surrounding tissue. At one end of the line, the two wires are extended
beyond the floating shield and bent to create the two arms of a dipole with a flare angle, α,
between them. Because of the inherently balanced nature of the antenna and its feed, the
antenna is expected to generate localized SAR patterns over a wide frequency band.
In the configuration shown in Fig. 6.1, the tip-to-tip distance between the dipole arms
may exceed the outer diameter of the floating shield. This, however, is not expected to
increase the invasiveness of the antenna. In practice, we envision that the dipole antenna
79
Floating shield
Balanced line
Dipole
d1
s
d2
d3
y
α
ld
y
x
(a)
dt
z
(b)
Figure 6.1: Topology of a balanced antenna comprised of a dipole fed by a two-wire shielded
balanced transmission line. (a) Cross-sectional view of the two-wire line. (b) Side view of
the feed line and the antenna. Black represents copper, light gray represents Teflon, and
white represents air. Dimensions for the shielded two-wire line are as follows: d1 = 0.72
mm, d2 = 2.04 mm, d3 = 2.50 mm, and s = 0.20 mm. The floating shield is embedded in a
Teflon catheter with an overall diameter of 3.0 mm. Dimensions for the dipole are as follows:
ld = 3.5 mm and α = 60◦ . The lateral tip-to-tip distance between the two dipole arms, dt ,
is 5.1 mm.
and the two-wire feed line will be guided through the floating shield to the ablation site and
the dipole arms will be deployed once in position. In this case, a metallic biopsy needle
may serve as the floating shield. Similar mechanisms have been developed for deployment
of expandable electrode applicators for RF ablation [100] and are currently available and
used in commercial RF ablation systems (e.g. RITA Model 70 coaxial probe, RITA Medical
Systems, Inc., Mountain View, California).
A prototype of the balanced antenna shown in Fig. 6.1 was designed for operation at 10
GHz in ex vivo porcine muscle (e.g. pork loin). The tissue was modeled as a homogeneous
rectangular cuboid of porcine muscle with the dimensions of 30×30×50 mm3 . The dielectric
properties of porcine muscle were obtained from open-ended coaxial probe measurements
performed on the surface of room-temperature pork loin samples from 0.5 GHz to 15 GHz.
The samples used to perform these dielectric characterizations were not the same pork loin
samples used in subsequent ablation experiments discussed in Section 6.3 and Section 6.4;
thus natural variations in dielectric properties from one tissue batch to another were present
60
30
50
25
40
20
30
15
20
10
10
5
0
0
5
10
σeff [S/m]
εr
80
0
15
Frequency [GHz]
Figure 6.2: Measurement results for the relative permittivity and effective conductivity of
pork loin. Each curve represents the average of nine measurements (three measurement sites
on each of three samples).
in our study. The relative permittivity and effective conductivity data that was imported
into CST Microwave Studio to model the tissue is shown in Fig. 6.2. The antenna was
inserted at an insertion depth of 30 mm along the longest edge of the tissue block. The
physical dimensions of the antenna and the feed line are given in the caption of Fig. 6.1.
The shielded two-wire transmission line with the dimensions as shown has a characteristic
impedance of 50 Ω. For comparison, a 50-Ω coaxial cable having an inner conductor with
the same diameter as one of the conductors of the two-wire line (i.e., 0.72 mm) will require
an outer conductor with an inner diameter of 2.4 mm. If the outer conductor of the coax has
the same wall thickness as the floating shield, then the shielded two-wire transmission line
used in this design will have a slightly narrower outer diameter for the same characteristic
impedance.
Initial simulations were performed by exciting the antenna with a discrete port that
differentially excites the parallel wires. The dimensions of the dipole arms and the angle
between them were optimized to achieve a good impedance match between the antenna
and the transmission line at 10 GHz. We varied the flare angle with an increment of 10◦
81
to investigate its effect on the input reflection coefficient of the antenna. In the dipole
configuration shown in Fig. 6.1, the flare angle should be less than 80◦ to prevent the dipole
arms from making electrical contact with the floating shield. Fig. 6.3(a) shows the input
VSWRs of the antenna for six different cases where the flare angle, α, is changed from 20◦
to 70◦ with an increment of 10◦ while the dipole arm lengths, ld are fixed at 3.5 mm. The
VSWR values at 10 GHz for these flare angles are specifically provided in the caption of Fig.
6.3. In general, the impedance matching of the antenna improves as the flare angle, α, is
increased. We chose the flare angle of 60◦ for the dipole since it provides a very low VSWR
and not a significant improvement in the VSWR is observed by increasing α beyond 60◦ .
Due to the inherently balanced nature of the antenna and its feed, the antenna provides
localized SAR patterns across a very wide frequency band. This is illustrated in Figs. 6.3(b)
and 6.3(c), wherein -30 dB SAR contours are compared at 7 GHz, 10 GHz and 13 GHz in
two orthogonal cut planes for the case where α = 60◦ and ld = 3.5 mm. Since the dipole
configuration is not rotationally symmetric with respect to the z-axis, there are differences in
the size and shape of the SAR patterns in one cut plane compared to the other. Specifically,
the -30 dB contour in the y-z plane for the same frequency is slightly larger than that in
the x-z plane. This can be attributed to the larger physical extension of the dipole in the
y-z plane. The wideband nature of this antenna suggests that it can maintain its good
impedance match and confined SAR pattern at a fixed frequency despite the large changes
in the dielectric properties of its surrounding environment that occur during ablation.
An external coax to two-wire transmission line adapter is needed because the power
amplifier that we use for our experiments has a standard 50 Ω coaxial output, which cannot
be directly connected to the two-wire balanced transmission line feeding the antenna. This
adapter, or transformer, is placed outside of the body and does not adversely impact the
response of the interstitial antenna or increase its invasiveness and dimensions. Fig. 6.4 shows
the topology of a well-known microstrip line to coplanar stripline transition that we use as
the transformer [101]. The transformer is implemented using a two-sided printed circuit
82
5
α = 20°
α = 30°
α = 40°
α = 50°
α = 60°
α = 70°
VSWR
4
3
2
1
6
10
7
8
9
10
11
Frequency [GHz]
12
13
10
7 GHz
10 GHz
5
7 GHz
10 GHz
5
13 GHz
13 GHz
y [mm]
x [mm]
(a)
0
−5
0
−5
(b)
−10
−20
(c)
−15
−10
−5
z [mm]
0
5
10
−10
−20
−15
−10
−5
z [mm]
0
5
10
Figure 6.3: Simulation results for the balanced antenna for the situation where the two-wire
shielded transmission line feeding the antenna is differentially fed using a lumped port in
CST Microwave Studio. (a) Input VSWRs for various values of the flare angle α. The VSWR
values at 10 GHz are 1.67, 1.35, 1.25, 1.10, 1.08, and 1.08 for the flare angle of 20◦ , 30◦ , 40◦ ,
50◦ , 60◦ , and 70◦ , respectively. (b)-(c) Normalized SAR patterns at 7 GHz, 10 GHz, and 13
GHz in the (b) x-z plane and (c) y-z plane for the case where α = 60◦ .
