HYPERTHERMIA PHYSICS

 The objecive of local or regional hyperthermia therapy is to achieve tumor temperatures in the range of 40℃ to 45℃ and to maintain that temperature for a time period on the order of 1 hour. Attaining this thermal dose goal throughout a tumor volume requires the delivery of power with sufficient spatial and temporal variation to balance the heterogeneous and time-varying heat transfer processes within the target volume. Heat transfer from thermal conduction and blood perfusion redistribution of energy within living tissue are complex, and achieving the desired thermal unniformity across large tissue regions is difficult.

 Thermal conduction is a well-understood process, but in practical delivery of clinical hyperthermia the thermal conductivity is complicated by the irregularities of patient anatomy and tissue interfaces and varies by more than factor of two between high-water-content tissues and fat or bone. An even larger uncertainty in heat transfer is caused by blood perfusion, which varies drramatically among tissue types, and as a function of time and local temperature.

 Even with excellent control of the power deposition, knowledge of tissue thermal properties is needed in order to preplan temperature uniformity within a given target volume. In addition, since blood flow changes rapidly, temperature should be measured continously during treatment in order to readjuvant the heating pattern for uniform temperature.

 

 Clinical hyperthermia can be accomplished using one of three modalities: thermal conduction(e.g., circulating hot water in a needle, a catheter, or a surface pad), nonionizing electromagnetic radiation(EM), or ultrasound(US). In each case, heating is sensitive to tissue geometry, heterogeneity of tissue properties(especially blood perfusion), and issues with coupling energy into tissue from practical-sized applicators. Because of the limited penetration of heat from a hot source, thermal conduction heating is largely restricted to interstitial applications with closely spaced implant arrays.

 Energy can be delivered deeper into tissue using EM or US fields that deposit energy directly in tissue, with effective penetration dependent on frequently and applicator type. For either source, the absorbed power distribution is commonly normalized by the respective tissue density and referred to as the specific absorption rate(SAR) distrubution, with units of watts per kilogram. The underlying physical principles are described in several excellent review articles and in books that cover the field of hyperthermia. Representitive examples of typical heating equipment for each modality are giving in the following subsecitons.

 

#. Electromagnetic Heating

 When a radiofrequency electric field(E field) is applied to tissue, heating occurs from resistive losses as a result of electric current in a resistive media(tissue). At higher microwave frequencies, heat is generated from mechanical interactions between adjacent polar water molecules aligning to thae alternating EM field. In either case, energy deposition is proportional to the electrical conductivity and square of the time-averaged E field. Electrical conductivity varies 50-fold depending on tissue type and frequency, with higher conductivity for high-water-content tissue like muscle and internal organs and much lower conductivity for fat and bone. The electrical properties of most tissue have been characterized over a large range of frequencies and published previously.

 Power deposition has better penetration with lower-frequency, longer-wavelength fields, but the longer the wavelength, the broader is the focus of heating. To obtain deep peneetration in the body, multiple-antenna arrays are required to diffuse surface heating, and in combination with the longer wavelengths this produces regional energy deposition that involves substantial volumes of tissue.

 

 EM heating devices can be separated into two categories: superficial HT applicators, with effective penetration into tissue in the range of 1 to 4 cm; and deep HT devices, which have effective penetration >4 cm. Superficial HT devices include waveguides, modified horns, capacitive and inductively coupled sheet antennas, and microstrip or pathc antennas that typically operate at 433, 915, or 2,450 MH.

 In recent years, multiple-antenna planar and conformal array applicators have become more common to inrease the size of tumor that can be heated, as well as to provide lateral adjustment of power deposition to a accommodate heterogeneous tissue. These devices provide lateral adjustment of the power deposition pattern to equilibrate hot and cold regions across a tumor. Superficial microwave applicators are usually coupled into tissue through a deionized water bolus to accommodate surface irregularity. In addition, the bolus is usually temperature controlled to help maintainn skin temperatrure below 44℃. Although current developmental multiantenna array applicators should be availabe commercially in the near future, the majority of superficial clinical HT is performed with BSD Medical waveguide applicators in the United States and Lucite cone applicators or contact flexible microwave applicators in Europe.

