I. INTRODUCTION
Hyperthermia (HT) is the elevation of tumor temperatures to a clinically relevant temperature range of 40–43°C. HT is an adjuvant therapy, used to enhance the effects of radiotherapy (RT) and chemotherapy. The widely accepted thermal dose is 1 h at 43°C [Reference Sapareto and Dewey1], though this temperature is often not achieved. The clinical outcome is strongly correlated with the thermal dose [Reference Kapp and Cox2]; therefore, realizing a sufficiently high and homogeneous tumor temperature is important. The added value of HT has been demonstrated for many tumor types [Reference Datta3, Reference Cihoric4]. HT is a standard treatment for breast cancer, the most common cancer in women. Local recurrences are a significant problem, occurring in ~10% of the patients. Treatment with RT combined with HT has proven effective, e.g. in the combined analysis of five randomized trials, which showed a complete response rate of 59% for RT + HT versus 41% for RT alone [Reference Vernon5]. Other randomized and large cohort studies have confirmed the significant radiosensitizing effect of HT in the treatment of recurrent breast cancer [Reference Jones6–Reference Datta, Puric, Klingbiel, Gomez and Bodis12]. Most local breast cancer recurrences are located a few centimeters below the skin and are routinely treated with many different systems. These include electromagnetic (EM) antennas operating at 434 and 915 MHz yielding tumor heating up to 4 cm depth, e.g. the 434 MHz Contact Flexible Microstrip Applicator (CFMA) and the 434 MHz ALBA ON4000 [Reference Gelvich and Mazokhin13–Reference Lamaitre, van Dijk, Gelvich, Wiersma and Schneider17] and the 915 MHz BSD-500 [Reference Datta18]. Non-EM radiation-based systems used for heating recurrent breast cancer utilize absorption of ultrasound [Reference Moros, Penagaricano, Novak, Straube and Myerson19] and infrared radiation [Reference Notter, Piazena and Vaupel20].
Tumors in intact breast require a different approach because of the shape of the breast and the different heating depth. Fenn et al. [Reference Fenn, Wolf and Fogle21] developed a double antenna system, which required breast compression due to the limited penetration depth of the 915 MHz waveguides used. Decreasing the operating frequency and/or increasing the number of antennas are logical steps to achieving tumor heating in intact breast where tumors generally extend deeper than 4 cm. A similar approach is also used for heating deep-seated tumors (e.g. cervix or bladder) using phased array systems such as the BSD-2000 system [Reference Turner22, Reference Turner, Tumeh and Schaefermeyer23] and the four-antenna array developed at the Academic Medical Center (AMC) [Reference van Dijk, Schneider, van Os, Blank and Gonzalez24, Reference Crezee25]. Novel 70 MHz CFMAs suitable to treat breast tumors infiltrating more than 4 cm were designed, which can also be used in multi-antenna arrays [Reference van Wieringen26, Reference Kosterev, Kramer-Ageev, Mazokhin, van Rhoon and Crezee27]. Wu et al. [Reference Wu, McGough, Arabe and Samulski28] developed a dedicated phased array of four end-loaded dipole antennas operating at 140 MHz for treatment of breast cancer where phase and amplitude steering is used to achieve preferential tumor heating. Stang et al. employed a large two-dimensional array of 915 MHz microstrip patch antennas [Reference Stang, Haynes, Carson and Moghaddam29]. Most of these EM-based breast HT systems did not make it into routine clinical use. Ultrasound-phased arrays for heating lesions in the breast were also developed [Reference Hynynen30, Reference Malinen, Huttunen, Hynynen and Kaipio31], which have the advantage that relatively small focal volumes can be achieved in comparison with EM systems. These systems are in clinical use, but mainly for rapid high-temperature (T > 50°C) ablation instead of moderate heating of the lesion at 43°C for 1 h to enhance the effectiveness of RT.
The purpose of this research is to design, construct, and clinically apply a dedicated HT system for treating tumors in intact breast with robust features to facilitate clinical use. This system is then evaluated in terms of:
-
(1) waveguide characteristics for a setup with a tissue equivalent phantom;
-
(2) ability to achieve an adequate and homogeneous thermal dose in the tumor >40°C;
-
(3) absence of treatment limiting normal tissue hot spots; and
-
(4) patient comfort and clinical feasibility.