board that creates a transition between a 50 Ω microstrip line and a coplanar stripline. On
the bottom side of the substrate, the two conductors of the coplanar stripline are connected
to the discontinued ground plane of the microstrip line. On the top surface, the microstrip
crosses the gap between the two strips of the coplanar stripline, creating a potential difference
between them and exciting balanced currents that flow in opposite directions in each strip. To
ensure this is done with maximum efficiency, we choose the length of the short-circuited slot
created between the two strips of the coplanar stripline, l2 , and the length of the open ended
microstrip line, l1 , to be approximately one quarter of the respective guided-wave wavelength
83
50 Ω SMA
connector
w2
External transformer
Balanced line
l3
Z0= 50 Ω
ws
w1
l1
l2
Zin= 135 Ω
Figure 6.4: Topology of an external transformer used to connect the two-wire transmission
line at the input of the antenna to the coaxial output of the power amplifier. The transformer is implemented on the double-sided printed circuit board. For the transformer, yellow
represents copper trace on one side of the substrate, and black represents copper trace on the
other side. For the balanced line, black represents copper, and light gray represents Teflon.
of each line. The width of the microstrip line, w1 , was selected to have a characteristic
impedance of 50 Ω. By changing the position of the crossing between the microstrip line and
the gap between the two strips of the coplanar stripline, the transformer can be matched to
a different input impedance of the balanced line.
From a fabrication standpoint, it is not practical to solder the parallel wires of the
balanced transmission line to the coplanar strips at the output of the transformer without
extending the wires out of the floating shield. However, an unshielded section of the two-wire
line has a characteristic impedance of 80 Ω rather than 50 Ω. The length of the unshielded
section of the two-wire line was selected to be 15 mm. This length is long enough to allow the
two conductors of the balanced shielded transmission line to be easily connected to the two
conductors of the coplanar stripline on the ground plane of the transformer. With this setup,
the input impedance of the entire arrangement of the antenna/feed line was determined via
simulation to be 135 Ω. Therefore, the transformer was designed to match this 135 Ω
input impedance to 50 Ω at the input of the transformer’s SMA connector. The optimized
dimensions for the transformer implemented on a Rogers RT/duroid 5880 substrate with
84
a thickness of 0.508 mm are as follows: w1 = 2.0 mm, w2 = 2.2 mm, ws = 0.4 mm,
l1 = 4.2 mm, l2 = 5.0 mm, and l3 = 0.8 mm. Fig. 6.5 shows the simulated transmission
and reflection coefficients of this transformer. These results demonstrate that the structure
provides minimal insertion loss and a good impedance match over the frequency band of 9
to 11 GHz.
0
−10
−0.5
−20
−1
−30
−1.5
−40
−2
−50
−2.5
−60
8
9
10
11
|S21 | [dB]
|S11 | [dB]
0
−3
12
Frequency [GHz]
Figure 6.5: Simulated reflection and transmission coefficients of the external transformer
shown in Fig. 6.4.
The simulated VSWR of the antenna, as seen from the coaxial input of the transformer,
is shown in Fig. 6.6. The overall input VSWR of the system remains below 2 over the
frequency band of 9-11 GHz and an excellent impedance match is obtained at the desired
frequency of operation of 10 GHz. Due to the finite bandwidth of the transformer, the overall
system bandwidth is narrower than the ultra-wide bandwidth achievable from the balanced
antenna alone (see Fig. 6.3(a)). However, even with the use of the external transformer,
the antenna still provides a wide bandwidth of more than 2 GHz over which both a good
impedance match and a compact SAR pattern is achieved. The simulated SAR patterns
produced by the antenna, when fed at the input of the external transformer, are presented
in Fig. 6.7. Fig. 6.7 shows the normalized SAR patterns at 9, 10, and 11 GHz in the x-z
and y-z planes. The antenna still provides highly localized SAR patterns in the frequency
85
5
Simulation
Measurement
VSWR
4
3
2
1
9
9.5
10
10.5
11
Frequency [GHz]
Figure 6.6: Simulated and measured input VSWRs of the balanced antenna when fed from
the input of the coax to two-wire line transformer. The curves represent the total input
VSWRs as seen from the input of the transformer.
band of 9-11 GHz despite the presence of the transformer. The SAR patterns at 10 GHz
are similar to those produced by the antenna simulated with the ideal balanced source as
shown in Fig. 6.3. This indicates that the transformer, which is located outside of the
biological tissue, can successfully be used to interface the two-wire shielded transmission line
feeding the antenna to the coaxial output of the power amplifier. Moreover, none of the SAR
patterns shown in Figs. 6.3 and 6.7 indicate any leakage of the EM fields from the shielded
transmission line into the lossy tissue surrounding it along the insertion path. This is further
evidence indicating the shielding effectiveness of the balanced transmission line used to feed
the antenna.
6.3
Ablation experiments
We fabricated a prototype of the proposed balanced antenna designed to work at 10 GHz
in ex vivo porcine muscle (e.g. pork loin). The shielded two-wire transmission line and the
86
−5
y [mm]
5
5
10
-30 dB
-20 dB
-10 dB
0
−5
−10
−20 −15 −10 −5 0
z [mm]
10
5 10
10 GHz
5
10
−10
−20 −15−10 −5 0
z [mm]
-30 dB
-20 dB
-10 dB
(e)
−10
−20 −15 −10 −5 0
z [mm]
5
11 GHz
0
−5
5 10
11 GHz
0
10
-30 dB
-20 dB
-10 dB
0
(d)
5
−5
(c)
5
−5
(b)
10
-30 dB
-20 dB
-10 dB
−10
−20 −15 −10 −5 0
z [mm]
y [mm]
−20 −15 −10 −5 0
z [mm]
9 GHz
0
−5
−10 (a)
10
5
10 GHz
x [mm]
0
10
-30 dB
-20 dB
-10 dB
y [mm]
5
9 GHz
x [mm]
x [mm]
10
5
10
-30 dB
-20 dB
-10 dB
(f)
−10
−20 −15 −10 −5 0
z [mm]
5
10
Figure 6.7: Normalized SAR patterns of the balanced antenna when the antenna is fed with
the coax to two-wire balanced line adapter shown in Fig. 6.4. (a)-(b) SAR patterns at 9
GHz in the (a) x-z plane and (b) y-z plane. (c)-(d) SAR patterns at 10 GHz in the (c) x-z
plane and (d) y-z plane. (e)-(f) SAR patterns at 11 GHz in the (e) x-z plane and (f) y-z
plane.
dipole antenna were fabricated manually in our laboratory. Each conductor of the two-wire
transmission line was implemented using a solid copper wire with a diameter of 0.72 mm,
which was covered with a Teflon coating with a shell thickness of 0.1 mm. The assembled
two-wire line was placed inside a hollow copper tube with the inner and outer diameters
of 2.05 mm and 2.5 mm, respectively. The floating shield was then embedded in a Teflon
catheter with the outer diameter of 3.0 mm. We emphasize that these specific dimensions
of the conductors and the floating shield were chosen primarily to simplify the fabrication
of this proof-of-concept prototype using the in-house fabrication facilities available to us.
As a result, the diameter of this specific prototype is larger than the diameter of a typical
commercial MWA antenna. However, this does not adversely impact the quality of the ex
vivo ablation experiments conducted in this work. For clinical applications, we envision
that smaller-diameter versions of the proposed antenna can be readily fabricated using wellestablished commercial processes used to fabricate multi-conductor cables and transmission
87
lines.
