|Home | About | Journals | Submit | Contact Us | Français|
This paper reviews systems and techniques to deliver simultaneous thermoradiotherapy of breast cancer. It first covers the clinical implementation of simultaneous delivery of superficial (microwave or ultrasound) hyperthermia and external photon beam radiotherapy, first using a Co-60 teletherapy unit and later medical linear accelerators. The parallel development and related studies of the SURLAS, an advanced system specifically designed and developed for simultaneous thermoradiotherapy, follows. The performance characteristics of the SURLAS are reviewed and power limitation problems at high acoustic frequencies (> 3MHz) are discussed along with potential solutions. Next, the feasibility of simultaneous SURLAS hyperthermia and IMRT/IGRT is established based on published and newly presented studies. Finally, based on the encouraging clinical results thus far, it is concluded that new trials employing the latest technologies are warranted along with further developments in treatment planning.
Conventional hyperthermia in cancer therapy can be defined as the elevation of tissue temperatures to 41~43°C for more than 30 minutes. Hyperthermia is typically used as an adjunct therapy to radiotherapy and/or chemotherapy [1–5]. Many well-conducted clinical trials, including phase III multi-institutional trials following quality assurance guidelines, have shown that hyperthermia can significantly increase both local tumor control rates and duration of local control in tumors that recur or persist after surgery, radiotherapy and/or chemotherapy [6–9]. Furthermore, hyperthermia has been shown beneficial in the treatment of residual microscopic disease in the management of local-regional breast cancer , in the treatment of soft tissue sarcomas preoperatively with radiation , and in the treatment of deep pelvic tumors [12, 13]. A significant survival benefit was also reported when hyperthermia was combined with brachytherapy in the treatment of glioblastoma multiforme  and with external beam radiotherapy in the treatment of recurrent head and neck cancer .
In fact, when clinical trials have been conducted under widely accepted quality assurance guidelines—which implies matching the heating technology to the target site along with the implementation of a reasonable thermometry strategy—radiotherapy plus hyperthermia has always resulted in statistically better outcomes than radiotherapy alone .
Patients with persistent and/or recurrent breast cancer and chest wall tumors have significantly benefited when hyperthermia has been added to their radiotherapy and/or chemotherapy regimens [17–26]. One of the physical reasons for the success of hyperthermia in this group of patients is the superficial location of the local disease. Superficial lesions are the least difficult to heat adequately because of their accessibility and proximity to external energy sources. The composition of the local anatomy may also contribute to better heating [27, 28]. Another possible reason may be that superficial lesions are more amenable to invasive thermometry and thermal mapping of temperature sensors, thus providing more temperature feedback data during treatment, which can be used to improve heat delivery and ensure treatment quality [29, 30].
Clinical research efforts have demonstrated that the response of cancerous tumors to sequential thermoradiotherapy (i.e., sequentially combined radiotherapy and hyperthermia) is well correlated with power deposition coverage and/or thermal dose coverage [24, 30–40]. These studies have also point to the challenge of achieving consistently biologically meaningful thermal doses in 100% of the target volume in routine clinical practice. Therefore, it is widely accepted that the benefits of hyperthermia as an adjunct to radiotherapy can be significantly augmented with improvements in treatment delivery techniques, better heating technology, advances in treatment planning and implementation of quality assurance guidelines [6, 13, 30, 41, 42]. Moreover, in vitro and animal studies have shown that when hyperthermia and radiation are administered simultaneously —rather than sequentially as it is conventionally done—heat-induced radiosensitization (HIR) is increased at thermal doses achievable in the clinic [43–47]. For instance, mild hyperthermia—41°C maintained for ~60 minutes—produces HIR if radiation and heat are delivered simultaneously, but not sequentially. This is significant because a minimum target temperature of 41°C is more clinically achievable than the > 42°C needed to produce HIR with sequential treatment [44, 47, 48]. The above facts support the development of clinical hyperthermia devices that permit simultaneous delivery of heat and ionizing radiation.
In this section, a review is given of the systems and techniques clinically implemented for the delivery of simultaneous thermoradiotherapy at Washington University in St. Louis. This research group has been the only one to complete several simultaneous thermoradiotherapy clinical trials with breast cancer being the site treated the most.