II. METHODS
The treatment system consists of a treatment bed fitted with a 50 cm × 40 cm × 16 cm temperature-controlled open water bolus and antenna (Fig. 1). The water bolus serves a dual purpose; it is a matching layer between antenna and patient as well as providing a controlled thermal boundary condition to achieve the desired skin temperature. The concept of an open water bolus was chosen to provide optimal contact, similar to the coaxial transverse electric and magnetic (TEM) regional HT system [Reference De Leeuw and Lagendijk32]. The patient lies in prone position on a water-filled mattress with the breast immersed in the water positioned in front of a 70 MHz waveguide placed at the bottom of the bolus. The ¼λ waveguide is 12 cm deep and has an aperture size of 20 cm × 34 cm, this waveguide is also used in the AMC-phased array system consisting of up to eight 70 MHz waveguides used for heating deep-seated tumors, including cervix, rectum, and bladder [Reference Crezee25]. The water-filled waveguide operates in the TE10 mode with the dominant E-field direction parallel to the skin surface and perpendicular to the long side of the aperture with a cut-off frequency
$f_c = 1/2a\sqrt {\mu \epsilon} $
≈ 50 MHz for a = 0.34 m−1. The EM field is absorbed in the tissue with a specific absorption rate (SAR):
$${\rm SAR} = \displaystyle{\sigma \over {2\rho}} \left \vert E \right \vert ^2 \;({\rm W/kg}),$$
with σ the conductivity and ρ the density of tissue.
Fig. 1. Treatment setup: patient in prone position, breast immersed in open water bolus. 70 MHz waveguide placed at the bottom.
The waveguide is connected to an analogous amplifier (Henry Radio, Los Angeles, California, USA) providing 500 W output power. A double-slug tuner between the amplifier and antenna is used for tuning. The variable impedance of the applicator in combination with the patient has to be tuned with the fixed 50 Ω amplifier output.
A) Waveguide characteristics
The emitted E-field pattern was measured in the setup shown in Fig. 2 using an elliptical phantom filled with a muscle-equivalent 3 g/l saline solution. This solution has a relative dielectric permittivity of 77 and a conductivity of 0.51 S/m at 70 MHz and a water temperature of 22°C. The phantom has a cross-section of 24 cm × 36 cm and a length of 115 cm. The wall was made of polyvinyl chloride with a thickness of 2 mm. The waveguide is placed on top of the phantom; a water bolus containing de-ionized water is placed between waveguide and phantom.
Fig. 2. Setup to determine E-field patterns of the 70 MHz waveguide in the transversal (XY) and sagittal (ZY) midplanes in the liquid saline phantom.
The E-field was measured using an electric field vector probe, connected to a measurement system. The output was sent to a DAQ-unit (Labjack U12, labjack cooperation, Lakewood, USA) connected to a PC. An in-house developed MATLAB application interprets and stores the signals. The probe is mounted on an in-house developed computer-controlled x–y scanner. E-field measurements were performed in the transversal XY (z = 0) and sagittal YZ (x = 0) midplane (Fig. 2). Measurements were normalized in the center of the phantom and center of the applicator. The amplitude of the E-field was defined to be 100% at this point.
Additionally, the reflection coefficient (S11), the Voltage Standing Wave Ratio and the relative reflected power were determined for this phantom setup.
B) Tumor temperatures
Patient selection for the clinical feasibility test of this system was based on tumor depth and whether the intact breast was sufficiently large to be immersed in the open water bolus. All treatment histologies were accepted. HT was applied using this system for a total of six patients. Patient characteristics are listed in Table 1. Four patients had adenocarcinoma, two had melanoma. Median age was 59 (42–75). RT dose was 32 Gy given in eight fractions of 4 Gy twice a week.
Table 1. Patient characteristics and treatment parameters.
HT was given once a week for 4 weeks in combination with a RT session, commencing within 1 h after RT. The protocol allowed for a maximum of 30 min of preheating in order to achieve the desired temperature level of at least 40°C and this steady state then continued for 60 min. The maximum duration of treatment was thus 90 min.
Treatment was performed using the setup shown in Figs 1 and 3. The temperature of the water in the bolus starts at 39°C and is gradually increased to 42°C at steady state to permit the skin to gradually get accustomed to higher temperatures. Tap water is used and the water is not recirculated but continuously refreshed from the tap. Waveguide was positioned facing the breast containing a tumor. Both breasts are immersed in the water, but the healthy breast is enveloped in fabric creating 1 cm thick air layer isolating it from the EM field to prevent heating of this non-target breast (Fig. 3).