Prior to each ablation, we inserted the balanced antenna into pork loin and measured
the input VSWR of the system, as seen from the input of the transformer, with a vector
network analyzer (Agilent E5071C). Fig. 6.6 shows the measured VSWR of the fabricated
prototype. The measurement result also demonstrates that the antenna achieves a good
impedance match over the 9-11 GHz frequency range. The slight discrepancies observed
between the measured and simulated results are attributed to the uncertainties involved in
the fabrication process as well as the possible deviation of the dielectric properties of the
pork loin from those used in the simulations (the dielectric characterizations and the VSWR
measurements were performed using different pork loin samples).
After measuring the VSWR of the antenna, ablation experiments were conducted in ex
vivo pork loin. The experimental setup consisted of a signal generator (HP 8350B Sweep
Oscillator) connected to a high-power traveling wave tube amplifier (IFI T186-40). The
output of the power amplifier was connected to the input of the antenna through the external
transformer shown in Fig. 6.4. Eight ex vivo ablation experiments were conducted using
an input power level of 18 W for an ablation duration of 10 minutes each. This specific
power level used in our experiments was dictated by the power handling capability of the
metal/metal connections used in the transition between the external transformer and the
two-wire shielded line. Specifically, the connection between these two components were made
by soldering the two wires of the shielded line to the two output strips of the transformer
shown in Fig. 6.4. Whenever the input power exceeded 20 W, the temperature of this
connection would get hot enough to melt the solder and sever the connection. Thus, the
input power levels used in the experiments were kept below 20 W. To perform ablations at
a higher power level using this specific antenna prototype, a more heat-tolerant method of
connecting the output of the transformer to the two-wire input of the antenna is needed
(see Fig. 6.4). Alternatively, one can use a commercially-available single-to-differential
transformer instead of the device used in our experiments. Nevertheless, this input power
88
(a)
(b)
(c)
(b)
(e)
(f)
(g)
(h)
Figure 6.8: Photographs of ablation zones created in ex vivo pork loin in the (a)-(d) x-z
plane and (e)-(h) y-z plane by applying 18 W to the input of the balanced antenna for 10
minutes.
level used here was sufficient to perform the necessary ex vivo ablation experiments and to
experimentally demonstrate the performance of the antenna.
Upon the completion of each ablation, we cut the pork loin to inspect ablation zones in
one of the two orthogonal cut planes, either parallel (y-z plane) or perpendicular (x-z plane)
to the plane of the dipole, and measured the dimensions of the ablation zones. Fig. 6.8
shows photographs of eight ablation zones, with four in the x-z plane and the other four in
the y-z plane. Each ablation zone consists of a small region of charred tissue (black/dark
brown) in the center and a coagulated region (white) surrounding it. According to [102],
pork loin heated to temperature levels in the range of 63◦ -82◦ C is fully cooked and changes
color to slightly pink/white. This change in color was the criterion used to determine the
boundaries of the ablation zone. The dimensions of the ablation zone were measured using
a ruler with an accuracy of ±0.5 mm. As can bee seen from Fig. 6.8, the ablation zones are
fairly localized, with minimal tail heating as expected. We noticed that the ablation zones
are slightly elongated along the feed line, despite the almost spherical shapes of the SAR
patterns produced by the antenna. This elongation of the ablation zones along the floating
shield can be explained by noticing that the shape of the SAR pattern of an antenna is not
89
the only parameter that impacts the shape or size of the ablation zone. A good example
demonstrating this is given in [87] where it was demonstrated that the dimensions of the
ablation zones produced by a 10 GHz MWA antenna were considerably different from those
of its SAR pattern. These differences were shown to be due to thermal conduction and heat
diffusion. Both of these factors can play a significant role in determining the final shape
and size of the ablation zones obtained from an antenna with a given SAR pattern. In the
experiments conducted here, part of the heat generated in the vicinity of the antenna is
carried away from that region through thermal conduction by the floating shield, which is a
good thermal conductor. Because of the temperature gradient existing between the two ends
of the floating shield and the heating of the tissue surrounding it, a tear drop ablation zone
can be created. Another reason for the deviation of the shape of the ablation zone from that
of the SAR pattern can be attributed to the heating of the floating shield due to the fields
of the two-wire shielded line. At the relatively high frequency of 10 GHz, the RF resistance
of the conductor is relatively high resulting in relatively high omhic loss and generation of
heat, which can raise the temperature of the tissue surrounding the shaft of the antenna. The
problem of shaft heating can be alleviated by using lower microwave frequencies or actively
cooling the floating shield. Such an active cooling mechanism can be implemented, e.g., by
using a custom-made floating shield that accommodates circulation of low-temperature gas
to cool its inner surface.
Table 6.1 summarizes the dimensions of the ablation zones in the x-z and y-z planes
obtained from these experiments. The dimensions of each ablation zone are measured as
the long-axis diameter and the short-axis diameter of the boundary of its coagulated zone
shown in white color in the photographs in Fig. 6.8. In overall, the dimensions of the
ablation zones in each plane are very close to each other. This illustrates that despite the
lack of rotational symmetry in dipole configuration with respect to the z-axis, the ablation
zones produced by the balanced antenna are almost rotationally symmetric and relatively
independent of the orientation of the dipole. We computed the aspect ratio, defined as the
90
ratio of the long-axis diameter to the short-axis diameter, for each ablation zone to quantify
the sphericity of the ablation zone. A larger aspect ratio represents a more elongated ablation
zone while an aspect ratio of 1 represents an ideally spherical shape. The average aspect
ratio of the ablation zones produced by the balanced antenna is 1.54 ± 0.06 in the x-z plane
and 1.50 ± 0.10 in the y-z plane as reported in Table 6.1.
Table 6.1: Dimensions of Eight Ablation Zones Generated with The Balanced Antenna
Cut plane
Size
AR
Photograph
4.0 × 2.5
1.60
Fig. 6.8(a)
4.0 × 2.6
1.53
Fig. 6.8(b)
3.6 × 2.4
1.50
Fig. 6.8(c)
3.7 × 2.5
1.48
Fig. 6.8(d)
4.0 × 2.5
1.60
Fig. 6.8(e)
3.8 × 2.5
1.52
Fig. 6.8(f)
4.0 × 2.6
1.54
Fig. 6.8(g)
3.5 × 2.5
1.40
Fig. 6.8(h)
[cm × cm]
x-z
y-z
AR: aspect ratio
6.4
Comparison with a Coax-fed Antenna
In this section, we compare the performance of the proposed non-coaxial-based balanced
antenna with that of a triaxial antenna. The triaxial antenna is chosen as a representative of
coax-fed interstitial antennas, since it has been well studied and used in an FDA-approved
commercial MWA system (Certus 140, NeuWave Medical Inc., Madison, Wisconsin). The
Certus 140 system has also been used in a number of clinical trials involving human subjects [103]- [104]. In the triaxial antenna design, a hollow metallic needle is used to adjust
the insertion depth of a monopole antenna in order to achieve good impedance matching
between the antenna and the feed line and to reduce outer-surface currents flowing on the
91
outer conductor of the coaxial feed line. The topology, principles of operation, and design
guidelines of this antenna are discussed in detail in [65] and will not be repeated here for
brevity. Following these design guidelines, we designed and fabricated a prototype of a
triaxial antenna operating at 10 GHz in ex vivo porcine muscle tissue.