Although it had been known from in vitro and in vivo experiments that simultaneous delivery resulted in larger thermal enhancement ratios than sequential delivery, simultaneous treatments had not been tried with human patients mainly due to formidable logistical problems . The research group set out to overcome these problems and embark in a long-term clinical research effort. The first approach consisted of heating with 915 MHz microwave applicators and irradiating with a Co-60 teletherapy unit . Both technologies enjoyed extensive clinical experience, were well-understood and had been used in sequential thermoradiotherapy [39, 40]. Therefore, any unusual response to the new therapy approach could be attributed to the simultaneous delivery of heat and gamma radiation. Hyperthermia waveguide applicators were used, one at the time. Two clinical setups were characterized. In one a microwave applicator was attached to a blocking tray and inserted into the tray slot in the head of the gantry of the teletherapy unit. Thus the gamma beam traveled through a mostly hollowed waveguide and into a treatment volume after transversing a mineral oil coupling bolus. This was called the en face setup. The other setup was with the microwave propagation vector perpendicular to the gamma beam central axis. This was called the orthogonal setup. Using these setups a 400 cGy radiation fraction of 4 to 7 minutes duration could be delivered in the middle of a 60 minute hyperthermia treatment without interruptions. Temperatures and power level were remotely monitored and recorded outside the teletherapy room. The microwave applicator was not mounted or in contact with the teletherapy unit in the orthogonal setup, so it did not affected the radiation dose distributions. However, for the en face setup the dose distributions were affected by the microwave applicator. Film measurements in a solid water phantom showed uniform dose within the radiation field, except for 10–18% attenuation under the metal tuning electrodes inside the waveguide. This dose defect was clinically smoothed using feathering techniques. This approach was used successfully without technical problems in a phase I/II clinical trial. Importantly, an analysis showed that the temperature distributions achieved during simultaneous delivery had the same general characteristics as those achieved during sequential delivery, and that the steady state distributions were maintained during the time of simultaneous irradiation. This first experience demonstrated that simultaneous superficial microwave hyperthermia and external beam radiation was technically feasible and safe .
The waveguide microwave applicators limited the size of the lesions we could treat with hyperthermia simultaneously with gamma rays [39, 40]. In order to be able to treat larger lesions, we decided to adapt a commercial planar ultrasound system to deliver simultaneous treatments. This task was much more challenging because the ultrasound applicators were not hollowed like the microwave waveguides and degassed water was needed to couple the sound waves into the treatment volume. The ultrasound applicators of the Labthermics 1000 system were also bulkier and heavier than the microwave applicators. An orthogonal setup similar to the one used with microwaves was relatively simple to mimic, but many tumor sites would require an en face setup for proper treatment. Since ultrasound is reflected efficiently and specularly from metal surfaces, we devised a reflecting system in which the ultrasound beam, initially directed perpendicularly to the gamma beam, was deflected 90° so that both beams travel through the same window of entry into the tumor while the ultrasound source remained outside the radiation beam . The reflecting system, which was mounted on the blocking tray, was filled with degassed water and made of water-equivalent materials, except for a 1 mm sheet of polished brass used as the reflector. With this en face approach the absorbed power patterns generated with and without the reflecting system at the same extended distance from the transducer demonstrated that the hyperthermia system remained able to control heating pattern s. The effect of the reflecting system on the gamma beam was minimal except for uniform attenuation, and the Co-60 beam had no effect on the performance of the thermocouples. Extensive testing followed by clinical treatments of patients on protocol demonstrated that the modifications made did not impair the ability to deliver ultrasound hyperthermia or teletherapy effectively. Again, this work showed that the implementation of ultrasound hyperthermia simultaneous with gamma irradiation was technically and clinically feasible without complications or hazards to patients.
In 1995 the Co-60 teletherapy unit at the Mallinckrodt Institute of Radiology was decommissioned. The only option for external beam radiation was to develop the devices and techniques to allow the use of a medical linear accelerator. Although excited about the use of modern radiation technology, the transfer to megavoltage photon beam necessitated additional developments and measurements. Both the orthogonal and the en face setups were successfully transferred, but had to be adapted to the different geometry and dimensions of the linear accelerator. Straube et al. (2001) reported on devices, techniques and dosimetry for simultaneous thermoradiotherapy delivered with a linear accelerator including detailed day-to-day simulation and treatment procedures .