Fig. 3. Waveguide faces breast with tumor, healthy breast is isolated.
Tissue temperatures are recorded using 14-sensor copper-constantan thermocouple probes with a diameter of 0.9 mm (Ella CS, Hradec Kralove, Czech Republic). Temperatures are recorded every 30 s, switching off power for 5 s to prevent electronic disturbance during measurement [Reference de Leeuw, Crezee and Lagendijk33]. Four probes are placed on the skin of the breast. In addition, two invasive probes are placed in lossless polyethylene catheters (diameter 1.3 mm), which are inserted in lateral and in cranial–caudal direction through the breast to obtain representative temperature profiles in the target region. The resulting tumor temperature distribution is characterized using T10, T50, and T90, the temperature achieved in 10, 50, and 90% of the treatment volume, respectively. The temperature parameters T10, T50, and T90 are averaged over the steady-state treatment period of 60 min per session and over all four sessions per patient.
C) Normal tissue hot spots
Incidence of normal tissue hot spots was determined from the recorded T10 values and by recording complaints about hot spots by the patients. Normal tissue temperatures exceeding 44°C were considered hot spots to be avoided associated with the risk of side effects.
D) Patient comfort
Patient comfort and overall clinical feasibility were determined based on the aforementioned temperature parameters (tumor temperature and normal tissue hot spots) and the ability of patients to complete treatment as scheduled. The opinion of the patients on treatment comfort was also recorded after treatment.
III. RESULTS
A) Waveguide characteristics
The measured E-field patterns are shown in Fig. 4 and show that the effective field size is about 15 × 20 cm2 and that the penetration depth is sufficient to heat intact breast up to 10 cm depth. The field patterns for this 70 MHz rectangular waveguide design match simulations using in house developed Finite Difference Time Domain software [Reference van Stam34]. The waveguide properties in this phantom setup: reflection coefficient (S 11) is 6.84 dB, the Voltage Standing Wave Ratio is 2.668:1, and the relative reflected power is 20.7%. These waveguide properties are very acceptable for our clinical needs to heat an intact breast.
Fig. 4. E-field patterns measured in the XY midplane with the setup of Fig. 2.
B) Tumor temperatures
In Figs 5 and 6, an example of an individual patient is shown, a 65-year-old patient treated with thermoradiotherapy for recurrent adenocarcinoma in the breast with both deep and superficial breast lesions. On this ground, the clinician decided to treat her complete breast with HT using the 70 MHz breast applicator. Prior to treatment, two 14-sensor thermocouple probes were placed in catheters inserted in lateral and in cranial–caudal direction through the breast at 5 and 3.5 cm depth, respectively (Figs 5 and 6).
Fig. 5. Invasive lateral (l) and cranial–caudal (c–c) thermocouple probes placed in catheters in the breast of patient 1.
Fig. 6. CT showing the target region and the location of the invasive lateral (l) and cranial–caudal (c–c) thermocouple probes of Fig. 5, as well as the superficial thermocouple probes on the skin surface (skin).
The combination of 300–400 W antenna power and a water temperature of 42°C was well tolerated for the entire session of 1 h and resulted in good tumor temperatures with T90 = 40.1°C, T50 = 41.6°C, and T10 = 42.1°C for the session of patient 1 shown in Figs 7 and 8.
Temperature profiles for the lateral and cranial–caudal catheters of patient 1 are shown in Figs 7 and 8. No toxicity or complaints were associated with the HT treatment. A water mattress and other measures were needed to assure a comfortable position throughout treatment.