The antenna was constructed based on a 50-Ω UT-085C semi-rigid coaxial cable from
Micro Coax with an outer diameter of 2.2 mm. The inner conductor was extended out of
the outer conductor of the coax to create the monopole antenna. A hollow copper tube, the
inner and outer diameters of which are 2.5 mm and 3.2 mm, was used to make the outermost
hollow needle of the triaxial design. The distance between the distal end of the hollow needle
and the distal end of the outer conductor of the coax is defined as the insertion depth of the
monopole antenna [65]. The length of the monopole and the insertion depth that provided
the best performance are respectively 15 mm and 6 mm.
The SAR pattern of the triaxial antenna at 10 GHz was calculated using full-wave EM
simulations in CST Microwave Studio. SAR values are reduced by 10 dB, 20 dB, and 30 dB
along the antenna shaft at a longitudinal distance of 20 mm, 23 mm, and 25 mm from the tip
of the monopole, respectively. The -30 dB SAR contour covers a volume with a maximum
lateral dimension of 16 mm and a maximum longitudinal dimension of 27 mm. The input
reflection coefficient of the fabricated prototype was measured while the antenna was inserted
in ex vivo pork loin. The measurement result agreed very well with the simulation result
and both show a good impedance match between the triaxial antenna and the 50-Ω feed line
at 10 GHz with the measured and simulated VSWR values of respectively 1.15 and 1.05.
We used the fabricated prototype of the triaxial antenna to conduct four ablation experiments in ex vivo pork loin with the same input power level (18 W) and duration (10
minutes) as the ones conducted with the balanced antenna. Fig. 6.9 shows photographs
of the ablation zones obtained from the four experiments. Each ablation zone consists of a
small region of charred tissue (black/dark brown), an ellipsoidal region of coagulated tissue
(white) surrounding the charred region, and an extended tail of coagulated tissue (white)
92
along the shaft of the outermost needle. We noticed that the excessive tail heating along the
antenna shaft was due to the lack of a shaft cooling mechanism that is incorporated in the
commercial version of the triaxial antenna [105]. However, this tail heating was not taken
into account in measuring the dimensions of the ablation zones obtained from these experiments. To determine the dimensions of each ablation zone, we estimated the intersection
point of the ellipsoidal part of the ablation zone and the shaft of the antenna and excluded
the tail beyond this point from the measurement of the maximum long-axis diameter of the
ablation zone. This scenario is shown in Fig. 6.9 where the fitted boundaries of the ablation
zones are highlighted with white dashed ellipses on the photographs. Doing this is justified
because the tail heating produced by the triaxial antenna can be reduced or eliminated by
using a shaft cooling mechanism [105]. Table 6.2 reports the dimensions of the ablation zones
produced by the triaxial antenna used in this study. In comparison with the ablation zones
created by the balanced antenna, the ones produced by the triaxial antenna have comparable short-axis diameters and larger long-axis diameters, and therefore larger aspect ratios.
Specifically, the average aspect ratio produced by the triaxial antenna is 1.98 ± 0.10, which
is significantly larger than that of the proposed balanced antenna (we emphasize that this
aspect ratio excludes the undesirable effects of tail heating). We can observe this trend by
comparing the photographs of the ablation zones of the balanced antenna (in Fig. 6.8) and
those of the triaxial antenna (in Fig. 6.9). As can be clearly observed, the ablation zones
produced by the balanced antenna are more spherical and compact. The comparative study
presented in this section shows that the proposed balanced antenna design has two distinct
advantages over the triaxial antenna. First, by eliminating the third coaxial conductor, the
proposed antenna can have a smaller shaft diameter than its triaxial counterpart. Secondly,
it provides a significantly more spherical ablation zone compared to what can be achieved
from a triaxial antenna under identical operating conditions.
93
(a)
(b)
(c)
(d)
Figure 6.9: Photographs of ablation zones created in ex vivo pork loin by applying 18 W to
the input of the triaxial antenna for 10 minutes. The photographs in (a), (b), (c) and (d)
are from four separate ablations. The white dashed ellipses show the fitted boundaries of
the ablation zones and are used for measuring the dimensions of the ablation zones.
Table 6.2: Dimensions of The Ablation Zones Generated with The Triaxial
Antenna
Size
AR
Photograph
5.0 × 2.5
2.00
Fig. 9(a)
4.7 × 2.5
1.88
Fig. 9(b)
4.5 × 2.3
1.96
Fig. 9(c)
5.2 × 2.5
2.08
Fig. 9(d)
[cm × cm]
AR: aspect ratio
6.5
Discussion
A non-coaxial-based balanced interstitial antenna was introduced for microwave ablation
applications. The antenna consists of a dipole fed with a balanced, shielded two-wire transmission line. The inherently balanced nature of the antenna and its feed line eliminates the
need for using any choking structure that is commonly used in coax-fed MWA antennas. This
allows for reducing the diameter of the antenna and hence, its invasiveness. We designed an
external transformer that interfaces between the coaxial output of the microwave source and
the balanced transmission line feeding the antenna. Upon integration with the transformer,
the antenna maintains an excellent impedance match and localized SAR patterns over a wide
bandwidth in the vicinity of the intended operating frequency of 10 GHz.
We fabricated a prototype of the antenna along with the balanced line and the exter-
94
nal transformer to conduct ablation experiments. The fabricated antenna demonstrated a
VSWR < 2 over the frequency range of 9-11 GHz. Several ex vivo ablation experiments in
porcine muscle were conducted with an input power of 18 W for a duration of 10 minutes. In
each case, localized ablation zones with aspect ratio ranging from 1.40 to 1.60 were obtained.
We compared the performance of the proposed balanced antenna and a triaxial antenna
designed for operation at 10 GHz in ex vivo porcine muscle. The triaxial antenna was chosen
as a representative of coax-fed antennas because it is widely studied and is used in an FDAapproved commercial microwave ablation system (Certus 140, NeuWave Medical Inc.). A
prototype of the triaxial antenna was fabricated and used to conduct ablation experiments
under the same conditions in which ablation experiments with the proposed balanced antenna
were conducted. The ablation zones obtained by the balanced antenna were found to be more
spherical with lower aspect ratios compared to those of the triaxial antenna while having
comparable lateral dimensions. The simulation and experimental results of the proposed
non-coaxial-based balanced antenna indicate its promise as a minimally-invasive interstitial
antenna for microwave ablation applications.
While the results of this study demonstrate the feasibility of using the proposed antenna
design concept for MWA applications, several future improvements are envisioned before
such antennas are used in clinical applications. First, we envision fabricating the antenna
using a technology that allows for it to be easily guided within the floating shield to the
desired ablation location and deployed there. Technologies similar to this exist and are
currently in use in RF ablation systems such as the LeVeen CoAccess Electrode system by
Boston Scientific [100]. Secondly, versions of this antenna can be developed that exploit a
metallic biopsy needle as the floating shield. Finally, the transition between the external
transformer and the input of the balanced line feeding the antenna should be optimized to
allow for a higher power handling capability.
95
Bibliography
[1] J. P. McGahan and G. D. Dodd, “Radiofrequency ablation of the liver: current status,”
Am. J. Roentgenol., vol. 176, no. 1, pp. 3-16, 2001.
[2] C. Boutros, P. Somasundar, S. Garrean, A. Saied, and N. J. Espat, “Microwave coagulation therapy for hepatic tumors: review of the literature and critical analysis,” Surg.