The above described developments have made possible the conduct of clinical trials since 1992. The first phase I trial was ended successfully in 1995, a second phase I/II dose escalation trial closed in 2001, and a third trial was initiated 1999. These clinical results have been summarized previously by Myerson et al. (1999, 2004) [44, 47]. Up-to-date, 119 patients have been enrolled on these three consecutive trials. The first two trials (60 cases) were for patients with macroscopic disease, usually recurrent after prior therapies, including radiotherapy, and escalated the number of simultaneous treatments from 3 to 8 in a course of 30–32 Gy. The third trial was to evaluate the short-and long -term effects and local control of simultaneous thermoradiotherapy in the treatment of high risk, but curative, subclinical (no visible disease after surgery and/or chemotherapy) breast carcinoma in patients with no prior radiotherapy. The simultaneous administration of radiotherapy and hyperthermia is of particular potential value for this patient population because tumor cells share normal tissue physiology. Fifty nine patients were accrued in this trial. The number of simultaneous treatments was escalated from 4 to 8 in a course of 60 Gy; most patients receiving 400 cGy per hyperthermia treatment—approximately 26% to 53% of the radiation dose was given under hyperthermic conditions. Elective hyperthermia was given to the primary tumor bed site plus an adjacent 6×12cm strip randomized to lie medial or lateral to the primary site. The adjacent strip served as an unheated but irradiated control. The first two studies demonstrated that the per treatment average thermal dose progressively improved with the number of hyperthermia treatments. In addition the product of radiation dose and total thermal dose was highly correlated with complete response . In the third trial the overall local control was 97% with no significant morbidity difference between heated and control sectors and no increase in morbidity between 4 and 8 simultaneous treatments.
Finally, the data clearly shows that chest wall lesions (recurrent breast cancer) have been the most responsive to simultaneous thermoradiotherapy with highly significant statistical correlation with radio-thermal dose factors (p < 0.05) . This exciting finding points to the importance of developing thermoradiotherapy treatment planning systems and continuing efforts to improve clinical thermal dosimetry techniques.
In parallel with the efforts to adapt existing commercially available hyperthermia technology to deliver simultaneous hyperthermia and external photon beam radiation, ideas were being discussed for the development of a system specifically designed for this purpose. During the decade of the nineties, electron beams were used commonly in the treatment of superficial lesions due to their limited depth of penetration, which meant sparing of underlying normal tissues. However, the existing hyperthermia devices were too bulky/dense to permit the use of electron beams. This was the main reason only high energy photon beams had been used when delivering simultaneous thermoradiotherapy. Eventually, the concept of the Scanning Ultrasound Reflector Linear Arrays System, or SURLAS, was born [52, 53]. By combining ultrasound linear arrays and scanning ultrasound reflectors, an applicator thin enough to permit the passage of high energy electron beams could be made to work using an en face approach. Obviously, photon beams would also work as in the past. The device concept is illustrated in Figure 1, which shows a computer-aided solid model of the clinical grade applicator that was developed [54, 55]. Figure 1 shows the basic components of a dual-frequency SURLAS applicator. The reflecting surfaces of a scanning double-faced V-shaped reflector make 45° angles with respect to the sound propagation coming from the (high and low frequency) arrays at the side ends of the applicator head (Figure 1 shows the windows for the arrays, not the arrays), thus both sound beams are deflected in parallel toward the treatment volume. As the reflector scans back and forth, the sound energy is distributed over the treatment volume producing therapeutic time-averaged temperature distributions with acceptable temperature fluctuations [56, 57]. Each array has several individually powered elements. Modulation of power input to the array elements as the reflector scans between the arrays provides control of the two-dimensional surface energy fluence pattern [55, 58]. This capability is referred to as Lateral Power Conformability or LPC. One of the arrays operates at a low frequency (1 ~ 2 MHz) while the other operates at a high frequency (3 ~ 5 MHz). This arrangement allows concomitant dual-frequency insonation. Penetration depth control, or PDC, is thus achieved by varying the power input to one array relative to the other’s, i.e., frequency mixing [59, 60]. For example, for areas where the target volume is deeper, the low frequency can be weighted more heavily than the high frequency. LPC plus PDC means that the SURLAS has 3D power deposition control, which can eventually be fully exploited with the implementation of temperature feedback control strategies based on temperatures measured during treatment .