Results for all six patients are shown in Table 1 and Figs 9 and 10. Tumor temperatures recorded during the steady-state period of 60 min averaged over all four sessions per patient and averaged over all six patients were T90 = 39.8°C, T50 = 41.1°C, and T10 = 42.2°C, which is quite acceptable both in terms of temperature uniformity and in terms of minimum temperature achieved. For all patients, the full 30 min preheating period was needed to achieve therapeutic temperature levels in the deepest measurement points in the breast. As the maximum temperature T10 was low in spite of consistently using maximum power for patients 1, 2, 3, and 5, we decided that even better temperatures should be possible by increasing the power level beyond 500 W. We therefore replaced after the fifth patient the 500 W 70 MHz Henry Radio generator with a more powerful generator, in this case the 2.5 kW 70 MHz generator formerly used for the coaxial TEM applicator [Reference De Leeuw and Lagendijk32]. The sixth patient was then treated with 900 W output power. This did result in higher than average tumor temperatures for the sixth patient. Most of this need for a relatively high output power can be attributed to losses in the water bolus, which utilizes tap water, not deionized water. But also tissue perfusion was thought to be high for all patients except patient 4, the only patient where high temperatures were achieved at a relatively modest power level. Power absorption in the breast is definitely not uniform with a three times higher SAR in the muscle-like areas (the denser, lighter grey shade areas in the CT scans of Figs 6–8) compared WITH the more fat-like tissue (the darker areas in the CT scan). Nonetheless, the resulting steady-state temperature distribution is relatively uniform, due to thermal conduction and blood perfusion. This good clinical temperature distribution was achieved with only a single antenna, where other proposed breast applicators always use an array of antennas [Reference Wu, McGough, Arabe and Samulski28, Reference Stang, Haynes, Carson and Moghaddam29], which gives good flexible steering to focus the energy but also adds complexity to the steering. We feel that our single-antenna system is a more robust approach for heating deep-seated breast tumors that is likely to provide adequate heating without the need for extensive pretreatment planning and online adaptive planning required for achieving optimal temperature distributions when using multi-antenna-phased array HT devices [Reference Paulides35–Reference Kok, Wust, Stauffer, Bardati, van Rhoon and Crezee37].
Fig. 7. Steady-state temperature profile in the lateral catheter of patient 1.
Fig. 8. Steady-state temperature profile in the cranial–caudal catheter of patient 1.
Fig. 9. Temperature achieved averaged over the four sessions per patient.
Fig. 10. Power during treatment averaged over the four sessions per patient.
C) Normal tissue hot spots
The combination of up to 900 W antenna power and a water temperature of 42°C was well tolerated for the entire session of 1 h. None of our six patients complained about pain associated with the antenna power. Also none of our temperature measurements gave any indication of normal tissue temperatures exceeding 45°C. This is remarkable as during locoregional heating in the pelvic region with similar waveguides, much lower power per antenna (typically 200–300 W) is applied to avoid normal tissue hot spots. The use of an open water bolus is likely to be a major reason for this remarkable finding. During locoregional heating, a water-filled plastic bag is placed between waveguide and the skin of the patient, and many normal tissue hot spots occur due to fringing fields at the edge of the water bolus [Reference Wiersma, van Dijk, Sijbrands and Schneider38]. This fringing field effect no longer occurs when an open water bolus is used, but distinct anatomical features can still be associated with a high risk of hot spots [Reference Wiersma, van Wieringen, Crezee and van Dijk39]. Apparently, these anatomical features are not present in the intact breasts we treated with our applicator.
D) Patient comfort
Patients tolerated the treatment extremely well, and all sessions were completed as scheduled. No complaints were reported about either the water temperature or the antenna power. The only issue that caused patient complaints was the need to undergo treatment in prone position for up to 90 min. In spite of the water mattress used, most of the patients had difficulty with that position, in particular when patients were more obese. This part of the treatment system will need to be redesigned to increase patient tolerance. However, the EM part of the system seems to function properly without affecting patient comfort.
IV. CONCLUSIONS
The 70 MHz breast applicator system performed well and tumor temperatures were good. Some modifications in the system setup are desired to make the prone position more comfortable for the patient. When this modification has been achieved, this system will be used in a phase I/II study for deep-seated tumors in intact breast treated with RT and HT.
Johannes Crezee received his M.Sc. degree in Experimental Physics from the Free University Amsterdam in 1986, and his Ph.D. degree from Utrecht University in 1993. From 1988 to 2000 he was with the University Medical Center, Utrecht, engaged on several hyperthermia projects, including treatment planning and development of interstitial hyperthermia methods. Since 2000, he has been with the Department of Radiation Oncology, Academic Medical Center (AMC), Amsterdam, as a principal investigator, with a special interest in hyperthermia research, focusing on the development of new hyperthermia equipment and on hyperthermia treatment planning. His current research projects at the AMC supported by grants of the Dutch Cancer Society (KWF) include the application of adaptive treatment planning, planning for bladder cancer, and for irreversible electroporation (IRE) and the use of MRI for hyperthermia treatment planning.