Oncol., vol. 19, pp. e22-e32, 2010.
[3] K. F. Chu and D. E. Dupuy, “Thermal ablation of tumors: biological mechanisms and
advances in therapy,” Nat. Rev. Cancer, vol. 14, no. 3, pp. 199-208, 2014.
[4] L. Tiong and G. J. Maddern, “Systematic review and meta-analysis of survival and
disease recurrence after radiofrequency ablation for hepatocellular carcinoma,” Br. J.
Surg., vol. 98, no. 9, pp. 1210-1224, 2011.
[5] M. R. Khan, R. T. Poon, K. K. Ng, A. C. Chan, J. Yuen, H. Tung, J. Tsang, and S. T.
Fan, “Comparison of percutaneous and surgical approaches for radiofrequency ablation
of small and medium hepatocellular carcinoma,” Arch. Surg., vol. 142, no. 12, pp. 11361143, 2007.
[6] M. Ahmed, C. L. Brace, F. T. Lee, and S. N. Goldberg, “Principles of and advances in
percutaneous ablation,” Radiology, vol. 258, no. 2, pp. 351-369, 2011.
96
[7] B. Dong, P. Liang, X. Yu, L. Su, D. Yu, Z. Cheng, and J. Zhang, “Percutaneous sonographically guided microwave coagulation therapy for hepatocellular carcinoma: results
in 234 patients,” Am. J. Roentgenol., vol. 180, no. 6, pp. 1547-1555, 2003.
[8] M. Lu, H. Xu, X. Xie, X. Yin, J. Chen, M. Kuang, Z. Xu, G. Liu, and Y. Zheng,
“Percutaneous microwave and radiofrequency ablation for hepatocellular carcinoma: a
retrospective comparative study,” J. Gastroenterol, vol. 40, pp. 1054-1060, 2005.
[9] H. Shiomi, S. Naka, K. Sato, K. Demura, K. Murakami, and T. Shimizu, “Thoracoscopyassisted magnetic resonance guided microwave coagulation therapy for hepatic tumors,”
Am. J. Surg., vol. 195, no. 6, pp. 854-860, 2008.
[10] P. Liang, Y. Wang, X. Yu, and B. Dong, “Malignant liver tumors: treatment with percutaneous microwave ablation—Complications among cohort of 1136 patients,” Radiology,
vol. 251, no. 3, pp. 933-940, 2009.
[11] A. R. Gillams and W. R. Lees, “Five-year survival following radiofrequency ablation of
small, solitary, hepatic colorectal metastases,” J. Vasc. Interv. Radiol., vol. 19, no. 5, pp.
712-717, 2008.
[12] C. Christophi, M. Nikfarjam, C. Malcontenti-Wilson, and V. Muralidharan, “Long-term
survival of patients with unresectable colorectal liver metastases treated by percutaneous
interstitial laser thermotherapy,” World J. Surg., vol. 28, n. 10, pp. 987-994, 2004.
[13] T. Gulesserian, A. H. Mahnken, R. Schernthaner, M. Memarsadeghi, M. Weber, A.
Tacke, and J. Kettenbach, “Comparison of expandable electrodes in percutaneous radiofrequency ablation of renal cell carcinoma,” Eur. J. Radiol., vol. 59, no. 2, pp. 133-139,
2006.
97
[14] F. J. Wolf, D. J. Grand, J. T. Machan, T. A. Dipetrillo, W. W. Mayo-Smith, and D. E.
Dupuy, “Microwave ablation of lung malignancies: effectiveness, CT findings, and safety
in 50 patients,” Radiology, vol. 247, no. 3, pp. 871-879, 2008.
[15] T. Hiraki, H. Gobara, T. Iishi, Y. Sano, T. Iguchi, H. Fujiwara, N. Tajiri, J. Sakurai,
and H. Date, “Percutaneous radiofrequency ablation for clinical stage I non-small cell
lung cancer: results in 20 nonsurgical candidates,” J. Thorac. Cardiov. Sur., vol. 134,
no. 5, pp. 1306-1312, 2007.
[16] R. Lencioni, L. Crocetti, R. Cioni, R. Suh, D. Glenn, D. Regge, T. Helmberger, and A.
R. Gillams, “Response to radiofrequency ablation of pulmonary tumours: a prospective,
intention-to-treat, multicentre clinical trial (the RAPTURE study),” Lancet Oncol., vol.
9, no. 7, pp. 621-628, 2008.
[17] M. R. Callstrom, J. W. Charboneau, M. P. Goetz, J. Rubin, T. D. Atwell, M. A. Farrell,
T. J. Welch, and T. P. Maus, “Image-guided ablation of painful metastatic bone tumors:
a new and effective approach to a difficult problem,” Skeletal Radiol., vol. 35, no. 1, pp.
1-15, 2006.
[18] C. L. Brace, “Microwave tissue ablation: Biophysics, technology and applications,” Crit.
Rev. Biomed. Eng., vol. 38, no. 1, pp. 65-78, 2010.
[19] M. G. Skinner, M. N. Iizuka, M. C. Kolios, and M. D. Sherar, “A theoretical comparison of energy sources–microwave, ultrasound, and laser–for interstitial thermal therapy,”
Phys. Med. Biol., vol. 43, no. 12, pp. 3535-3547, 1998.
[20] C. L. Brace, “Radiofrequency and microwave ablation of the liver, lung, kidney, and
bone: what are the difference?” Curr. Probl. Diagn. Radiol., vol. 38, no. 3, pp. 135-143,
2009.
98
[21] J. L. Hinshaw, M. G. Lubner, T. J. Ziemlewicz, F. T. Lee, and C. L. Brace, “Percutaneous tumor ablation tools: microwave, radiofrequency, or cryoabaltion–what should
you use and why?” Radiographics, vol. 34, no. 5, pp. 1344-1362, 2014.
[22] A. Andreano and C. L. Brace, “A comparison of direct heating during radiofrequency
and microwave ablation in ex vivo liver,” Cardiovasc. Intervent. Radiol., vol. 36, pp.
505-511, 2013.
[23] A. Andreano, Y. Huang, M. F. Meloni, F. T. Lee, and C. Brace, “Microwave create
larger ablations than radiofrequency when controlled for power in ex vivo tissue,” Med.
Phys., vol. 37, no. 6, pp. 2967-2973.
[24] C. L. Brace, J. L. Hinshaw, P. F. Laeseke, L. A. Sampson, and F. T. Lee, “Pulmonary
thermal ablation: comparison of radiofrequency and microwave devices by using gross
pathologic and CT findings in a swine model,” Radiology, vol. 251, no. 3, pp. 705-711,
2009.
[25] W. Fan, X. Li, L. Zhang, H. Jiang, and J. Zhang, “Comparison of microwave ablation
and multipolar radiofrequency ablation in vivo using two internally cooled probes,” AJR
Am. J. Roentgenol., vol. 198, no. 1, pp. W46-W50, 2012.
[26] P. F. Laeseke, F. T. Lee, L. A. Sampson, D. W. van der Weide, and C. L. Brace,
“Microwave ablation versus radiofrequency ablation in the kidney: high-power triaxial
antennas create larger ablation zones than similarly sized internally cooled electrodes,”
J. Vasc. Interv. Radiol., vol. 20, no. 9, pp. 1224-1229, 2009.
[27] T. Tanaka, S. Westphal, P. Isfort, T. Braunschweig, T. Penzkofer, P. Bruners, K.
