old 24_01

Rationale for Clinical Use of Heat

The effect of heat on malignant tumors was first mentioned by Hippocrates. In 1866, Busch [ref: 30] described the disappearance of a soft tissue sarcoma after high fever in a patient with erysipelas. Later, Coley [ref: 40] induced fever by injecting bacterial toxins; Warren [ref: 412] and Westermark [ref: 422] used localized hyperthermia to produce tumor regression in patients.

In the past 20 years interest has been rekindled in the clinical application of heat, encouraged by biologic reports that there may be a significant advantage in the use of heat combined with radiation and cytotoxic drugs to enhance the killing of tumor cells. [ref: 57,59,65] The clinical use of heat has been hampered by a lack of adequate equipment to effectively deliver heat in deep-seated and even large superficial lesions and of thermometry techniques that provide reliable information on heat distribution in the target tissues. However, significant progress has been made. [ref: 134,187]

In vitro and in vivo biologic experiments suggest that heat may be more damaging to tumors than to normal tissues for several reasons: chronically hypoxic cells may have an increased sensitivity to heat [ref: 59] (they are at least as thermosensitive as oxygenated cells); cells with a low pH (less than 6.8) that are metabolically deprived (as in a tumor) are more heat sensitive; heat affects cells in S phase, which are known to be resistant to irradiation [ref: 349,385]; and blood flow in the tumor is reduced. [ref: 58,59,195] Heat causes a greater degree of mitotic delay than radiation, and this factor may affect the distribution of cells in the cell cycle after exposure to heat or radiation. [ref: 60,157] The sensitivity of hypoxic cells to heat is complicated by the possible association of low oxygen tension with nutrient deficiency or reduced pH. As Dewey and associates [ref: 59] pointed out, the response of the tumor may be affected by physiologic changes associated with lowering of the blood flow and oxygen tension produced by hyperthermia. The differential heat sensitivity of tumors is a consequence of tumor physiology, with nutrient deprivation [ref: 91,171,305] and lower pH [ref: 108] being the main contributing factors and not a consequence of the intrinsic state of malignancy of the cells. [ref: 111,112,121,194,273,283]

The biologic rationale for combining hyperthermia and irradiation in the treatment of cancer rests in two biologic mechanisms: radiosensitization and direct hyperthermia cytotoxicity. It can be hypothesized that hypoxic cells in the center of a tumor are relatively radioresistant but thermosensitive, whereas well-vascularized peripheral portions of the tumor are more sensitive to irradiation. [ref: 277,282] This supports the use of combined radiation and heat; hyperthermia is especially effective against centrally located hypoxic cells, and irradiation eliminates the tumor cells in the periphery of the tumor, where heat would be less effective. In experiments on a transplanted mammary carcinoma, Overgaard [ref: 274] reported no cures with 16 Gy (single dose), 22% with heat alone (43 degrees C, 60 minutes), and 77% when the two modalities with the same parameters were applied.

Biologic Aspects of Hyperthermia

Despite the publication of numerous observations of heat-induced alterations of subcellular structures and systems, [ref: 313] no consensus concerning the molecular mechanisms of cell kill has emerged. Most commonly postulated mechanisms involve damage to three major cellular structures:

Plasma membrane. Hyperthermia produces numerous alterations in the plasma membrane, including effects on membrane (e.g., receptor proteins and transport proteins), [ref: 31,180,206,304,362,407] extensive bleeding, [ref: 24,46] and regions of altered cholesterol content. [ref: 308] The involvement of damage to the plasma membrane in the lethal event is supported by the observation that membrane-active agents (local anesthetics) [ref: 435] and aliphatic alcohols [ref: 199] act synergistically with heat, and cell kill by the action of these agents alone is strikingly similar to the action of heat by itself. [ref: 200] In addition, several studies suggest a relationship between membrane lipid composition and cell kill. [ref: 117,143,176,177,250,435]

Cytoskeleton/cytosol. In tissue, culture cells contain stress fibers resulting from bundling of actin-containing microfilaments. Within 5 minutes of exposure of Chinese hamster ovary (CHO) cells to 45 degrees C, 90% of cells do not contain observable stress fibers. [ref: 109] Spindle microtubules are disorganized and disassociated (15 minutes) on exposure of CHO cells to 45 degrees C, [ref: 46] suggesting that this effect may be responsible for the increased thermal sensitivity of mitotic cells. The vimentin-containing intermediate filaments collapse on heat exposure to form a pernicular cap. [ref: 96,386,421] Most of the effects of heat on the cytoskeleton are reversible with postheat incubation at 37 degrees C. Hyperthermia induces alterations in both the structure and function of numerous cytoplasm elements: mitochondria, [ref: 421] lysosomes, [ref: 151,282] and protein synthesis apparatus. [ref: 137,224] Hyperthermia induces disruption of respiration and glycolysis, [ref: 36,68,233] which appears to be related to morphologic changes in mitochrondria. Polysomes are destroyed, [ref: 133,421] and protein synthesis is inhibited at the incubation step [ref: 75,137,284] and may be mediated by phosphorylation of initiation factors. [ref: 265] The association of polysomes and initiation factors with cytoskeletal structures is altered, [ref: 150] indicating the possibility of a functional correlate between heat-induced cytoskeletal changes and protein synthesis.

Nucleus. In addition to inducing a number of structural changes, hyperthermia alters or disrupts many nuclear functions. [ref: 313] One of the distinct substructures within the nucleus is the nucleolus, a heat-sensitive organelle, which undergoes marked changes at heat exposure that leave cytoplasmic organelles largely unaffected. [ref: 348] An increase in the overall median protein content is observed in nuclei isolated from cells exposed to hyperthermia. [ref: 315,387] The increased nuclear protein content is a large and rapid effect [ref: 315] and appears to result from the heat-induced association of specific proteins (including heat shock protein 70 [hsp70]) with the nuclear matrix [ref: 417] and the nucleus. [ref: 267] When cell kill, as a surviving fraction, is plotted as a function of excess nuclear protein content immediately after hyperthermia, a linear quadratic-type correlation curve is obtained, which holds even when cells are thermal-sensitized or thermal-protected by chemical modifiers. [ref: 337] When time of association between these proteins and the nucleus is included, the correlation becomes a simple exponential one (linear on the semilog plot) and applies for all conditions tested. [ref: 158] Additional studies implicate the association of excess proteins in the nucleus with inhibition of DNA replication [ref: 413,425] and DNA repair. [ref: 37,228,414] DNA replication is inhibited at all defined steps in the process: initiation, elongation, and assembly of replicated DNA into mature chromatin structure. However, these steps have different thermal sensitivities. [ref: 313]

A model for the mechanism of heat-induced cell kill based on the foregoing considerations proposed by a number of investigators [ref: 57,313,314,426] is as follows: disruption of critical plasma membrane structures (e.g., plasma membrane-cytoskeleton attachment points and protein channels); collapse of the cytoskeleton toward the nucleus; absorption of protein onto the nuclear matrix; disruption of nuclear functions, possibly involving inhibition of DNA supercoiling changes; and damage to critical nuclear structures.

When mechanisms of cell kill are considered, at least three modes of cell death can be identified. One mode of cell death involves apoptosis. The other two do not. One of these modes (to which the foregoing model applies) appears to occur after either at least one S-phase transit (not necessarily complete) [ref: 217] or at least one mitosis. [ref: 37] The other mode is rapid necrosis without cell-cycle progression. Part of the confusion regarding different mechanisms of cells arises because different workers were studying different modes of cell death without characterizing them. This situation is particularly telling when it is realized that different modes can occur for the same cell type. For example, L5178Y cells heated at 43 degrees C die by apoptosis, but when heated at 45 degrees C, they die by rapid necrosis. [ref: 218] Also, the cell death mode in CHO cells varies with the level of cell killing at the same temperature. [ref: 410] Regardless of the different modes, cell kill resulting from heat exposure can be represented typically as a surviving fraction plotted as a function of heating time (Fig. 24-1).