Geertjan van Tienhoven received his M.D. in 1985 and had his training in Radiation Oncology at the Dutch Cancer Institute in Amsterdam between 1987 and 1991. He became a staff member at the Department of Radiation Oncology at the Academic Medical Center of Amsterdam (AMC) in 1991. He received his Ph.D. degree from the University of Amsterdam in 1997 with a thesis on Quality Assurance, Apex Biopsy and Local Control in Breast Cancer. Current research interests are breast cancer and pancreatic cancer. He is co-principal investigator of several large international phase III clinical trials.
Merel W. Kolff received her M.D. in 2001. After 2 years of working at the departments of surgery and internal medicine, she had her training in Radiation Oncology at the Academic Medical Center of Amsterdam (AMC) between 2004 and 2008. She is staff member at the Department of Radiation Oncology at the AMC since 2008 and head of the clinical department of hyperthermia since 2014. She is principal investigator of a phase II randomized study of reirradiation and hyperthermia versus reirradiation and hyperthermia plus chemotherapy for locally recurrent breast cancer in previously irradiated area.
Jan Sijbrands works as a mechanical engineer at the Department of Radiation Oncology of the Academic Medical Center in Amsterdam.
Gerard van Stam received his Bachelor in Electronic Engineering from the Technical College Amsterdam in 1979. He worked from 1987 to 2014 at the Department of Radiation Oncology of the Academic Medical Center in Amsterdam, focusing on the development of electronic computer controlled equipment capable of achieving hyperthermia and the equipment quality control. He is currently working on optimization of the Quality Assurance procedure for superficial and deep hyperthermia equipment at the Department of Radiation Oncology of the Kanton Spital Aarau, Switzerland.
Sabine Oldenborg received her M.Sc. degree in Medical Biology at the University of Amsterdam in 2000 and works since 2001 at the Department of Radiation Oncology of the Academic Medical Center in Amsterdam. She is currently working toward her Ph.D. degree collecting and analyzing treatment data of more than 800 breast cancer patients treated with radiotherapy and hyperthermia at the Academic Medical Center in Amsterdam and the Verbeeten Institute in Tilburg.
Elisabeth D. Geijsen received her M.D. degree in 1996. After 2 years of residency at Deventer Ziekenhuizen Deventer, she had her training in Radiation Oncology at the University Medical Centre Groningen (UMCG) between 1998 and 2003. In 2002, she had an internship at Duke University, Durham, North Carolina, with subject “Adhesion of leukocytes and endothelial cells in tumour vessels”. She is staff member at the Department of Radiation Oncology at the Academic Medical Center of Amsterdam (AMC) since 2004. She is principal investigator of a phase III randomized trial of intravesical chemotherapy versus thermochemotherapy in intermediate risk non-muscle invasive cancer which is financially supported by the Dutch Cancer Society KWF.
Maarten C.C.M. Hulshof had his training in Radiation Oncology at the Dutch Cancer Institute in Amsterdam and is staff member at the Department of Radiation Oncology at the Academic Medical Center of Amsterdam (AMC) since 1988. He received his Ph.D. degree from the University of Amsterdam in 2002 with a thesis on radiation effects in glioblastoma patients. Current research interests are urological oncology and upper gastrointestinal (GI) tumors, and he is principal investigator of several phase II and III clinical trials. He is chair or board member of several Dutch upper GI and uro-oncology tumor groups.
Henny P. Kok received her M.Sc. degree in Computational Science at Utrecht University in 2002. Thereafter, she worked as a Ph.D. student at the Department of Radiation Oncology of the Academic Medical Center (AMC) in Amsterdam on a project to develop a dual modality heating technique for hyperthermia treatment of esophageal cancer. She received her Ph.D. from the University of Amsterdam in 2007 and continued as a post doc at the AMC on optimization of locoregional hyperthermia delivery. Her research was awarded five young investigator awards and an Editor's Award of the International Journal of Hyperthermia. Currently she works as a researcher on advanced hyperthermia treatment planning and biological thermoradiotherapy planning to model the effect of combined radiotherapy and hyperthermia, supported by grants of the Dutch Cancer Society (KWF).