Kichikawa, T. Schmitz-Rode, and A. H. Mahnken, “Microwave ablation compared with
radiofrequency ablation for breast tissue in an ex vivo bovine udder model,” Cardiovasc.
Intervent. Radiol., vol. 35, no. 4, pp. 914-920, 2012.
99
[28] M. G. Lubner, C. L. Brace, J. L. Hinshaw, and F. T. Lee, “Microwave tumor ablation:
mechanism of action, clinical results and devices,” J. Vasc. Interv. Radiol., vol. 21, pp.
S192-S203, 2010.
[29] C. Gabriel, Dielectric properties of biological materials. Bioengineering and Biophysical
Aspects of Electromagnetic Fields 1, CRC Press, pp. 87-136, 2006.
[30] S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of biological tissues:
III. Parametric models for the dielectric spectrum of tissues,” Phys. Med. Biol., vol. 41,
pp. 2271-2293, 1996.
[31] K. R. Foster, J. L. Schepps, R. D. Stoy, and H. P. Schwan, “Dielectric properties of
brain tissue between 0.01 and 10 GHz,” Phys. Med. Biol., vol. 24, no. 6, pp. 1177-1187,
1979.
[32] L. Chin and M. Sherar, “Changes in the dielectric properties of rat prostate ex vivo at
915 MHz during heating,” Int. J. Hyperther., vol. 20, no. 5, pp. 517-527, 2004.
[33] M. Lazebnik, M. C. Converse, J. H. Booske, and S. C. Hagness, “Ultrawideband
temperature-dependent dielectric properties of animal liver tissue in the microwave frequency range,” Phys. Med. Biol., vol. 51, pp. 1941-1955, 2006.
[34] L. Chin and M. Sherar, “Changes in dielectric properties of ex vivo bovine liver at 915
MHz during heating,” Phys. Med. Biol., vol. 46, pp. 197-211, 2001.
[35] P. R. Satuffer, F. Rossetto, M. Prakash, D. G. Neuman, and T. Lee, “Phantom and
animal tissues for modelling the electrical properties of human liver,” Int. J. Hyperther.,
vol. 19, no. 1, pp. 89-101, 2003.
[36] Z. Ji and C. L. Brace, “Expanded modeling of temperature-dependent dielectric properties for microwave thermal ablation,” Phys. Med. Biol., vol. 56, pp. 5249-5264, 2011.
100
[37] V. Lopresto, R. Pinto, G. A. Lovisolo, and M. Cavagnaro, “Changes in the dielectric
properties of ex vivo bovine liver during microwave thermal ablation at 2.45 GHz,” Phys.
Med. Biol, vol. 57, pp. 2309-2327, 2012.
[38] C. L. Brace, “Temperature-dependent dielectric properties of liver tissue measured during thermal ablation: toward an improved numerical model,” Conf. Proc. IEEE Eng.
Med. Biol. Soc., Vancouver, British Columbia, Canada, 2008, pp. 230-233.
[39] T. R. Larson, D. G. Bostwick, and A. Corica, “Temperature-correlated histo pathologic
changes following microwave thermoablation of obstructive tissue in patients with benign
prostatic hyperplasia,” Urology, vol. 47, no. 4, pp. 463-469, 1996.
[40] S. Thomsen, “Pathologic analysis of photothermal and photomechanical effects of lasertissue interactions,” Photochem. Photobiol., vol. 53, no. 6, pp. 825-835, 1991.
[41] M. Nikfarjam, V. Muralidharan, and C. Christophi, “Mechanism of focal heat destruction of liver tumors,” J. Surg. Res., vol. 127, no. 2, pp. 208-223, 2005.
[42] S. N. Goldberg, G. S. Gazelle, and P. R. Mueller, “Thermal ablation therapy for focal malignancy: a unified approach to underlying principles, techniques, and diagnostic
imaging guidance,” AJR Am. J. Roentgenol., vol. 174, no. 2, pp. 323-331, 2000.
[43] C. Christophi, A. Winkworth, V. Muralihdaran, and P. Evans, “The treatment of malignancy by hyperthermia,” Surg. Oncol., vol. 7, pp. 83-90, 1999.
[44] L. F. Fajardo, B. Egbert, J. Marmor, and G. M. Hahn, “Effects of hyperthermia in a
malignant tumor,” Cancer, vol. 45, no. 3, pp. 613-623, 1980.
[45] H. M. Sweetland, A. Wyman, and K. Rogers, “Evaluation of the effect on normal liver
of interstitial laser hyperthermia using artificial sapphire probes,” Lasers Med. Sci., vol.
8, pp. 99-105, 1993.
101
[46] W. J. Wiersinga, M. C. Jansen, I. H. Straatsburg, P. H. Davids, J. M. Klaase, D. J.
Gouma, and T. M. van Gulik, “Lesion progression with time and the effect of vascular
occlusion following radiofrequency ablation of the liver,” Br. J. Surg., vol. 90, no. 3, pp.
306-312, 2003.
[47] T. Ohno, K. Kawano, A. Sasaki, M. Aramaki, T. Yoshida, and S. Kitano, “Expansion
of an ablated site and induction of apoptosis after microwave coagulation therapy in rat
liver,” J. Hepatobiliary Pancreat. Surg., vol. 8, no. 4, pp. 360-366, 2001.
[48] S. L. Sankari, K. M. K. Masthan, N. A. Babu, T. Bhattacharjee, and M. Elumalai,
“Apoptosis in cancer—an update,” Asian Pac. J. Cancer Prev., vol. 10, pp. 4873-4878,
2012.
[49] K. Ahmed, Y. Tabuchi, and T. Kondo, “Hyperthermia: an effective strategy to induce
apoptosis in cancer cells,” Apoptosis, vol. 20, no. 11, pp. 1411-1419, 2015.
[50] G. Heuff, H. S. A. Oldenburg, H. Boutkan, J. J. Visser, R. H. J. Beelen, N. Van Rooijen,
C. D. Dijkstra, and S. Meyer, “Enhanced tumor growth in the rat liver after selective
elimination of Kupffer cells,” Cancer Immunol. Immun., vol. 37, no. 2, pp. 125-130, 1993.
[51] N. Watanabe, Y. Niitsu, H. Umeno, H. Kuriyama, H. Neda, N. Yamauchi, M. Maeda,
and I. Urushizaki, “Toxic effect of tumor necrosis factor on tumor vasculature in mice,”
Cancer Res., vol. 48, no. 8, pp. 2179-2183, 1988.
[52] J. Overgaard, “Ultrastructure of a mouse mammary carcinoma exposed to hyperthermia
in vivo,” Cancer Res., vol. 36, no. 3, pp. 983-995, 1976.
[53] J. A. Dickson and S. A. Shah, “Hyperthermia: the immune response and tumor metastasis,” Natl. Cancer Inst. Monoqr., vol. 61, pp. 183-192, 1982.
102
[54] P. R. Geraghty, S. T. Kee, G. McFarlane, M. K. Razavi, D. Y. Sze, and M. D. Dake,
“CT-guided transthoracic needle aspiration biopsy of pulmonary nodules: needle size
and pneumothorax rate,” Radiology, vol. 229, no. 2, pp. 475-481, 2003.