 

 In order to heat at depth >4 cm, one must use lower frequencies in the radiofrequency(RF) band from 50 kHz to 150 MHz. There are three basic techniques for EM power deposition deep in tissue: magnetic induction heating, capacitively coupled RF current heating, and RF phased-array heating.

 Magnetic fields penetrate to the body axis but induce local eddy current loops that are governed by paths of least resistance and cause power deposition peaks and nulls that cannot be controlled externally. Although regional heating with magnetic inductions failed to show sufficient control clinically, there is interest in using magnetic fields to couple energy into implanted ferromagnetic seeds and nanoparticles.

 

 The capacitive coupling technique uses RF fields from 5 to 30 MHZ to drive current between two or more saline pad electrodes. Heat is concentrated under the electrodes in tissues with high resistance(e.g., fat). The electrodes are generally aggressively cooled to prevent hot spots on the skin surface and superficial fat. Although there is lack of real-time adjustment of heat distribution during treatment, energy can be concentrated on one side using a smaller electrode. The technique has been used effectively in the clinic for superficial and moderate depth tumors, predominantly in Asian patients with thin fat layers that can be cooled sufficiently with surface cooling. Another possible approach is to drive RF current between an interstitial needle and a large-surface-area return electrode to focus heating at depth around the implanted electrode. Similarly, this can be done with an implanted balloon electrode for intracavitary applications(e.g., esophagus).

 

 The third option for noninvasive deep heating is the RF phased-array technique, which uses a body-concentric array of dipole antennas or large waveguides. Driving multiple antennas with equal phase produces a heat focus centrally in a concentric array, providing much deeper penetration than is possible operating the antennas noncoherently. By applying a phase delay to some of the antennas, one can steer the central focus laterally.

 Systems in the 1990s includes four separate phase-and amplitude-controlled antennas in a single concentric array around the body. Since then, both dipole and waveguide array systems have been expanded to include two or three rings of antennas to provide axial as well as lateral adjustment of heating at depth.

 In general, RF phased arrays have more flexibility in adjusting the SAR pattern than magnetic induction and capacitive heating techniques. The capacitive approach, such as Yamamoto Thermotron RF-8, and RF pahsed-array applicators like the BSD Medical Sigma Ellipse and Sigma Eye are the most frequently used deep heat systems. 

 

#. Ultrasound Heating

 Energy transfer from an acoustic field results from the mechanical losses of viscous friction. The penetration of a US field decrease with increasing frequency, just as for EM energy. However, the wavelength of US is orders of magnitude smaller than that of an EM field of similar penetration capabilities, so that US energy can be focused into small tissue volumes. The use of small US applicators is an advantage for producing controllable sizes and shapes of focal regions at depth. However, anatomic geometry and tissue heterogeneity limit the use of US in many regions of the body because air reflects and bone preferentially absorbs US energy. The availability of an adequate "acoustic window"(path for US beam unobstructed by bone or air proximal or distal to the target) is the primary consideration for clinical applications.

 

 Superficial tissue disease may be heated with a single unfocused US tranducer, but array applicators are preferred for larger volumes typical of human disease. One popular clinical system involves a 4 x 4 transducer planar array applicator that treats up to 15-cm-square area. Effective penetration of this lightly focused array is 2 to 6 cm using frequencies in the 1- to 3.5-MHz range. Tranducers are coupled with a temperature-regulated conformal water bolus to control skin temperature, and good contact is ensured with US coupling gel. With all 16 tranducers at the same frequency, the power deposition pattern can be adjusted laterally by varying power to the transducers, but only minor adjustment of penetration depth is possibly by changing the water cooling temperature.

 

 Deep heating with US may be obtained usting stationary arrays of focused tranducer, eletronically phase focused arrays, or mechanically scanned focused arrays in the 0.5- to 2-MHz range. At these frequencies, penetration in soft tissue is substantial, and care must be taken to avoid problems with bony structures both in front of and behind the target volume. In addition, careful consideration of the beam path both entering and exiting the body is required to avoid potential surfaces burns from reflections at body/air interfaces. Ultrasound arrays have been used successfully for treating relatively small tissue regions at depth in the body, which is particularly appropriate for image-guided thermal ablation. Extension of ultrasound arrays for treating large tumors deep in the body has been attempted but has not proven practical for clinical use.

 

#. Interstitial Hyperthermia

 Interstitial heating shares many of the characteristics of interstitial brachytherpay-highly localized and inhomogeneous dose distribution, invasiveness, and sensitivity to specific site and approach. Heat treatments are usually combined close in time with brachytherapy, making dual use of implants for both hyperthermia and radiation. Some techniques permit simultaneous delivery, but most clinical experience has been with heat and radiation delivered sequentially. As reviewed previously, there are numerous technologies for interstitial hyperthermia.

 

 The simplest approach involves heating the implants themselves with resistive heating elements, circulating hot water, or coupling energy into ferromagnetic implants from an external magnetic field. Dense implant spacing(~1.0 cm between sources) is required because heat transfer is by thermal conduction only. More commonly, interstitial heating is accomplished within closely spaced arrays of 0.5- to 30-MHz radiofreauency electrodes or 433- to 2,450-MHz microwave antennas. With RF electrodes, heating occurs from current between the exposed metal, and there is no mechanism to adjust heating along the electrode length during treatment. High current density in tissue near the electrode surface requires close electrode spacing(1 to 1.5 cm). Heterogeneous tissue conductivity or nonparallel RF electrodes compromise temperature uniformity. By raising the freaquency from about 500 kHz to 10 to 30 MHz, current is coupled capacitivvely through the catheter wall, and electrodes may be segmented to enable adjustment of heating along the implant length.

 

 Interstitial microwave antenna have been designed for operation at frequencies of 433, 915, and 2,450 MHz, which produce different heating lengths in tissue(approximately 9, 4, and 2 cm, respectively). Although the antenna radiation pattern cannot be adjusted during treatment to accommodate different tumor dimensions, some disigns offer the ability to select antennas of different heating lengths. Although heating is generally no uniform along the length of interstitial microwave antennas, the radial penetration is higher than with RF or thermal conduction sources, permitting slightly larger spacing(e.g., 1.5 to 2 cm). In addition, EM fields from adjacent coherently phased antennas can interact constructively, resulting in maximum temperature rise entrally in an array rather than adjacent to the antenna surface.

 

 In recent years, interstitial US has emerged as the most controllable interstitial heating approach, demonstrating the potential to adjust power deposition in tissue axially along the implant length, radially into tissue, and directionally around the implant. Small tubular piezoelectric transducers with lengths on the order of millimeters may be aligned in a linear array to match any tumor length. The cylindrical elements radiate US pressure waves in the frequency range of 3 to 10 MHz. Penetration at these frequencies permits implant array spacing of ≥2 cm with better temperature uniformity than with other interstitial power sources, due to improved penetration, directional control of ultrasound radiation, and independent power control of each segment of tumor-length arrays.

 

 Although considerable effort has gone into the development of interstitial heating technology, interstitial hyperthermia has lost its popularity due to decline in brachytherapy use and nonuniform distributions. Becasuse of the excellent localization, the same technologies are seeing rapidly expanding use for thermal ablation procedures, wihch have a larger range of therapeutic temperature.

 

#. Determination of Thermal Dose

 Accurate real-time thermal dosimetry has been the limiting factor in the delivery of quality hyperthermia treatments. Temperature distributions are highly non ununiform and cannot be predicted accurately with present bioheat transfer modeling because the parameters used as input do not accurately model the spatial and temporal variation of actual properties. Clinical measurements of complete three-dimensional(3D) temperature distribution have not been technically feasibel because most treatment have been monitored with only a limited number of invasive temperature measurements.

 

 Numerous publications report correlation of various temperature-related parameters to clinical response based on sparse sampling with invasive probes. This effort has not been definitive in identifying quantitative parameters for prospective delivery of quality hyperthermia treatments. Minimum tumor temperature is one of the most-quoted prognostic temperature descriptors. Biologically, it is reasonable to expect that treatment outcome will be associated with minimum temperature attained by all tumor cells. However, the true minimum of a distribution is never recorded with typical sparse sampling. Reconizing this limitation, investigators have looked at surrogated for biologically significant minimum temperature, such as the 90th-percentile temperature(T90) or the temperature eceeded by 90% of measured points. Although they are less sensitive to clinical outcome than actual minimum tumor temperature, some investigators have found these surrogates to be reliable predictors for prescribing efffective hyperthermia prospectively. Going forward, it is expected that as our knowledge of minimum tumor temperature improves, parameters like fraction of tumor achieving temperatures greater than various indices(e.g., T40℃, T41℃, T42℃,etc.) will be increasingly predictive for response.

 

 Typical clinical treatments are characterized with invasive thermometry, which smaples 8 to 16 points continuously during heating. In some cases, 30 or more points are sampled using multisensor probes or by mechanically translating sensors through implanted catheters. Thermal mapping to achieve a higher density of measured points is now a quality assurance requirement for multi-institutional clinical trials. Even so, the number of measurement points is extremely low and placement within the tumor ill defined. Strategies for choosing the spatial sampling of measurement points are highly variable, ranging from uniform spacing, to centrally or peripherally enhanced spacing, to just randomly placed sensors. Invasive thermometry has been firmly established as a basic requirement for monitoring and control of slinical hyperthermia, but there are serious limitations of random sampling as regards assurance of minimum prescribed dose. Clearly, future clinical HT will benefit more complete noninvasive volumetric thermal dosimetry.

 

The obvious limitations of invasive thermometry for characterizing complex in vivo temperature distributions have stimulated the development of noninvasive thermometry approaches, which include backscatter ultrasound, electrical impedance tomography, active microwave imaging, passive microwave radiometry, and magnetic resonance thermal imaging(MRTI).

 Of these approaches, the volumetric average reading possible with microwave radiometry show most promise for control of superficial heat applicator distributions, and the complete 3D characterization of tissue temperature distributions possible with multiple-slice MR thermal imaging shows most promise for monitoring and control of deep hyperthermia. Although several MR parameters are sensitive to temperature, proton resonance frequency shift-based MRTI has been shown to provide excellent temperature sensitivity and stability, with approximately 0.3℃ to 0.5℃ resolution in 1-㎤ volumes in phantom studies and 0.5℃ to 1℃ per 1 ㎤ in clinical monitoring of deep-tissue hyperthermia. Although research is ongoing to correct image distortion and motion artifacts caused by breathing, organ movement, or circulating cooling water, MR thermal imaging has clearly demonstrated its usefulness for monitoring clinical HT.

 

 Numerical modeling of SAR and thermal distributions has experienced a dramatic improvement in accuracy in recent years due to advances in EM and thermal modeling software, tissue segmentation programs, and available computing power. Because the electrical properties of tissue are known with more certainty than thermal properties, current numerical modeling approaches are more accurate in calculating SAR patterns than steady-state temperature distiributions.

 Investigators are reporting increasingly accurate results of patient treatment planning from SAR optimization based on high-resolution, patient-specific anatomy derived from computed tomography or MRI scans. Current software can complete 3D SAR distribution calculations in a matter of hours. Until recently, validation of calculated SAR patterns has been performed primarily in simplified phantom models, using infrared thermography, Schottky diode sheet array, or fieroptic thermal monitoring sheet array measurements of 2D cross-sectional planes in split phantom layered tissue models or 3D scanning of electric field probes through liquid muscle tissue-equivalent materials.

 With the increasing availability of MR thermal imaging to provide complete noninvasive 3D characterization of thermal distirbutions, it is now possible to characterize heating patterns in vivo with high resolution during heating. This dynamic vision allow validation of preplanned SAR distributions or real-time adjustment of heating parameters to rectify errors in preplanned heating from unexpected tissue properties or misalignment of patient position in the applicator.

 

 Along with the advent of more controllable heat applicators like the Academic Medical Center three-ring waveguide array and the BSD Sigma Eye three-ring dipole array, pretreatment optimization of phase and amplitude drive parameters has become increasingly complex. Even with high-resolution thermal feedback, rapid manual adjustment of multiple power sources for correction for unintended hot spots is challenging.

 Thus, new approaches are under investigation that use MR thermal feedback to correct errors in preplanned power excitation parameters via real-time adjustments to the power control algorithm. The path is now clear for increasing use of noninvasive thermometry to provide high-resolution volumetric thermal feedback for real-time power adjustment, as well as adaptive correction of preplanned control algorithms. This includes increasing use of microwave radiometry for monitoring/control of superficial HT and MRTI for monitoring/control of deep HT. This new vision, combined with improving theoretical models, should lead to significantly enhanced uniformity of thermal dose, which should in turn produce large gains in efficacy of thermal therapy for cancer in the coming years.