Several laboratory prototypes of the SURLAS were developed over the years in order to demonstrate feasibility of the overall concept and of particular features such as LPC and PDC. Finally, a clinical grade system was developed that could be used under approval of the Federal Food and Drug Administration (FDA) via the Investigational Device Exemption (IDE). However, manufacturing a clinical device proved to be a much more rigorous process than assembling prototypes for laboratory studies. Important requirements concerning reliability, patient and operator safety, and quality control/assurance presented great challenges to there search group. In addition to these rigorous requirements, the following were specific design criteria for the clinical SURLAS applicator:
These design criteria were met for the most part as reported in a paper published in early 2005 that focused mainly on the hardware components of the clinical grade SURLAS . A second paper published later that same year reported on the design, development, and testing of the personal computer based treatment delivery software that coordinated the interactions between the operator, the SURLAS applicator and several peripheral devices . There were two important tasks in software development worth mentioning here. The first was the coordination of the input power sequences to the elements of the high (4.9 MHz) and the low (1.9 MHz) frequency arrays (eight 1.5 cm × 2.0 cm elements/array) with the position of the dual-face scanning reflector. This was achieved by dividing the treatment window in up to 64 sectors (minimum size of 2cm × 2 cm) and controlling the power delivery to each sector independently by adjusting the output power from the 16 channels of a programmable radio-frequency generator. Upon completion of the software development process, it was integrated with the hardware followed by extensive testing. The second important task was safety, which was a paramount concern and design criterion. To ensure safety, a failure mode and effects analysis, or FMEA, was applied to the entire system in order to identify safety issues and rank their relative importance. The FMEA analysis led to the implementation of a software structure where each peripheral device communicated independently and directly with the controlling personal computer, so that in case of a malfunction in any component of the system or any violation of a pre-defined safety criterion, the software can terminate treatment immediately.
As mentioned before the original motivation and goal was to develop an ultrasound hyperthermia device that would enable simultaneous hyperthermia and electron beam radiation . Consequently, we tried to make the profile (height) of the SURLAS as small as possible in order that a high energy electron beam could travel through it and still be energetic enough to penetrate into superficial lesions as deep as 3 cm (electron beams are attenuated in water about 2 MeV/cm and the 80 to 90% isodose cloud is commonly used for tumor coverage). This meant that we implemented the smallest arrays possible that our previous studies indicated suitable in terms of the required emittance (W/cm2) to induce adequate hyperthermia. Note that the smaller the array the higher the required continuous wave (CW) emittance. This choice and other factors discussed below, however, created problems in practice that required time to figure out and resolve and de railed the clinical translation of the SURLAS.
The major problem was a decrease in electrical-to-acoustic efficiency with increasing transducer-to-reflector distance . This was particularly obvious for the high frequency array as shown in Figure 2, which presents radiation force balance measurements for the low (2a) and high (2b) linear arrays. This loss of ultrasonic power with distance was determined to be non -linear propagation, a subject of much research in recent years. Essentially, nonlinear propagation is the conversion of ultrasonic energy from the fundamental frequency to higher frequency harmonics as the waves are distorted while propagating from sinusoidal-like to sawtooth-like . The wave distortion increases with the acoustic intensity, increasing distance from the source and increasing frequency. It also depends on the characteristics of the medium. This phenomenon has been described by the shock parameter σ ,
where β is the coefficient of nonlinearity of the medium, ε is the mach number, k is the wave number, and x the travel distance. Based on calculations using the shock parameter we determined that the high frequency array was suffering significant power loss due to non-linear propagation. For the clinical SURLAS, the maximum travel distance of the waves from the array to the distal water bolus membrane (including reflection) was about 23 cm. For the high frequency waves(~ 4.9 MHz), nonlinear propagation was calculated to occur at acoustic intensities as low as 0.7 W/cm 2at the maximum distance for σ =1 (onset of nonlinear propagation). The acoustic intensities required for hyperthermia are higher than this value . Moreover, note that scanning of the ultrasound energy over the treatment window requires higher CW intensities in comparison to a single transducer of comparable size to the treatment window [53, 64]. When σ = 3, the waveforms become sawtooth causing intensity losses of 6 dB with respect to the fundamental frequency. Assuming plane wave behavior, a series of simulations were generated to