[55] J. M. Bertram, D. Yang, M. C. Converse, J. G. Webster, and D. M. Mahvi, “A review of
coaxial-based interstitial antennas for hepatic microwave ablation,” Crit. Rev. Biomed.
Eng., vol. 34, no. 3, pp. 187-213, 2006.
[56] W. Hurter, F. Reinbold, and W. J. Lorenz, “A dipole antenna for interstitial microwave
hyperthermia,” IEEE Trans. Microw. Theory, vol. 39, no. 6, pp. 1048-1054, 1991.
[57] R. D. Nevels, G. D. Arndt, G. W. Taffoul, J. R. Carl, and A. Pacifico, “Microwave
catheter design,” IEEE Trans. Biomed. Eng., vol. 45, no. 7, pp. 885-890, 1998.
[58] J. C. Lin and Y. Wang, “The cap-choke catheter antenna for microwave ablation treatment,” IEEE Trans. Biomed. Eng., vol. 43, pp. 657-660, Jun. 1996.
[59] S. Pisa, M. Cavagnaro, P. Bernardi, and J. C. Lin, “A 915-MHz antenna for microwave
thermal ablation treatment: Physical design, computer modeling and experimental measurement,” IEEE Trans. Biomed. Eng., vol. 48, no. 5, pp. 599-601, May 2001.
[60] K. Saito, S. Hosaka, S.-Y. Okabe, H. Yoshimura, and K. Ito, “A proposition on improvement of a heating pattern of an antenna for microwave coagulation therapy: Introduction
of a coaxial-dipole antenna,” Elec. Commun. Jpn., vol. 86, no. 1, pp. 16-23, 2003.
[61] G. Schaller, J. Erb, and R. Engelbrecht, “Field simulation of dipole antennas for interstitial microwave hyperthermia,” IEEE Trans. Microw. Theory Tech., vol. 44, pp.887-895,
1996.
[62] H. M. Chiu, A. S. Mohan, A. R. Weily, D. J. Guy, and D. L. Ross, “Analysis of a novel
expanded tip wire (ETW) antenna for microwave ablation of cardiac arrhythmias,” IEEE
Trans. Biomed. Eng., vol. 50, pp. 890-899, 2003.
103
[63] D. Yang, J. M. Bertram, M. C. Converse, A. P. O’Rourke, J. G. Webster, S. C. Hagness,
J. A. Will, and D. M. Mahvi, “A floating sleeve antenna yields localized hepatic microwave
ablation,” IEEE Trans. Biomed. Eng., vol. 53, no. 3, pp. 533-537, Mar. 2006.
[64] K. Ito, M. Hyodo, M. Shimura, and H. Kasai, “Thin applicator having coaxial ring
slots for interstitial microwave hyperthermia,” IEEE AP-S Antennas Propagation Soc.
Int. Symp. Dig., vol. 3, pp. 1233-1236, 1990.
[65] C. L. Brace, P. F. Laeseke, and D. W. van der Weide, “Microwave ablation with a
triaxial antenna: Results in ex vivo bovine liver,” IEEE Trans. Microw. Theory Tech.,
vol. 53, pp. 215-220, 2005.
[66] K. Saito, H. Yoshimura, K. Ito, Y. Aoyagi, and H. Horita, “Clinical trials of interstitial
microwave hyperthermia by use of coaxial-slot antenna with two slots,” IEEE Trans.
Microw. Theory Tech., vol. 52, pp. 1987-1991, 2004.
[67] C. L. Brace, “Dual-slot antennas for microwave tissue heating: Parametric design analysis and experimental validation,” Med. Phys., vol. 38, pp. 4231-4240, 2011.
[68] M. Kuang, M. D. Lu, X. Y. Xie, H. X. Xu, L. Q. Mo, G. J. Liu, Z. F. Xu, Y. L.
Zheng, and J. Y. Liang, “Liver cancer: Increased microwave delivery to ablation zone
with cooled-shaft antenna—Experimental and clinical studies,” Radiology, vol. 242, no.
3, pp. 914-924, 2007.
[69] Y. Wang, Y. Sun, L. Feng, Y. Gao, X. Ni, and P. Liang, “Internally cooled antenna for
microwave ablation: results in ex vivo and in vivo porcine livers,” Eur. J. Radiol., vol.
67, no. 2, pp. 357-361, 2008.
[70] I. Longo, G. B. Gentili, M. Cerretelli, and N. Tosoratti, “A coaxial antenna with miniaturized choke for minimally invasive interstitial heating,” IEEE Trans. Biomed. Eng.,
vol. 50, no. 1, pp. 82-88, Jan. 2003.
104
[71] A. U. Hines-Peralta, N. Pirani, P. Clegg, N. Cronin, T. P. Ryan, Z. Liu, and S. N.
Goldberg, “Microwave ablation: Results with a 2.45-GHz applicator in ex vivo bovine
and in vivo porcine liver,” Radiology, vol. 239, no. 1, pp. 94-102, 2006.
[72] S. Labonte, A. Blais, S. R. Legault, H. O. Ali, and L. Roy, “Monopole antennas for
microwave catheter ablation,” IEEE Trans. Microw. Theory Tech., vol. 44, No. 10, pp.
1832-1840, Oct. 1996.
[73] G. Carrafiello, A. M. Ierardi, F. Piacentino, N. Lucchina, G. Dionigi, S. Cufarri, and
C. Fugazzola, “Microwave ablation with percutaneous approach for the treatment of
pancreatic adenocarcinoma,” Cardiovasc. Intervent. Radiol., vol. 35, no. 2, pp. 439-442,
Apr. 2012.
[74] M. Cheng, M. A. Haider, M. J. Dill-Macky, J. M. Sweet, J. Trachtenberg, and M.
R. Gertner, “MRI and contrast-enhanced ultrasound monitoring of prostate microwave
focal thermal therapy: An in vivo canine study,” J. Magn. Reson. Im., vol. 28, no. 1, pp.
136-143, Jul. 2008.
[75] W. Zhou, X. Zha, X. Liu, Q. Ding, L. Chen, Y. Ni, Y. Zhang, Y. Xu, L. Chen, Y. Zhao,
and S. Wang, “US-guided percutaneous microwave coagulation of small breast cancers:
A clinical study,” Radiology, vol. 263, no. 2, pp. 364-373, May 2012.
[76] Y. Kanaoka, C. Yoshida, T. Fukuda, K. Kajitani, and O. Ishiko, “Transcervical microwave myolysis for uterine myomas assisted by transvaginal ultrasonic guidance,” J.
Obset. Gynaocol. Res., vol. 35, no. 1, pp. 145-151, Feb. 2009.
[77] D. A. Hodgson, I. B. Feldberg, N. Sharp, N. Cronin, M. Evans, and L. Hirschowitz,
“Microwave endometrial ablation: Development, clinical trials and outcomes at three
years,” Br. J. Obstet Gynaec, vol. 106, pp. 684-694, Jul. 1999.
105
[78] S. A. Jack and K. G. Cooper, “Microwave endometrial ablation: an overview,” Reviews
in Gynaecological Practice, vol. 5, pp. 32-38, 2005.
[79] R. P. Jones, N. R. Kitteringham, M. Terlizzo, C. Hancock. D. Dunne, S. W. Fenwick,
G. J. Poston, P. Ghaneh, and H. Z. Malik, “Microwave ablation of ex vivo human liver
and colorectal liver metastates with a novel 14.5 GHz generator,” Int. J. Hyperther., vol.