The resulting survival curves can be analyzed by either the target theory equation, S/So = 1 (1 - e**D/Do**n, or the linear-quadratic equation, S/So = e**alphaD betaD2. The former relationship has formed the basis of the current concept of thermal dose. [ref: 169]

Thermal Dose

The concept of thermal dose has arisen from attempts to convert heat exposures at various temperatures to equivalent time at a reference temperature (usually 43 degrees C) based on the biologic effectiveness of the actual temperature. [ref: 138] Sapareto and Dewey [ref: 324] developed a thermal dose concept that converts a thermal exposure to an equivalent exposure at an arbitrarily chosen reference temperature of 43 degrees C. This phenomenon has been observed in vitro and in vivo. [ref: 41,59,100,102,324]

These time-temperature conversions were shown by Dewhirst and associates [ref: 66] to be good prognostic indicators for spontaneous tumor treatment in dogs and cats when the thermal dose calculated for the coolest part of the treated tumor was used. These results indicated that equivalent minutes of exposure were the best predictor of long-term tumor response. However, Dewhirst and colleagues [ref: 67] stressed the importance of three-dimensional (3-D) dose mapping. A workshop on thermal dose demonstrated that a number of factors must still be evaluated and their importance resolved before this concept is generally accepted. These factors include the temperature of the transition or breakpoint, the R value below the breakpoint, the effect of step-down heating, the effect of thermotolerance, the effect of the interaction and radiation on cell survival, and the importance of blood flow and other physiologic factors. The workshop concluded that "any attempt to plan thermal treatments based on any of these dose concepts or to predict clinical response from a calculated dose is very inappropriate and premature." [ref: 322,323] However, Sapareto and Dewey's model is the most practical method available for the comparison of clinical treatments. Therefore, the method should be clinically tested to fully determine its potential usefulness.

Thermotolerance and Step-Down Heating

It has been frequently reported that mammalian cells are substantially more resistant to heat after prior heat exposure. [ref: 135] Henle and Leeper [ref: 136] noted that cells initially exposed to 45 degrees C became resistant to subsequent exposure to 45 degrees C if allowed a 10- to 20-hour period at 37 degrees C between treatments (Fig. 24-2).



Thermotolerance is a transient phenomenon and thus does not represent a selection of genetically resistant cells, which occurs at too low a frequency to account for this phenomenon. [ref: 131] The mechanism for thermotolerance is not known; however, protein or RNA synthesis must occur before thermotolerance can develop. [ref: 189,327]

Both Henle [ref: 134] and Li and associates [ref: 198] have indicated that prior exposure to temperatures above 43 degrees C sensitizes cells to lower temperatures. This phenomenon has been termed "step-down heating." Also, Li and colleagues [ref: 198] have shown that thermotolerance is inhibited for several hours immediately after exposure to 45 degrees C.

Studies indicate that thermotolerance is accompanied by enhanced ability to repair (restore) certain types of heat-induced cellular alterations. The following effects are repaired more rapidly in thermotolerant cells: nuclear localization of heat shock protein (hsp70), [ref: 267] disassembly of the cytoskeleton, [ref: 176] increased nuclear protein binding, [ref: 158] and inhibition of certain types of DNA repair. [ref: 431] Further studies may show that enhanced ability to repair heat-induced damage contributes more significantly to heat-resistance than does protection from damage.

Heat Shock Proteins

Thermotolerance appears to be closely related to the induction of a class of protein polypeptides of molecular weights of 25 to 110 kDa. [ref: 184,202,374] These heat shock proteins have been well characterized and occur as a result of gene transcription induced by thermal stress in Drosophila melanogaster [ref: 229] and most other living systems [ref: 337]; however, their function is unknown. A good correlation exists between the increased induction and degradation of these constitutive proteins and the induction and decay of thermotolerance (Fig. 24-3), whether induced by heat shock or other toxic stress phenomena. [ref: 203]



Particular interest has been generated by hsp70, [ref: 201] which migrates to the nucleus during heat shock. [ref: 267,420] Different types of thermotolerance appear to be associated with a period of enhanced synthesis of hsp70 and the period during which cells have elevated levels of hsp70. [ref: 185] Although it appears that increased levels of hsp70 are sufficient to cause increased heat resistance, it is becoming clear that they are not necessary. A series of heat-resistant cell lines from radiation-induced fibrosarcoma (RIF) tumors do not express increased levels of hsp70. [ref: 3] Also, transfection of cells with genes for hsp27 leads to increased heat resistance. [ref: 35] Although the cellular response to thermal stress appears to be one of the most conserved biologic mechanisms in nature, [ref: 337] the function of the various heat shock proteins remains by and large a mystery despite numerous eloquent studies.

Long-Duration Moderate Hyperthermia

It was long believed that chronic tolerance contraindicated treatment of tumors with moderate (41 to 42 degrees C) temperatures for long durations (i.e., more than 48 hours) with any effective cell killing. However, recent results have caused workers in the field to rethink this assertion. Mackey [ref: 215] found that, unlike CHO cells, which arrest in the G1 phase of the cell cycle and maintain chronic thermotolerance, HeLa cells progressed into S phase and were killed. Although there is debate regarding the role of cell-cycle progression, the fact that human tumor cells can be killed under these conditions has been confirmed by other groups. [ref: 9,303] The mode of cell death induced by this treatment is not apoptosis, but rather appears to be an alternative mode to those described earlier. [ref: 380] Thus, it appears that human tumors, where practical, might be treatable by this method. However, a word of caution should be advanced: some human colon carcinoma cell lines have been shown to survive and proliferate at 41.1 degrees C. [ref: 432] Thus, any clinical trial with moderate hyperthermia alone should include measurement of the thermal resistance of the tumor cells in question.

Heat Interaction with Radiation

Several good reviews are available on this subject. [ref: 59,65,275,375] The first and most generally observed phenomenon is that heat radiosensitizes cells. [ref: 59,326,363,364] Most reports note that the maximum increase in the slope of the radiation survival rate curves is 25% and that cells in S phase are more radiosensitized by heat than are cells in G1. [ref: 325]

The cause of this radiosensitization has not been firmly established; however, it is believed that the accumulation of proteins in the nucleus, which bind to the nuclear matrix after heat treatment, prevents the cell from repairing radiation damage. [ref: 37,228] The ability for enzymatic excision of radiation-induced thymine damage is inhibited in chromatin isolated from heated cells even when the repair enzymes are obtained from control cells, whereas the repair enzymes from heated cells were able to excise damage in chromatin from control
cells. [ref: 415] A correlation exists between the amount of excess nuclear protein and the inhibition of exogenous nucleases [ref: 416] and digestion of DNA in chromatin, which suggests that access to the DNA damage is blocked, thereby inhibiting repair. [ref: 159,416]

Another factor of possible clinical relevance is that cells in G1 are less sensitive to heat than are cells in S phase, whereas the opposite is true for cellular sensitivity to radiation (Fig. 24-4).




This effect often has been cited as one of the principal factors contributing to the "biologic rationale" for the clinical use of hyperthermia. In fact, a marked complementary synergism across the cell cycle was observed when heat (45 degrees C for 30 minutes) and radiation (4 Gy) were combined. [ref: 175] It can be argued that in a tumor the proliferating cells are likely be a small fraction of quiescent cells that are nutrient deprived [ref: 171] or at low pH [ref: 107] and are heat sensitive, and that the heat sensitivity is not affected by acute lypoxia. [ref: 106]

Enhanced cell kill by irradiation and heat has been defined as the thermal enhancement ratio (TER), expressed as follows:

Equation 1

The relationship of the TER in the tumor and the normal tissues is called the therapeutic gain fagtor (TGF):

TGF = TER in tumor/TER in normal tissue

Both TER and TGF should have values greater than 1 to have a positive therapeutic implication.

Urano and associates [ref: 395] and Stone [ref: 364] have reported a greater enhancement of early damage than late damage with a combination of irradiation and heat in mice.

Timing of Irradiateon and Heat Administration

A number of groups have considered how best to sequence heat and radiation. Dewey and co-workers [ref: 59] have hypothesized that if heat is delivered 3 hours before radiation, cells with a low pH will have minimal aoility to repair heat damage and therefore may be sensitized to the effects of subsequent radiation. Hill and Denekamp, [ref: 144] Field, [ref: 99] and Overgaard [ref: 276] strongly suggest that when preferential (selective) heating of the tumor exists in relation to normal tissues, it is best to administer the two modalities simultaneously. [ref: 47,87,167,173,237,279,281,312] However, when the temperature in the tumor and in the normal tissues is the same, some animal model studies suggest that the optimum therapeutic gain occurs when the heat is delivered 4 hours or more after exposure to radiation (Fig. 24-5).



These animal studies used tumor temperatures that may not be practical in humans. In human tumors treated with radiation and heat, the impact of timing of both modalities has varying influence on the effects of the treatment on the tumor or the normal tissues. [ref: 4,5,280]

The interaction between heat and radiation does not exhibit the same time-temperature relationship as does heat alone. This is proven by both in vitro and in vivo studies. [ref: 186,326] A maximum effectiveness is seen near 43 degrees C for simultaneous treatment, suggesting that there may be an optimal temperature for combined-modality treatment. However, when the modalities were given sequentially, little effect of temperature variations was found. [ref: 275] Mittal and co-workers [ref: 230] conducted experiments in transplanted RIF-1 fibrosarcomas in the flanks of C3H mice. Tumors were treated with fractionated x-rays (4 Gy twice weekly x 10) alone or in combination with heat (radiofrequency [RF] currents, 43 degrees C, twice weekly). Animals treated with irradiation alone or with heat and irradiation delivered with sequential fractionation (all heat sessions given before or after the radiation exposures) exhibited a 20% cure rate. In the animals treated with simultaneous combination of both modalities (heat delivered same day as irradiation, immediately after x-ray exposure), the cure rate was 70%. Another aspect of heat-induced radiosensitization comes from the long-duration moderate (41 to 42 degrees C) hyperthermia studies described earlier. Three separate groups [ref: 9,216,303] have shown that long-duration moderate hyperthermia produces thermal enhancement. Raaphorst and colleagues [ref: 303] applied long-duration moderate hyperthermia, combined with low dose-rate (LDR) irradiation, to produce thermal enhancement.

These considerations open the possibility of a wide variety of methods to combine heat and irradiation.

Interaction of Heat with Chemotherapeutic Agents

Hahn and associates [ref: 120,122] and other clinicians [ref: 139,262] have reported on the interaction of heat with a variety of cytotoxic agents. The type of drug, dose, temperature, and time of administration of the agents are important factors in determining cell kill by combination of these agents. [ref: 25,121,221] The interaction of heat and drugs in patients is not completely understood. Since Hahn's original work, [ref: 120] there have been numerous reports on this subject, and some excellent reviews have been published. [ref: 49,94,121]

Enhanced cytotoxicity of drugs at elevated temperatures is not a predictable process. The vinca alkaloids and most antimetabolites have only additive cytotoxicity with hyperthermia. [ref: 121,141] In two anticancer agents, AMSA and Ara-C, cell kill was actually inhibited at elevated temperatures [ref: 139] (B.E. Magum, personal communication). Doxorubicin (Adriamycin) and dactinomycin (actinomycin D) exhibit complex interactions with hyperthermia, and both increased killing and protection have been observed with these agents, depending on the scheduling of drugs and heat. [ref: 72,123]

A summary of interactions between anticancer drugs and hyperthermia is shown in Table 24-1. In addition, a category of agents not normally considered to be of therapeutic value at 37 degrees C shows significant killing ability at elevated temperatures. These agents include alcohols, amphotericin-B, cysteine, cysteamine, and AET (2-amino-ethyl-isothiourea).

The wide variety in mechanisms of drug cell kill precludes the idea that heat and drug interaction is a simple unilateral phenomenon. In fact, heat can enhance drug resistance in the case of doxorubicin [ref: 120] and dactinomycin. [ref: 72] As heating duration is increased, cells in culture become highly resistant to killing by either drug. This may be caused by heat-induced alteration in drug transport into the cell. [ref: 101]

The sensitizing effects of heat in combination with cisplatin have been confirmed in vivo. [ref: 73,262] Herman and co-workers [ref: 140,142] demonstrated in experimental models (murine RIF tumor) the substantial enhancement of antitumor effect obtained with the combination of radiation, heat, and cytotoxic agents (cisplatin). This hypothesis is being evaluated in prospective clinical trials. [ref: 141,142]


Physiologic Mechanisms in Hyperthermia

Microvasculature of Normal Tissue

There is great variation in the microcirculation of different tissues such as striated muscle, skin, and so on. Nevertheless, there is regularity in distribution within a specific tissue. [ref: 178] In a typical model, all exchanges between blood and parenchymal cells take place at the capillary level (microcirculation). True capillaries in normal tissues have a diameter close to that of an erythrocyte (7 to 10 micrometers).

Microvasculature of Tumors

At an early stage of tumor development, the tumor cells probably proliferate by using energy and nutrients supplied through the host's blood vessels. As the tumor grows, host vessels are occasionally incorporated into the tumor mass. As the demand for nutrients and oxygen exceeds the supply capacity of the host vessels, neovascularization in the tumor begins (formation of "buds" and, by confluence, "sprouts"). It has been suggested that certain humoral factors are important for the initiation of this process (e.g., tumor angiogenesis factor [TAF] and endothelial proliferating factor [EPF]). The capillaries formed by random fusion of sprouts are tortuous, elongated, and dilated, and they lack basement membranes. [ref: 80] A large proportion of tumor blood does not exchange with blood in the general circulation (stasis). This intermittent circulation – periods of stasis followed by resumption of blood flow -- is probably a normal feature of the intravascular transport system of neoplastic tissues. [ref: 118,404] The histologic patterns and functional status of vascular networks in malignant tumors vary, depending on the type, age, and size of the tumor. [ref: 404]

Tannock [ref: 383] suggested that the longer turnover time of endothelial cells (their slower proliferation relative to that of neoplastic cells) accounts for the decline in vascular density. Reduced vascular density together with the sluggish perfusion of blood through the capillaries may account for the decrease in total blood flow. [ref: 352,404] In general, tumor blood flow is vigorous in the periphery and sluggish in the center. [ref: 305,351]

Hyperthermia and Normal Tissue Microcirculation

Song and associates [ref: 353] observed that the blood flow of skin overlying the tumor and of muscle near the tumor is more than twice that of the skin and muscle far from the tumor. They attribute this phenomenon to inflammatory processes near the tumors. A significant increase in the blood flow occurred in skin and muscle both near and far from the tumor when heated to 43 degrees C for 1 hour. [ref: 350] It should be noted that the magnitude of increase was higher in the normal tissues adjacent to the tumor than in the tissues far from the tumor.

The dynamic changes of skin and muscle blood flow are both time dependent and temperature dependent. Peak blood flows are different for various times and temperatures; similar trends have been observed in muscle blood flow. Blood vessels in mouse gut were particularly sensitive to heat: a temperature of 41 degrees C for 1 hour to the lower body resulted in a sizable reduction in the visible venous tree. [ref: 95] The microcirculation of connective tissue was investigated in rats [ref: 61] and in rabbit ear chambers. [ref: 74]

Hyperthermia and Tumor Microcirculation

Several excellent reviews have been published on the subject of hyperthermia and tumor microcirculation. [ref: 81,91,305,370,405] The complex relationship between temperature, exposure time, and physiologic response for tumors and for normal tissue has been depicted in a simplified diagram in Fig. 24-6.



The data are based almost entirely on rodent studies (mostly murine models). Tumor vasculature seems to be less able to show vasodilation to elevated temperature and is more heat labile than the vasculature of normal tissue. Both tumor and normal tissues increase their blood flow as a result of hyperthermia exposure. However, the vasodilatory effect in normal tissue may be greater. Even in normal tissues, however, if the temperature exceeds 46 degrees C for a few minutes, vascular destruction occurs, and this leads to direct tissue damage due to ischemia. The temperature threshold for this type of vascular destruction in tumors (rodent models) is lower than in normal tissues (40 to 43.5 degrees C). [ref: 17,69,80,87,93,306,350,379,406,408] Confirmatory data for a similar effect in human tumors are yet to be provided. Song and associates [ref: 354] noted that vascular damage after hyperthermia is progressive even after the heating is completed. This damage makes the tissue highly sensitive to a second heat treatment given a few hours after the first. Extremely important information was provided by Reinhold and van den Berg-Block, [ref: 306] who studied the response of microcirculation to hyperthermia in five different tumors, growing in "sandwich" observation chambers in
the back of the rat. Their conclusions were as follows:

The various tumors required significantly different exposure times for inducing 50% stoppage of the tumor microcirculation (ST50). This seems to indicate that differences in the characteristics of the tumor cells were more important for causing microcirculatory stoppage than in the sensitivity of the cells of the blood vessels.

An increase in surface (volume) was observed in all four tumors examined; however, the rate of increase was significantly different among various tumors.

The relative velocity of the erythrocytes in selected capillaries in the tumors decreased as a result of hyperthermic treatment and was probably related to the tumor-specific ST50.

Several different mechanisms have been proposed to explain the vascular events occurring in tumors during hyperthermia (Fig. 24-7). [ref: 74]



However, none of the proposed single causative factors has been fully explained.

Hyperthermia and Intratumor pH

The pH of arterial blood is 7.4 and that of venous blood and interstitial fluid is 7.35. Intracellular pH usually ranges from 6.0 to 7.4 in different cells, averaging about 7.0. Studies show that no significant difference exists between the intracellular pH of normal cell lines and that of their malignant counterparts. [ref: 156]

Song and co-workers [ref: 353] and Bicher and associates [ref: 19] have shown that hyperthermia triggers an immediate and significant decrease in the pH of tumors. Bicher reported that when heating was terminated, the pH increased to 6.78 but decreased to 6.5 to 6.6 when the tumors were reheated.

Ryu and associates [ref: 317] observed that the lactic acid content in mouse tumors significantly increased with heating. Streffer and van Beuningen [ref: 370] also reported that hyperthermia caused an increase in the amount of both lactic acid and beta-hydroxybutyric acid in mouse tumors. The pH of human tumors is significantly lower than that of normal tissue [ref: 424]; there is no significant difference among different types of human tumors in this respect. [ref: 423] Also, human tumor pH appears to consistently increase after treatment with a combination of localized hyperthermia and radiation therapy. [ref: 401,423]

Basic Principles of Physics and Instrumentation

Principles of Power Deposition and Heat Transfer

The physical agents used for power deposition in clinical hyperthermia are electromagnetic (EM) fields at very high microwave (MW) frequencies (300 to 2450 MHz), low-frequency microwaves (60 to 120 MHz), EM fields at radiofrequencies (0.1 to 27 MHz), and ultrasound (US) at frequencies of 0.3 to 10 MHz. The main characteristics of these modalities are summarized in Table 24-2.

Electromagnetic Power Deposition

Temperature elevations with RF electric fields may be produced through conductive, dielectric, or inductive heating of tissue. Conductive or resistive heating refers specifically to heating with RF currents driven between pairs of external electrodes in electrical contact with body surfaces (Fig. 24-8).



The resulting power deposition is effected through the movement of free conduction charges through the medium in response to an applied electric field. With dielectric or capacitive heating, the power is deposited in tissue through the interaction of the electric fields produced by electrodes not in contact; this mechanism is usually called dielectric polarization (Fig. 24-8). Inductive heating occurs when energy is lost to tissue by eddy currents induced by an RF magnetic field produced by an external coil (current loop) applicator adjacent to or surrounding the tissue (Fig. 24-9).



Inductive applicators can be considered to be magnetic dipoles. [ref: 125]

The average power deposition in tissues during EM heating is proportional to the square of the magnitude of the electric field times the tissue electrical conductivity. The most common concept used to quantify both the magnitude and the distribution of EM energy absorbed in tissue is the specific absorption rate (SAR), [ref: 125] which is the rate of energy absorbed in a given tissue volume normalized by the mass in that volume. For tissue with density, the SAR is given by

Equation 2

The SI units of SAR are W/kg. Although the above equation appears to be simple, the determination of the electric field vector (E) in heterogeneous tissue is generally very complex and must be calculated solving Maxwell's equations using advanced computational techniques subject to pertinent boundary conditions. Boundaries of particular importance in hyperthermia are those with sudden changes in dielectric constants (e.g., fat-muscle interfaces) that are perpendicular to the direction of the principal component of the electric field vector. The electric field component parallel to the interface is continuous but the perpendicular component is governed by

e(1)E(per1) = e(2)E(per2)

where e(1) and e(2) are the complex dielectric constants of tissue 1 and tissue 2, respectively, and E(per1) and E(per2) are the magnitude of the perpendicular component of the electric field in tissue 1 and 2, respectively. In short, sudden changes in the dielectric constant imply sudden changes in the electric field; hence, even greater changes of SAR are present at these interfaces. This physical phenomenon has important clinical consequences that will be discussed later. For further details of the physical characteristics and basic principles of EM energy absorption in hyperthermia, see the review by Hand. [ref: 125]

Ultrasonic Power Deposition

Mechanical waves of frequencies above the audible range are called ultrasound. US is usually generated by the conversion of electric voltages into mechanical motion in a piezoelectric crystal (Fig. 24-10).



The physical characteristics of US wave propagation are mostly determined by the characteristics of the transducer and by the acoustic properties of the tissues involved. An important advantage of US is its short wavelength in soft tissues (0.15 cm at 1 MHz), which allows deeper penetration, negligible beam dispersion, and great flexibility in beam forming and manipulation when compared with EM heating. A typical example of this can be seen in Fig. 24-10, which illustrates a spherically curved US transducer generating a focused beam.


Penetration depth with US is not a function of applicator size but is strongly dependent on frequency. Another advantage is that US transducers can be manufactured in a wide range of sizes and shapes. This is especially helpful when multiple individually powered crystals are grouped to form arrays. Arrays can be used to electrically focus and scan US beams. Mechanical scanning of single or multiple transducers also offers many advantages for both superficial and deep hyperthermia.

In soft homogeneous tissue, the US energy is attenuated exponentially. Assuming linear plane wave behavior, the US intensity I(z) at some depth z can be written as

I(z) = I(0)e**-2muz

where I(0) is the intensity at the tissue surface (usually the skin) and mu is the acoustic attenuation coefficient of the medium with SI units of 1/m. The units of intensity are W/m**2.

From the intensity distribution one can calculate the power deposited in tissue as

Q(z) = 2alphaI(z)

where alpha is the acoustic absorption coefficient. Notice that this expression does not account for acoustic scattering, which contributes to the total energy attenuation because it is usually assumed to be negligible. Q has units of W/m**3.

As in the case with EM SAR, the US power deposition Q is usually complicated, and it must be estimated using computer programs or measured in water phantoms. US propagation is also affected by tissue interfaces, especially if there is a sudden change in acoustic impedance Z (Z = density times speed of sound) across a given interface, which causes them to undergo partial reflection and transmission. Tissue interfaces of clinical relevance are tissue-air and soft tissue-bone interfaces. At tissue-air interfaces a sound beam is completely reflected, causing the power deposition to double upstream of the interface; consequently, high temperatures may be induced. At tissue-bone interfaces most of the beam is reflected, but some is transmitted into the bone. Because US attenuation in bone is about an order of magnitude greater than in soft tissues, all of the transmitted energy is absorbed in the first few millimeters. Very high temperatures at soft tissue-bone interfaces are possible and may result in treatment-limiting pain. The reflective/transmission fraction, as well as the angle of reflection and transmission, is dependent on the angle of incidence that the US beams makes with the interface.

The coupling bolus in US hyperthermia is usually degassed water. Acoustic coupling gel, commonly used in diagnostic US imaging, is usually used to couple the bolus membrane to the treatment site. Care must be taken to avoid air bubbles trapped between the membrane and the patient's surface because these can block US propagation and contribute to inadequate heating.

For a more detailed treatment of the physical characteristics and basic principles of US energy absorption in hyperthermia, see the review by Hynynen. [ref: 153]

Hyperthermia Devices

Superficial Hyperthermia Devices

Superficial hyperthermia traditionally refers to the treatment of superficial tumors of less than 6 cm in depth -- most commonly no deeper than 3 cm. The most commonly used clinical devices have been radiative EM applicators operating at 915 MHz and US applicators operating at between 1 and 5 MHz. Heating is induced by coupling externally applied EM or US waves into tissue. Many technologic advances have been made in the past 15 years in response to clinical needs. Temperature uniformity has been the most important requirement in applicator design and clinical delivery techniques. It is widely accepted today that to achieve this requirement applicators need to have the ability to locally vary their power deposition patterns in real time. This requires multisource or scanning source(s), and, in many cases, site-specific applicators in combination with extensive thermometry and temperature feedback control. Unfortunately, it is also clear that, even with highly advanced systems, the most important limiting factor may be implementing adequate thermometry in the clinic.

Superficial Electromagnetic Hyperthermia Devices

The simplest waveguide MW applicator consists of a hollow metallic rectangular box with an open end. The metallic cavity connects to the outer conductor of a coaxial feed line. The radiating antenna, which is connected to the feed line inner conductor, is positioned inside the cavity parallel to the aperture and perpendicular to one of the walls. The physical dimensions of the waveguide are carefully chosen to maximize radiative efficiency. A waveguide applicator has its dominant electric field component perpendicular to the propagation direction. This implies that the dominant electric field is parallel to fat-muscle interfaces near the skin. This is one advantage of waveguide radiators in comparison with capacitive EM systems, which have a dominant field direction that is perpendicular to the interface between subcutaneous fat and underlying muscle. This field orientation may produce excessive heating of the subcutaneous fat.

Based on theory, physical measurements, and careful study of clinical data, it has been established that the maximum depth of therapeutic heating achievable with 915 MHz MW applicators is no more than 3 cm. [ref: 34,71,253,291,369] It is important to point out that, as depth increases, the volume receiving adequate SAR (SAR > 25%) decreases, and the areas of the 25% (or even 50%) SAR contours are significantly smaller than the aperture of the applicator (Table 24-3). The maximum SAR is induced on the surface. Skin cooling is often used to avoid undesirable high temperatures on the skin. Other important factors affecting SAR distributions are bolusing techniques, thickness of the overlying fat layer, patient contour, electric field orientation, and applicator size. [ref: 34] Even for a given applicator these factors can dramatically change SAR patterns. Another important clinical consideration is the possible concentration of fields at the aperture corners that may produce undesirable hot spots on the skin surface. This is of particular concern with air-coupled applicators.

The higher the water content of the tissue, the higher the rate of energy absorption from EM fields. Therefore, microwaves are far more penetrating in fat than in muscle. Penetration also decreases with increasing frequency. In practice, however, the decrease in penetration over the 300- to 1000-MHz range is far less pronounced than is suggested by plane wave analysis and can almost be considered negligible if the size of the applicator is to remain practical (less than 20 x 20 x 20 cm). Some applicators have special geometric designs or are loaded with dielectric materials to reduce physical size at higher frequencies.

The drawbacks of the first-generation MW machines were recognized early and motivated a great deal of research and development. The design requirements for most of the second-generation applicators have been (1) to induce more uniform SAR/temperature fields across the applicator aperture, (2) to increase the effective heating area, (3) to be able to treat extensive superficial malignancies (e.g., areas of greater than 100 cm**2), (4) to implement or improve power deposition control, and (5) to reduce applicator size.

Another common applicator is the microstrip antenna, which is a transmission line consisting of a thin planar dielectric sandwiched by a metallic ground plane on one side and a radiating metallic pattern on the other. Samulski and co-workers [ref: 319] successfully constructed and clinically tested microstrip applicators having Archimedian spirals as the radiating pattern. These operated at 400 to 1200 MHz with diameters ranging from 3.5 to 8.5 cm (Fig. 24-11).





These antennas, when loaded with a 1- to 2-cm thick deionized water bolus, coupled well with tissue and induced gaussian SAR distributions. They have a dominant transverse electric field component, a very desirable characteristic to avoid overheating of fat overlaying muscle. Their compact size makes them practical in the clinic and allows treatment of regions inaccessible to bulky waveguide applicators. A device with two spirals that are mechanically scanned in a circular reciprocating path with independent power control to each microstrip has been tested clinically. [ref: 319] The ability to modulate power as a function of the antenna position along the scan trajectory allows for a higher degree of temperature control that was not previously possible with stationary single-antenna large-aperture applicators. The use of multiple microstrips in an array configuration conformable to body contours is presently under investigation. [ref: 188]

A compact design that has been used with success in the clinic is a magnetically coupled applicator typically referred to as a current sheet applicator (CSA). [ref: 116,190] A CSA consists of one wide current loop or several current loops to generate a magnetic dipole parallel to the body surface (Fig. 24-12).



The SAR from a single CSA is similar to that of a waveguide applicator. However, CSAs have a higher ratio of heating area to physical aperture area, and their dimensions can be chosen almost independently of frequency. The clinical system at the University of Arizona operated at 434 MHz. [ref: 190] As with the microstrip applicators, their small size is advantageous. CSAs are good candidates for large area arrays for treating diffuse superficial disease (Fig. 24-12).

Some investigators have tried to work at lower frequencies to gain penetration depth. The main problem in designing applicators at frequencies of less than 150 MHz is their bulky size and a loss in spatial resolution. [ref: 197]

None of the second-generation systems described have been produced commercially. A new commercial 16-element (4 x 4) MW system has been introduced. [ref: 71] Each array element is a 3.8-cm square dipole antenna waveguide aperture, making the overall aperture size 15.2 x 15.2 cm. The small element size was possible by loading with high dielectric constant material. The array operates at 915 MHz and provides independent power control to each array element. The dominant electric field is parallel to the surface to be treated. A large distensible bolus of deionized water is used to provide conformability to body curvatures. The ratio of effective heating area to aperture area is very good with the 50% iso-SAR contour following closely the outer edge of the array. [ref: 71]

Several other designs have been tested clinically. For a more comprehensive review, the reader is referred to the works of Hand [ref: 125] and Lee. [ref: 187]

Superficial Ultrasound Hyperthermia Devices

The first generation of US applicators consisted of single circular planar crystals mounted on metal housings with long water columns terminated in coupling membranes. [ref: 43,165] These operated at frequencies of 0.3 to 3 MHz with typical diameters of 3 to 10 cm. The main advantage of US was deeper therapeutic heating. Nevertheless, clinical experience indicated that single-aperture devices were no more effective than their MW counterparts [ref: 14] because the first-generation transducers possessed suboptimal spatial distribution of power deposition and often induced treatment-limiting hot spots.

Improvements in power deposition control and effective heating area and, consequently, in temperature distributions have been made by electrically or mechanically segmenting the piezoelectric crystal into individually powered elements. [ref: 266,316] The system described by Underwood and associates and later tested in vivo by Ogilvie and co-workers is now commercially available and has been used extensively in the clinic. [ref: 320] It has two applicators: a four-element (2 x 2) and a 16-element (4 x 4) array. The 3.8-cm square elements operate either at 1 or 3.4 MHz. Typical simulated and measured power deposition patterns are shown in Fig. 24-13.



The initial clinical evaluation was characterized as disappointing [ref: 320]; however, the lesions treated were relatively large and were considered not to be better treated with other devices. Simulations and clinical experience indicate that this multielement system has the ability to more uniformly heat larger volumes of tissue and thus is superior to previous commercial hyperthermia devices. [ref: 237] For a more comprehensive review of US technology, see the review of Hynynen. [ref: 153]

Interstitial Hyperthermia Devices

Interstitial techniques offer several potential advantages over external heating techniques: better heat localization, ability to heat tumors deeper than 3 cm, more uniform temperature distributions theoretically possible, greater opportunity for power deposition control due to the multiplicity of sources, and greater potential for extensive thermometry. Stauffer and colleagues [ref: 356] identified nine different approaches used in the past two decades for interstitial heating:

MW antennas operating between 400 and 2450 MHz

RF electrodes for local current field (LCF) heating operating in the range of 0.3 to 3 MHz

Capacitive coupled RF electrodes operating between 8 and 27 MHz (CC-LCF)

Internal LCF electrodes coupled inductively to external power sources (6 to 13 MHz)

Tubular US radiators driven between 5 to 12 MHz

Lasers

Hot water tubes

Resistance wire heating

Thermoregulating ferromagnetic seeds

Several other comprehensive reviews are available. [ref: 105,128,340,355-357,371,394] The present discussion is limited to devices that have been clinically tested or that show a promising future.

Interstitial Microwave Antennas

Interstitial MW hyperthermia may be produced through the use of miniature coaxial MW antennas (diameter less than 1.5 mm) inserted in insulating plastic catheters implanted in the tissue with a spacing of to 2 cm. A cross-sectional view of a typical first-generation antenna is shown in Fig. 24-14.



This simple dipole antenna is made from 50-ohm coaxial cable. The outer conductor stops at the junction, and beyond the junction the inner and outer conductor are soldered, forming an extension of the inner conductor of length hA. At an operating frequency of 915 MHz, hA is approximately 3.5 cm. The distance from the air-tissue interface to the junction is hB. When implanted, this design radiates most efficiently when hB + hA = 2 hA; hence, at 915 MHz the optimal depth of insertion is 6 to 8 cm. The heating pattern of a single dipole antenna is ellipsoidal, with the longer axis coincident with the antenna axis. In general, both the longitudinally and transverse extent of the therapeutically significant SAR depend on the depth of insertion (hB), the length of the exposed inner conductor extension (hA), and the operating frequency. If the depth of insertion is not sufficiently deep, the antenna may not be coupled well to tissues leading to poor SAR. For a given dipole MW antenna, the potentially therapeutic volume is approximately 3.0 to 4.5 cm longitudinally and 1.0 to 1.5 cm transversally. The greatest absorbed power density and therefore the greatest amount of heating occur in the immediate vicinity of the antenna junction. The therapeutic heating potential of an array of four, single-junction coaxial antennas operating at 915 MHz, placed parallel and intersecting the corners of a 2-cm square, is shown in Fig. 24-15.



The distribution is normalized to its maximum value at the center of the array. On this central plane, nearly 85% of the 4 cm**2 area bounded by the antennas lies within the 50% iso-SAR. This area is smaller at all other transverse planes. For arrays, the longitudinal heating pattern strongly depends on the depth and angle of insertion as well as the separation between antennas. A major problem with MW antennas is the so-called dead space at their distal ends, which is 1 to 2 cm long, depending on design. Theoretic studies show that even for a four-antenna array (2 x 2 cm) a minimum dead length of 6 mm exists along the central axis of the array. The clinical significance of this is that in certain situations, such as in the treatment of brain tumors, the implanted volume must be significantly larger than the tumor volume in the longitudinal dimension at the expense of normal tissue.

Interstitial Radiofrequency Heating

Interstitial heating with RF electric fields is produced by currents driven between electrically connected arrays (or electrode pairs) of metallic needles in direct electrical contact with tissue. The frequencies used in such applications are usually in the range of 0.3 to 1.0 MHz, although capacitively coupled devices have been used at frequencies as high as 30 MHz. [ref: 356] At these frequencies the voltage along an electrode can be considered constant; consequently, the axial electric field at a given radial distance between an electrode pair is uniform. The maximum electric field is in the vicinity of the needles and falls off as the inverse of the radial distance; thus, the SAR falls off as the inverse radial distance squared. For a parallel needle pair in homogeneous tissue, the minimum SAR is located midway between the needles; closely spaced arrays are required. The optimal spacing between electrodes is 1 to 1.5 cm. Heating effectiveness declines with increasing array size, increasing tissue heterogeneity, and increasing skewness of electrodes relative to each other. Despite the highly nonuniform SAR patterns, simulation models predict uniform temperature distributions. Traditionally, closed-ended needles (1.5 to 2 mm outer diameter) have been used clinically, which allows for measuring of the maximum temperatures inside the needles for sequential or simultaneous interstitial irradiation. Selected tissues can be spared by insulating sections of the needles. [ref: 167] Other important technologic improvements include the implementation of template-mounted circuit boards and electrode water cooling. [ref: 44,301] The templates greatly reduce the time of electrode connections and facilitate temperature control. RF-LCF heating has more potential for controllable heating than MW dipole antennas. [ref: 358]

Capacitive coupling of currents from insulated metallic electrodes is possible at frequencies of around 27 MHz. The current density at the outer insulating surface is very uniform and insensitive to tissue heterogeneities. Therefore, SAR distributions with CC-LCF techniques are much less dependent on electrode separation or relative orientation than with RF-LCF techniques. However, the radial fall-offs are very similar. [ref: 55] These characteristics may have a great clinical impact because they counteract the ever-present difficulties of obtaining a geometrically satisfactory implant. In addition, dual-source electrodes have been successfully developed for longitudinal SAR control, thus increasing the potential for inducing adequate hyperthermia. [ref: 182]

Interstitial Conductive (Hot Source) Heating

With this technique, no power is deposited directly in tissue. Rather, the tissue is heated by thermal dissipation within the implanted array of heated sources. One advantage to this approach is that the maximum temperature can never exceed the temperature of the sources; this eliminates concerns about unanticipated hot spots. Source spacing must be approximately 1 cm. Some of the most common approaches are reviewed below.

Inductively Heated Ferromagnetic Seeds

Here the tissue is heated by thermal diffusion of heat from the implanted ferromagnetic seeds. The seeds are heated by inducing eddy currents on the surface of the metal by means of an externally applied magnetic field. The field is produced by a coil operating below 500 kHz to minimize direct inductive heating of tissues. [ref: 357] Magnetic field strengths of up to about 4500 A/m are needed; hence, treatments must be performed in shielded rooms. The ferroseeds are made of special alloys of highly permeable materials (iron or nickel). The seeds pass from a ferromagnetic state at low temperatures through a Curie transition at therapeutic temperatures (Curie point) to a nonmagnetic state. Power absorption from the magnetic field is essentially zero above the Curie point; thus, the seeds are self-thermoregulating. The Curie point for various alloys designed for hyperthermia applications is between 40 and 55 degrees C. [ref: 26,124] The ferromagnetic transition depends on both the alloy materials and the manufacturing process. [ref: 225] The main advantages of this technique are that (1) variable-length sources without external connections and seeds with various Curie points can be used in the same implant to optimize temperature distributions; (2) maximum temperature inside implant volume cannot be higher than the maximum Curie point; (3) temperature distributions are dependent only on thermal properties and blood flow or perfusion; and (4) ferroseeds have great potential for use in permanent implants. The main disadvantage is the need for closely spaced implants with generous margins.

Resistance Wire Heaters

As with the ferroseed technique, tissue heating with high-resistance implant needles is accomplished by thermal diffusion of heat from cylindrical wire sources. The basic heating mechanism is by ohmic loss from direct electric currents flowing through high-resistance coiled wires inside insulating catheters. In comparison with ferroseeds, resistance wire heaters have the disadvantage of needing external electrical connections that complicate treatment logistics and add the risk of electric shock. This technique, however, provides the opportunity to respond to real-time changes in heat dissipation. An existing commercial system with a temperature control algorithm has been used successfully to treat brain and prostate cancers. [ref: 53]

Hot Water Tubes

A variant conductive heating technique is the so-called hot water tube technique, which is currently being explored at a few centers. [ref: 126] Tumors are implanted with tubes through which hot water is circulated and maintained at a fixed, desired temperature. This is the easiest and probably safest technique to use from the engineering point of view. The main disadvantage is the diminished longitudinal temperature control along a given tube.

Interstitial Ultrasound Hyperthermia

The use of implantable US needles is not yet a clinical reality. This field, however, is advancing rapidly and offers significant advantages over other approaches. Recent prototypes use multielement arrays of tubular piezoelectric crystals inside water-filled plastic catheters. [ref: 70] Transducers operating at 7 to 12 MHz with diameters ranging from 1 to 2.5 mm have been tested. The energy emitted by a given multielement array is well collimated and well distributed along the length of each element. Power to each element can be controlled separately, and it would be possible to combine in one array elements operating at various frequencies. The main advantages of this technique are larger implant spacing (fewer needles needed) and the potential for better volumetric control of temperature distribution.

Practical Clinical Considerations in Interstitial Thermoradiotherapy

Interstitial hyperthermia is usually combined with interstitial irradiation. Afterloading techniques are also well suited for this type of combined therapy. The MW antennas should be implanted 1 to 1.5 cm apart. The RF electrodes must be implanted at similar uniform distances, and it is extremely important to line up the electrodes in straight rows. Every implant also should include catheters for thermometry probes and thermal mapping. It should be emphasized that a good implant geometry does not necessarily ensure adequate temperature distribution; however, subtherapeutic temperature-time distributions will most likely occur if a tumor is poorly implanted. Although much research effort has produced several improved interstitial applicator designs (356), clinical trial results strongly suggest that clinical factors may have a greater impact on the therapy than do technical improvements. Moros and colleagues [ref: 236] analyzed various quality assurance parameters for an interstitial Radiation Therapy Oncology Group (RTOG) phase III trial and found that the volume of the implant was consistently smaller than the volume of the tumor. This finding implies obvious negative consequences not only for hyperthermia but also for brachytherapy effectiveness. Improvements in technology may not translate into better temperatures unless adequate implants are concurrently achieved.

Intracavitary Hyperthermia Devices

Intracavitary applicators are used to deliver local hyperthermia to tumor sites in and adjacent to body cavities, such as in gastrointestinal, female reproductive, and genitourinary systems. [ref: 105,340] EM, US, and hot-water applicators have been developed and clinically tested. [ref: 312] Among the EM approaches, MW antennas have been the most widely used, followed by RF electrodes. US applicators, given that coupling is possible, are especially appropriate for intracavitary applications because they can be manufactured in almost any size and shape, they can be operated at various frequencies to target various depths, and they can be fitted with US imaging for therapy aiming and monitoring. [ref: 153]
Figure 24-16 shows a recent design consisting of 24 elements on a cylindrical surface.



The applicator allows both longitudinal as well as angular power deposition control.

Deep/Regional Hyperthermia Devices

The EM devices used for heating deep-seated tumors all use low-frequency (less than 120 MHz) EM radiation to allow practical energy penetration. [ref: 255,403,428] In the low-frequency range, the wavelength in tissue is 30 to 50 cm, which implies, in turn, that EM deep hyperthermia devices cannot concentrate energy deposition in a volume less than about 15 cm in diameter. US devices have much shorter wavelengths in tissue (less than 3 mm); therefore, in principle, they can localize heating more precisely to very small volumes. [ref:
125,127,392]

Magnetic Induction Devices for Deep/Regional Hyperthermia

RF range magnetic fields can penetrate tissues essentially without attenuation or coupling. However, the absorbed energy is not a direct result of the magnetic fields but is generated by electric fields that are induced by the time-varying magnetic fields. Unfortunately, the geometry of most magnetic induction devices is such that the electric field vanishes toward the center of the patient (center of coils), and most of the energy is deposited in the superficial tissues. [ref: 125,271]

Capacitive Devices for Deep/Regional Hyperthermia

Capacitive devices consist of parallel opposed plates on opposite sides of the patient. [ref: 125,145] A dielectric coupling medium is required. Electric currents run between the plates through the patient. The dimensions of the plates must be comparable with their separation. Some degree of steering of the power distribution, either anteriorly or posteriorly, can be achieved by using a smaller plate on the side where one wishes to achieve greater SAR. The capacitive devices do not have the problem of zero SAR at the core of the patient in contradistinction to inductive devices. Unlike both inductive and low-frequency MW devices, however, capacitive devices generate an electric field oriented perpendicular to the interface between the subcutaneous fat and deeper muscle. Therefore, hot spots may develop. This problem can be overcome with surface cooling only when the
subcutaneous fat is less than 1 cm thick.

Radiative Devices for Deep/Regional Hhyperthermia

Low-frequency MW devices use radiators operating in the 60- to 120-MHz range; a single applicator can heat tumors to a depth of about 5 to 6 cm. Coupling with a dielectric bolus medium is required. Arrays of MW antennas arranged around the circumference of the patient are commonly used to induce regional hyperthermia. The most recently developed commercial phased-array device with considerable clinical experience (BSD-2000 with Sigma 60 applicator) uses eight dipole antennas (Fig. 24-17). [ref: 254,257,388,430]



When the electric fields from the eight antennas interfere constructively at depth, the resultant SAR can be increased by a factor of 8. This can compensate for the effects of attenuation. Steering of the SAR distribution can be achieved by adjusting the relative phases and amplitudes among the antennas. The system is equipped with interactive software to allow easy power and phase settings to each quadrant. These features, however, need comprehensive quality assurance. [ref: 368,428] The BSD-2000 succeeded the BSD-1000 and incorporated significant technologic improvements that have had a definite clinical impact. [ref: 85,255,299]

The sites most amenable to regional hyperthermia are the pelvis, extremities, and abdominal or lower chest wall areas. [ref: 165] Recent technical developments permit a fair degree of confidence in determining beforehand where the EM energy will be absorbed. [ref: 196,377,378] However, until advanced treatment planning systems become available, the treating physician and physicist must assume that any problems with severe discomfort during a hyperthermia session are due to excessive temperatures at unmonitored locations.

Ultrasound Devices for Deep (Local) Hyperthermia

The short wavelengths of US in soft tissues (0.15 to 3.0 mm) allow sharp focusing, accurate steering/scanning, and deep penetration. Focusing and scanning can be achieved electrically, geometrically, or mechanically. Acoustic lenses, overlapping beams from several planar or spherically curved (focusing) transducers, and various phased-array designs have been developed for deep localized hyperthermia applications. [ref: 77,78,127,153,154,183,191,263,392] These systems are complex, with several degrees of freedom in motion of the focal spot(s) and are highly flexible in selection of various physical parameters (number of transducers, f-numbers, crystal sizes, frequencies, orientations, scanning patterns, scanning speeds, duty cycles, phase settings, etc.). These features make them ideally suited for advanced feedback temperature control algorithms. The intensity gain due to focusing is very high and proportional to the ratio of the source radiating area over the focal diameter area; thus, deep lesions can be treated despite attenuation losses in intervening tissues. US beams suffer weak scattering in tissue; that is, most of the attenuated energy is locally absorbed. [ref: 239] However, scanning patterns or power setting(s) of the small focal spot(s) (less than 1 cm in diameter in most cases) must be chosen carefully to avoid the creation of high or low temperature regions outside or within the target volume. [ref: 240] Due to high temporal peak temperatures along the focal path(s), proper scanning patterns or field conjugation techniques must be used to compensate for heat removal by thermal conduction and high blood perfusion. [ref: 79,241,411] Inadequate scanning of the beam(s) also may induce average temperatures higher in prefocal normal tissues than the target volume temperatures. [ref: 241] When deep lesions are treated, it is particularly important to select scanning trajectories unobstructed by bone or gas cavities (selection of an acoustic window). The high absorption and reflection by bone and the complete reflection at air-tissue interfaces lead to the creation of hot spots that may be treatment limiting. [ref: 153] The complex nature of deep local hyperthermia with US necessitates further incorporation of sophisticated technology to safely, accurately, and precisely deliver therapy. Advances in feedback control, imaging-aided aiming, and 3-D treatment planning should greatly expedite the advent of these systems into clinical practice. The advantages offered by US are far from being completely realized. Among the several designs, only scanned focused ultrasound systems (SFUS) have been tested in clinical trials (Fig. 24-18A and Fig. 24-18B). [ref: 119,127,130,191] Clinical results are encouraging.









Devices for Whole-Body Hyperthermia

Whole-body hyperthermia (WBH) results from directly or indirectly heating the circulating blood to temperatures of 41 to 42 degrees C for several hours, sometimes in combination with cytotoxic drugs. Because energy not externally applied is deposited with the patient, the maximum temperature is well understood, allowing the physician to use sedatives, analgesics, and any other medications needed to keep a patient comfortable and stable during the session. Devices usually used to administer WBH are radiant heat enclosures, thermal suits with circulating hot water, or a heat exchanger for extracorporeal blood circulation. [ref: 29,309,310]

Devices for Simultaneous Themoradiotherapy

For more than 25 years it has been known that the simultaneous application of heat and ionizing radiation enhances cell killing due to thermoradiosensitization. [ref: 56] However, most thermoradiotherapy clinical trials have used sequential delivery regimens. Simultaneous thermoradiotherapy (STRT) has been discouraged for several reasons: thermoradiosensitization occurs equally in both cancerous and normal tissues, there is a lack of specific or special hyperthermia devices, and various logistical difficulties have arisen when combining these therapies. Nevertheless, if heating can be confined to tumor volumes, improved clinical outcome should result. STRT offers the opportunity for more cell kill with limited temperatures than does sequential treatment. To achieve well-localized and controlled heating while simultaneously delivering ionizing radiation, new site- and approach-specific devices are needed. In recent years, a renewed interest in STRT has motivated the development and clinical testing of several devices and techniques. Montes and Hynynen [ref: 234] described a system for the simultaneous delivery of intraoperative orthovoltage irradiation and US hyperthermia, but it has not yet been tested clinically. The group at Okayama University has developed a system for simultaneous superficial capacitive hyperthermia and external photon or electron beams from a linear accelerator. [ref: 179] This approach has been tested clinically with encouraging results. [ref: 382] There have also been several interstitial approaches to STRT. Concomitant conductive heating and LDR brachytherapy approaches have been developed and used clinically. [ref: 104,129] Mallinckrodt Institute of Radiology had a very encouraging experience in the treatment of prostatic cancers with simultaneous long-duration, low-temperature hyperthermia and LDR brachytherapy. [ref: 104] New interstitial US technologies suitable for afterloading techniques are under development. [ref: 356]

Several STRT approaches have been developed and are under development at the Mallinckrodt Institute of Radiology for the treatment of superficial cancerous tumors by external means. Three different designs have been attempted. The first device uses single waveguide MHz) MW applicators for superficial hyperthermia and a **60Co unit (Fig. 24-19).



As shown in Fig. 24-19, the radiation beam and the MW propagation are aimed at the patient either parallel (en face setup) or perpendicularly (orthogonal setup) to each other. When the en face setup is used, the radiation beam travels through the hollow waveguide. This system was extensively tested in a phase I clinical trial. [ref: 242] The second system consists of a modified commercial multielement planar US applicator and a linear accelerator (Fig. 24-20).



This approach takes advantage of the properties of US to remove all nonuniformly perturbing or attenuating parts out of the radiation beam path; it also can be used in an orthogonal setup. [ref: 367] This system also has been successfully used in a phase I-II trial. The clinical experience and the lack of complications attributable to the simultaneity of therapies have demonstrated technical and clinical feasibility. The third system is a novel design under development that combines a US linear array and a scanning acoustic reflector to allow the use of electron instead of photon beams (Fig. 24-21).



The reflector is transparent to the ionizing radiation beam and is scanned to distribute the acoustic energy from the array over the target volume. [ref: 243] This system has a higher degree of power deposition conformability than commercially available systems. [ref: 240,245,246] Another version of this system is a dual-frequency, dual-array design with a V-shaped reflector that enables depth penetration control.

Thermometry in Hyperthermia

Invasive Thermometry

Currently, direct and continuous monitoring of temperatures in clinical hyperthermia with invasive probes constitutes the only reliable method of treatment monitoring, control, and dose calculation. Temperature probes should be able to measure temperatures to both an accuracy and precision of 0.1 degrees C in phantoms and 0.2 degrees C when used in vivo. [ref: 32,318] It is desirable for clinical probes to function accurately (nonperturbingly) in intense EM or US fields. The performance goals of clinical thermometers are summarized in Table 24-4.

Invasive thermometers fall into three basic categories: electrically conducting, minimally conducting, and nonconducting (optical) probes. They are typically designed to fit in 20- to 29-gauge plastic catheters or hypodermic needles. Conducting probes include standard thermistor and thermocouple sensors with metallic leads. The standard thermistor has a sensor that is a semiconductor, the resistance of which decreases with increasing temperature. Thermocouples use the junction between these two metals; redistribution of charge across the junction leads to the establishment of a potential difference of known temperature dependence. A minimally conducting probe commonly used in hyperthermia is the high-resistivity thermistor with carbon-impregnated plastic leads. The high-resistivity material permits accurate measurement in very strong EM fields. Nonconducting optical probes use sensors composed of gallium arsenide (GaAs) or a mixture of two rare earth phosphors. The leads are optical fibers. The GaAs sensor is a semiconductor for which the band gap is a known function of temperature. In the rare earth biphosphor probe a pulse of incident light excites both phosphor materials that subsequently fluoresce. The ratio of the intensities of a pair of fluorescence emission lines -- one from each phosphor -- is a known function of temperature.

Incorrect high temperature readings can be produced by direct heating of either the probe or the catheter. The presence of such artifacts is indicated by rapid decay of temperature with time immediately after power shut-off followed by a slower decay commensurate with the surrounding tissues. Such artifacts have been effectively eliminated for EM systems by using either high-resistive thermistors or optical probes. For US, however, artifact problems remain. US-induced temperature artifacts can be the result of direct sound absorption by catheters, viscous heating, and thermal conduction of heat along metallic leads. In clinical practice, artifacts of several degrees can be easily produced, having immediate and adverse effects on the treatment. Thin-needle thermocouple probes and the measurement of temperature during short interruptions of power are both used to minimize US artifacts.

It is important to check for thermometry artifacts every time that new technologies are combined. Straube and colleagues [ref: 366] reported artifacts when measuring temperatures with a fiberoptic probe near a high dose-rate iridium source. The design of special sensors and the development of methods for estimating and avoiding artifacts have been the subject of numerous reports. [ref: 419]

Thermometry considerations became extremely important after several studies and clinical trials indicated that thermal doses calculated from a few measurement locations were misleading and usually overestimated the actual dose. [ref: 63] Moreover, advanced heating devices started to demand extensive invasive thermometry for proper utilization of temperature feedback power control features. The use of many invasive probes is not usually feasible in the clinic. Multisensor probes, for both EM and US applications, thermal mapping devices and techniques, and descriptors of the four-dimensional (4-D) temperature-time data were developed to improve this situation. [ref: 419] Further improvements in hyperthermia delivery, monitoring, control, and quantification of dose necessitate a more complete knowledge of temperature fields during treatment. This has motivated a great deal of research in the areas of 3-D temperature estimation and noninvasive thermometry.

Noninvasive Thermometry

Noninvasive thermometry (NIT) uses physical properties that have temperature dependence and the spatial-temporal distribution of which can be imaged by various methods. NIT methods can be divided into two general groups: passive methods, which measure natural temperature-dependent EM or acoustic emission from a body, and active methods, which measure temperature-dependent EM or acoustic signals resulting from a probing radiation. [ref: 23]

Among the passive methods applied in hyperthermia, MW radiometry is the most advanced with encouraging results from an initial clinical experience in the treatment of superficial tumors. [ref: 23,231]

Among the active techniques, magnetic resonance imaging (MRI) has made tremendous strides in the past few years in improving spatial and temporal temperature measurement resolutions and in estimating blood flow or perfusion. [ref: 38,54,214,247,321] MRI is generally regarded as the most promising approach for NIT. Temperature information has been extracted using several methods and pulse-signatures such as water diffusion, [ref: 321] proton resonance frequency, [ref: 54] and diffusion-weighted echo-planar imaging. [ref: 214]

Electrical impedance tomography (EIT) actively measures the temperature-dependent electrical conductivity of tissues at low frequencies (50 to 100 kHz). [ref: 22,42,248,249,258] This technique is not as powerful as MRI, but it is considerably cheaper and more practical. Although improved temperature-retrieving algorithms are needed for accurate and reliable NIT, EIT may be a good method to augment temperature information from invasive thermometry.

NIT using US echonography attempts to measure changes in acoustic properties of tissues and relate these to temperature changes. [ref: 258,260,365] The useful information is encoded in the backscatter signal. Experimental systems and algorithms have been tested, but no clinical applications are expected soon. [ref: 365,391] Active MW radiometry is in a similar state. [ref: 23]