determine 1) the CW intensity required for the clinical SURLAS to induce and maintain hyperthermic temperatures, and 2) the maximum deliverable CW intensity for a shock parameter equal to3 (σ = 3). The frequency, the treatment widow length (~ scanning distance), and height of the linear array were variables. Selected results are presented in Table 1, which shows the maximum fundamental frequency (fmax) for four treatment window lengths (WL) and three array heights (AH). The AH for the clinical SURLAS is 1.5 cm and the thickness of the water coupling bolus is approximately2 cm. The developed clinical SURLAS has a maximum window length of 16 cm. Table 1 indicates that to utilize a window this size fmax could not be higher than 1.89 MHz. The frequency of the high frequency array is ~ 4.9 MHz, hence significant nonlinear propagation effects are to be expected per equation (1) and as shown in Figure 2. We had seen this effect before with our commercial ultrasound system for the en face setup at the frequency of ~3.5 MHz , but the source to skin distance there was over 50 cm, and despite the ~40% reduction of power adequate hyperthermia treatments were possible. This indicated that other factors were contributing to low efficiencies. These were:
The message from Table 1 and the above listed factors is clear, to avoid nonlinear propagation and improve electrical-to-acoustic efficiency, a new SURLAS must be constructed considering the following: lower frequency for the high frequency array, larger surface areas arrays (larger AH), shorter treatment window length (smaller WL), reflecting areas > array projected areas, and a more powerful radiofrequency generator with flat frequency response. The lower frequency for the high frequency array means some reduction in PDC. The smaller treatment window implies a proportional reduction in the size of the treatment area. A larger array height means a higher profile applicator—less or not suitable for electron beams. And larger scanning (possibly metallic) reflector precludes electron beams. That leaves only photon beams for the delivery of simultaneous thermoradiotherapy with the SURLAS. As discussed in the next section, this was actually welcome news due to advances in modern radiotherapy, especially intensity modulated radiation therapy or IMRT, as it is commonly called.
Intensity modulated radiation therapy (IMRT) has become clinically widespread worldwide in the last decade . IMRT’s appeal is its ability to focus dose to the targeted volume (commonly called the planning treatment volume or PTV) while minimizing dose to surrounding normal tissues. Our first working hypothesis was that IMRT can produce clinically acceptable dose distributions on a recurrent breast cancer tumor while that same tumor receives ultrasound hyperthermia using the SURLAS. To test this hypothesis two research projects were initiated: a radiation treatment planning study to compute dose distributions in recurrent breast cancer PTVs with the SURLAS applicator in place ; and an experimental study, where a radiation treatment plan was computed and delivered to a phantom with the SURLAS applicator on top of the phantom simulating a clinical setup, to evaluate the effects of the presence and potential misplacement of the SURLAS applicator on measured dose distributions .
The above studies were carried out using the treatment planning station and treatment unit of a Helical Tomotherapy (HT) system (Tomotherapy Inc., Madison, WI). This technology was chosen over conventional gantry-based medical linear accelerators (both available at our institution), because we thought that the multiplicity of beam angles and the helical motion would greatly minimize the perturbing effects of the SURLAS applicator on dose distributions. In addition, the mega voltage computed tomography (MVCT) feature of Tomotherapy, which is used for image-guided radiotherapy (IGRT) to verify the patient’s position right before treatment delivery, could also be used for treatment planning purposes without major image distortions or artifacts due to the metallic parts of the SURLAS applicator. Moreover, the MVCT images could be also used to verify the position of the SURLAS applicator with respect to the patient and from treatment to treatment. All these ideas were tested successfully [66, 67].
The main conclusion from the treatment planning study was that simultaneous treatment with the SURLAS and HT IMRT/IGRT is feasible as demonstrated by the clinically acceptable radiation treatment plans generated . Likewise, the main conclusion of the experimental study was that the delivered (measured) and planned radiation dose distributions were in excellent agreement when the SURLAS applicator was positioned as it was planned. Misplacement of the applicator from its planned position was found to have small effects on the measured dose distributions . Potentially large discrepancies between planned and delivered doses due to gross misplacement of the SURLAS can be further minimized by the implementation of routine quality assurance procedures using a position-and-orientation magnetic tracking device which was successful in precisely and reproducibly positioning of the applicator. Moreover, it was shown that there is great potential for the use of this tracking device for monitoring of the SURLAS applicator position and orientation during simultaneous treatments, and not just for patient/SURLAS set-up prior to daily treatment using MVCT .
Do the above findings apply if a conventional medical linear accelerator is used? This is important question to answer because gantry-based linacs remain the most commonly used technology in radiation therapy departments. Here we present new treatment planning results that indicate that Linac-based IMRT can also deliver clinically acceptable radiation dose distributions with the SURLAS applicator on the patient’s chest wall. The clinical setup and planning MVCT can be seen on the axial plane of Figure 3. The PTV was the superficial case in Penagaricano et al. 2008, with maximum dimensions of 1 cm deep by 11.5 cm wide (lateral) by 11 cm long (superior-inferior) on the lower chest/upper abdominal wall mimicking a chest wall recurrence from breast cancer. The results of Figures 3 and and44 were generated using the Pinnacle3 inverse IMRT treatment planning system (P3IMRT, Philips, Andover, MA) to model megavoltage photon beam dose deposition delivered by a Varian 2100 EX linear accelerator with 120-leaf dynamic multi-leaf collimator (Varian Medical Systems, Palo Alto, CA). A 12 beam plan was used, 8 beams were coplanar (linac central axis on the axial plane) and 4 beams were non-coplanar (linac central axis on the sagittal plane). Of the 8 coplanar beams, 6 were aimed from above the SURLAS applicator and 2 were aimed laterally and tangentially to the chest wall. The noncoplanar beams were all aimed from above. The prescription was 50 Gy in 25 fractions to 95% of the PTV using 6 MV photons. It was assumed that every fraction was given under hyperthermic conditions for the purpose of illustration. Cardiac toxicity would be solely dependent on radiation dose because none of the existing or forth coming devices would be able to induce significant hyperthermia in heart tissue due to 1) the depth of heart tissue and 2) the very high cardiac perfusion rates.
Figure 3 shows computed dose distributions on the axial, sagittal and coronal planes through the PTV and SURLAS applicator. It can be readily seen that the mass was nicely covered by the prescription dose with low doses to surrounding normal tissues. Dose volume histograms (DVH) for the PTV and organs at risk are shown in Figure 4. These DVHs are clinically acceptable for the PTV and surrounding organs and are comparable to the ones obtained using the Tomotherapy system . The PTV mean dose, the PTV D95%, the heart, lung and abdomen mean doses, and the spinal cord maximum dose were 50.7, 50.4, 16.7, 3.7, 12.9 and 10.0 Gy, respectively. These metrics are in good agreement with those reported by Penagaricano et al. 2008 for the Tomotherapy system (Table IV in ). In short, IMRT with either radiation therapy system is suitable for simultaneous thermoradiotherapy using the SURLAS.
Besides more conformal radiation doses, there are other advantages of using IMRT with the SURLAS. First, IMRT is expected to be highly forgiving of the modifications needed to the SURLAS applicator to minimize non-linear propagation and improve electrical-to-acoustical efficiencies. Second, the clinical experience thus far has only used one radiation beam concomitantly with hyperthermia; consequently the time of simultaneous delivery is short compared to the heating time. With IMRT and the SURLAS, the simultaneous time is expected to be longer for hyperthermia can be given during an entire multi-beam IMRT fraction. Third, the application of IGRT and magnetic position-orientation tracking technologies, to ensure reproducible applicator positioning prior and during the treatments, may encourage physicians to prescribe hyperthermia simultaneously with more if not every IMRT fraction. Finally, more simultaneous fractions should maximize the radiosensitizing and physiological effects of hyperthermia even if IMRT is delivered in shorter times using the emerging arc therapy paradigm .
The clinical implementation of simultaneous thermoradiotherapy has been challenging but proven to be feasible and safe. Clinical trial results have been encouraging and breast cancers are thus far the most responsive, with statistically significant correlations (p < 0.05) with radio-thermal dosimetric factors . Therefore, new technological developments such as combining the SURLAS and IMRT/IGRT with thermo-radio dose treatment planning should translate into improvements in radiothermal dose coverage and consequently in better treatment outcomes. Furthermore, several research groups have been developing advanced simultaneous heat and irradiation technologies that are moving toward clinical deployment while others have attempted different clinical approaches using existing hyperthermia devices [70–76]. In conclusion, new clinical protocols combining hypethermia and radiotherapy simultaneously are now possible and warranted.
The work reviewed in this paper was supported by NIH Grants CA 63121 and CA 71638, and by a grant from the Whitaker Foundation. We gratefully acknowledge partial support from the Central Arkansas Radiation Therapy Institute (CARTI) and a Medical Research Endowment pilot award from College of Medicine of the University of Arkansas for Medical Sciences, Little Rock. We thank our cancer patients for their courage and participation in the clinical research. Finally, we gratefully acknowledge the contributions of many collaborators over the years as reflected in the authorship of previously published papers and abstracts.