28, no. 1, pp. 43-54, 2012.
[80] C. P. Hancock, N. Dharmasiri, M. White, and A. M. Goodman, “The design and development of an integrated multi-functional microwave antenna structure for biological
applications,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2230-2241, May
2013.
[81] J. Yoon, J. Cho, N. Kim, D. Kim, E. Lee, C. Cheion, and Y. Kwon, “High-frequency
microwave ablation method for enhanced cancer treatment with minimized collateral
damage,” Int. J. Cancer, vol. 129, pp. 1970-1978, 2011.
[82] J. Vrba, C. Fraconi, and M. Lapes, “Theoretical limits for the penetration depth of
intracavitary applicators,” Int. J. Hyperther., vol. 12, no. 6, pp. 737-742, 1996.
[83] F. A. Duck, Physical Properties of Tissue: A Comprehensive Reference Book. London,
Academic Press, 1990.
[84] J. Chiang, P. Weng, and C. L. Brace, “Computational modeling of microwave tumour
ablation,” Int. J. Hyperther., vol. 29, no. 4, pp. 308-317, 2013.
[85] P. Prakash, M. C. Converse, J. G. Webster, and D. M. Mahvi, “An optimal sliding
choke antenna for hepatic microwave ablation,” IEEE Trans. Biomed. Eng., vol. 56, no.
10, pp. 2470-2476, Oct. 2009.
106
[86] M. Cavagnaro, C. Amabile, P. Bernadi, S. Pisa, and N. Tosoratti, “A minimally invasive antenna for microwave abaltion therapies: Design, performances, and experimental
assessment,” IEEE Trans. Biomed. Eng., vol. 58, no. 4, pp. 949-959, Apr. 2011.
[87] H. Luyen, F. Gao, S. C. Hagness, and N. Behdad, “Microwave ablation at 10.0 GHz
achieves comparable ablation zones to 1.9 GHz in ex vivo bovine liver,” IEEE Trans.
Biomed. Eng., vol. 61, no. 6, pp. 1702-1710, 2014.
[88] Harbour Industries, “Power handling capability of coaxial cable,” [Online]. Available:
http://harbourind.com/harbourdev/images/technical_pdfs/Power_Handling.
pdf. [Accessed Sep. 9, 2016].
[89] M. A. El-Brawany, D. K. Nassiri, G. Terhaar, A. Shaw, I. Rivens, and K. Lozhken,
“Measurement of thermal and ultrasonic properties of some biological tissues,” J. Med.
Eng. Technol., vol. 33, no. 3, pp. 249-256, 2009.
[90] H. Luyen, S. C. Hagness, and N. Behdad, “A balun-free helical antenna for minimally
invasive microwave ablation,” IEEE Trans. Antennas Propag., vol. 63, no. 3, pp. 959-965,
2015.
[91] H. Luyen, S. C. Hagness, and N. Behdad, “A minimally invasive, coax-fed microwave
ablation antenna with a tapered-slot balun,” IEEE Int. Symp. Antennas and Propagation,
Fajardo, 2016, pp. 1163-1164.
[92] J. W. Duncan and V. P. Minerva, “100:1 bandwidth balun transformer,” Proceedings of
the IRE, vol. 48, no. 2, pp. 156-164, 1960.
[93] J. G. Choi, S. H. Yi, and K. H. Kim, “Development of a novel tapered balun coupler,”
IEEE Power Modulator Symp., 2004, pp. 493-496.
[94] R. W. Klopfenstein, “A transmission line taper of improved design,” P. IRE, vol. 44.
no. 1, pp. 31-35, 1956.
107
[95] D. Yang, M. C. Converse, D. M. Mahvi, and J. G. Webster, “Expanding the bioheat equation to include tissue internal water evaporation during heating,” IEEE Trans.
Biomed. Eng., vol. 54, no. 8, pp. 1382-1388, 2007.
[96] S. R. Guntur, K. I. Lee, D. Paeng, A. J. Coleman, and M. J. Choi, “Temperaturedependent thermal properties of ex vivo liver undergoing thermal ablation,” Ultrasound
Med. Biol., vol. 39, no. 10, pp. 1771-1784, 2013.
[97] B. T. McWilliams, E. E. Schnell, S. Curto, T. M. Fahrbach, and P. Prakash, “A directional interstitial antenna for microwave tissue ablation: theoretical and experimental
investigation,” IEEE Trans. Biomed. Eng., vol. 62, no. 9, pp. 2144-2150, 2015.
[98] D. Haemmerich, F. T. Lee, D. J. Schutt, L. A. Sampson, J. G. Webster, J. P. Fine,
and D. M. Mahvi, “Large-volume radiofrequency ablation of ex vivo bovine liver with
multiple cooled cluster electrodes,” Radiology, vol. 234. no. 2, pp. 563-568, 2005.
[99] H. T. Luyen, Y. Mohtashami, J. Sawicki, J. D. Shea, S. C. Hagness, and N. Behdad,
“Recent advances in designing balun-free interstitial antennas for minimally-invasive microwave ablation,” IEEE Int. Symp. Antennas and Propagation, Hobart, Tasmania, Australia, 2015, pp. 1-4.
[100] S. Mulier, Y. Miao, P. Mulier, B. Dupas, P. Pereira, T. de Baere, R. Lencioni, R.
Leveillee, G. Marchal, L. Michel, and Y. Ni, “Electrodes and multiple electrode systems
for radiofrequency ablation: a proposal for updated terminology,” Eur. Radiol., vol. 15,
no. 4, pp. 798-808, 2005.
[101] R. Li, T. Wu, B. Pan, K. Lim, J. Laskar, and M. M. Tentzeris, “Equivalent-circuit
analysis of a broadband printed dipole with adjusted integrated balun and an array for
base station applications,” IEEE Trans. Antennas Propag., vol. 57, no. 7, pp. 2180-2184,
2009.
108
[102] R. Lien, M. C. Hunt, S. Anderson, D. H. Kropf, T. M. Loughin, M. E. Dikeman, and
J. Velazco, “Effects of endpoint temperature on the internal color of pork loin chops of
different quality,” J. Food Sci., vol. 67, no. 3, pp. 1007-1010, 2002.
[103] T. J. Ziemlewicz, S. A. Wells, M. A. Lubner, A. I. Musat, J. L. Hinshaw, A. R.
Cohn, and F. T. Lee Jr., “Microwave ablation of giant hepatic cavernous hemangiomas,”
Cardiovasc. Inter. Rad., vol. 37, pp. 1299-1305, 2014.
[104] T. J. Ziemlewicz, J. L. Hinshaw, M. G. Lubner, C. L. Brace, M. L. Alexander, P.
Agarwal, and F. T. Lee, “Percutaneous microwave ablation of hepatocellular with a gascooled system: initial clinical results with 107 tumors,” J. Vasc. Interv. Radiol., vol. 26,
pp. 62-68, 2015.
[105] M. G. Lubner, C. L. Brace, J. L. Hinshaw, and F. T. Lee Jr., “Microwave tumor
ablation: mechanism of action, clinical results, and devices,” J. Vasc. Interv. Radiol.,
vol. 21, pp. S192-S203, 2010.
Документ
Категория
Без категории
Просмотров
0
Размер файла
8 749 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа