Cellular Response to Heat

Exposure of mammalian cells to temperatures higher than 40° C leads to reproductive cell death. This effect of hyperthermic treatment on cells’ reproductive capacity depends on both the applied temperature and the duration of the exposure, and can be represented as survival curves resembling x-ray survival curves (Fig. 74-1). When the surviving cell fraction (capable of reproduction) is plotted on a logarithmic scale versus the duration of heating on a linear scale, the survival curve is characterized by an initial shoulder region followed by an exponential decrease in the surviving fraction.⁵ The time required to reduce the survival in the exponential region to 37% of its initial value is defined as D₀.

FIGURE 74-1 • Survival curves for asynchronous Chinese hamster ovary cells heated at different temperatures for varying time.

(From Hall EJ: Radiobiology for the radiologist. Philadelphia, 1994, JB Lippincott Company, p 260.)

Several mathematical models have been postulated to describe heat-induced cell killing. The most uncomplicated model assumes a one-step reaction:

in which S is survival, t is the treatment time, and k is the inactivation rate at the treatment temperature.¹⁹ The term k may also be replaced by 1/D₀ as used in RT biology.⁶ This model does not account for the survival curve “shoulder.”

The most commonly used model to describe survival curves for heat killing is the multitarget single hit model:

in which n is the number of targets to be inactivated before the cell is killed.^6,20 The survival curves for this model are characterized by the parameters n, D₀, and D_q (the quasithreshold dose) in which D_q = (ln n)D₀. Using this model, one can construct Arrhenius plots to demonstrate the kinetics of heat killing, where 1/D₀ is plotted against the inverse of the absolute temperature (Fig. 74-2). For most cell lines studied, the Arrhenius plots of cell inactivation seem to be composed of two segments, with a break at or near 43° C.⁴ Above the break point, the activation energy is between 110 and 150 kcal/mol, which is the energy required for protein denaturation,²¹ suggesting that protein damage is mainly responsible for cell death caused by HT. At lower temperatures, the activation energy is higher, approximately 300 to 400 kcal/mol. On the basis of the difference in activation energies, it has been suggested that there may be different primary modes of heat-induced cell death, one dominant higher than 43° C and the other lower than 43° C. However, the change in the slope of Arrhenius plots lower than 43° C could simply be a manifestation of the ability of the cells to develop thermotolerance, the transient resistance of cells to heat killing.²² The Arrhenius plot for a given cell line can be modulated by thermotolerance or environmental parameters such as pH.

FIGURE 74-2 • An Arrhenius plot for heat inactivation of Chinese hamster ovary cells plotting the reciprocal of D₀ (the inactivation rate) versus the reciprocal of the absolute temperature. Note the break point at approximately 43° C.

(From Hall EJ: Radiobiology for the radiologist. Philadelphia, 1994, JB Lippincott Company, p 261.)

Using the activation enthalpy and entropy for hyperthermic cell killing (determined from Arrhenius plots based on D₀s), Lepock and colleagues²³ introduced a thermodynamic model for hyperthermic cell killing based on the existence of a critical target. The rate of cell killing is thought to correspond with the rate of inactivation and denaturation of this hypothetical critical target. The calculated transition temperature of this critical target appeared to be 46.0° C for V79-WNRE cells when the temperature was raised by 1° C per minute. When the scan rates of 0.1° C per minute or 10° C per minute were used for differential scanning calorimetry (DSC), the calculated transition temperature was 43.5° C or 48.5° C, respectively. Thus, transition temperatures are only meaningful when the scan rate used in the measurements is specified.²³ Cellular fractionation to localize the critical target showed that protein denaturation occurred at 46° C in all subcellular fractions.

A fourth model, which nicely describes heat killing under various conditions (single heating, thermotolerance, and step-down heating), was proposed by Jung.²⁴ This model assumes a two-step process. The first step is the production of nonlethal lesions that can be converted into lethal lesions upon further heating in the second step. After the cells are heated for a time t at a certain temperature, the surviving fraction S is given by the equation:

in which p is the rate constant for the production of nonlethal lesions per cell per unit of time and c is the rate constant for the conversion of a nonlethal lesion into a lethal lesion per unit of time. One lethal lesion is sufficient to kill a cell. Jung assumed that the production and conversion of nonlethal lesions were random and the rate constants depended only on the temperature of the heat treatment.

The intrinsic thermal sensitivity of different cell lines varies significantly. So far, no consistent difference has been demonstrated between normal and malignant cells. On the other hand, the extracellular milieu can modify cells’ thermal sensitivity. In general, cells that are at low pH or are nutrient-deprived are more sensitive to heat. These cells may be hypoxic cells that are three times more resistant to radiation than are normally oxygenated cells. However, cells exposed to low pH for long periods adapt to pH changes and lose their increased heat-sensitivity.^25,26 Sensitivity to heat shock as a function of cell cycle has also been extensively studied. It is well established that the age response of thermal sensitivity complements that for x-irradiation. The phase of the cell cycle most resistant to x-rays (late S phase), is most sensitive to heat shock response.²⁷

When more than one heat treatment is given, cellular thermal sensitivity to the combined treatments can be modified depending on the order of the heat shock temperatures involved. Step-up heating occurs when the second heat shock is at a higher temperature than the first heat exposure at 39° C to 42° C. Conversely, step-down heating occurs when the first treatment is a short exposure at a high temperature, greater than 43° C, immediately followed by a subsequent exposure to less than 43° C. During step-up heating, thermotolerance may be induced if the first treatment temperature is less than 43° C. Therefore, the cytotoxic effect of combined heat treatments is much less than the effect without preheat treatment. On the other hand, during step-down heating, preheat treatment at greater than 43° C for a short period can sensitize cells to a subsequent heat shock treatment at a lower temperature, which may have been nonlethal by itself without the preheat treatment. However, a high temperature pulse does not abolish any thermotolerance developed earlier.⁶

Thermal Dose

Because heat killing is a function of both time and temperature, a thermal dose unit is needed to combine both factors and account for the different effect of heat higher and lower than 43° C. Equivalent minutes at 43° C (EM 43° degrees) was proposed as the thermal dose unit by Sapareto and Dewey in 1984:²⁸

in which t is the time at temperature T, the constant R = 0.5 for temperatures higher than 43° C, and R is 0.25 less than 43° C. Thermal dose for a given treatment can be obtained by summing thermal doses for each minute of treatment, and the thermal dose from multiple treatments can be summed to obtain the cumulative equivalent minutes (CEM) at 43° C for an entire treatment course, assuming that the interval between heat treatments is long enough to allow for decay of thermotolerance. To describe the variations in thermal dose throughout a tumor, minimum, maximum, and average values can be given as well as thermal dose values for the T₉₀ and T₅₀ temperatures (the temperatures exceeded by 90% or 50% of the measured intratumoral points, respectively), and so on.

This concept of thermal dose has been validated in multiple pet animal and human clinical trials. Multivariate analysis of a randomized trial in pet animals with spontaneous tumors treated with RT alone or RT plus HT showed that, along with tumor volume, the non–site-specific average minimum EM 43° was the most important predictor of complete response (CR) (P < 0.001) and local control (P < 0.05). The CR rate was only 32% for 28 tumors with an average minimum CEM 43° (CEM 43° T_min) of 1.0 minute or less, but 68% for 79 tumors with greater than 1.0 average CEM 43° T_min.²⁹ A thermal dose analysis of a prospective, randomized trial in patients with recurrent breast cancer showed a significant association between CEM 43° T_min and response, with CR rates of 43% for CEM 43° T_min less than or equal to 10 minutes versus 77% for CEM 43° T_min greater than 10 minutes.³⁰ Retrospective analyses of HT patients treated at Duke University showed that the cumulative minutes of treatment for which T₉₀ exceeded 39.5° C was highly associated with CR of superficial tumors,³¹ and that the cumulative minutes of treatment for which T₅₀ exceeded 41.5° C was highly associated with greater than 80% necrosis of soft-tissue sarcomas treated preoperatively with thermoradiotherapy.³² Thermal dose-response formulas based on these data showed that T₉₀ and T₅₀ temperatures routinely achieved need to be improved by 1.2° C to 1.5° C, or treatment duration increased threefold to fivefold, to increase thermal dose sufficiently to justify phase III trials.¹⁶ There were also significant associations between CEM 43° T₉₀ and response in a study of preoperative regional HT combined with conventional RT and low-dose 5-fluorouracil chemotherapy and leucovorin for 37 locally advanced primary rectal cancers (P = 0.006)³³ and between CEM 43° T₉₀ and CR and local control in patients with recurrent breast adenocarcinoma treated at Stanford with RT and HT.³⁴ These parameters, CEM 43° T_min and CEM 43° T₉₀, have become standard thermal dose descriptors for superficial HT.

Mechanisms of Action of Heat

Unlike ionizing radiation, HT does not result in the localized deposition of high levels of energy in cells. The thermal energy is more or less evenly absorbed by all molecules in the cell. As mentioned previously, the activation energy obtained for cultured cell lines and transplantable tumors in the temperature range from 42.5° C to 47° C is approximately 150 kcal/mol,^6,27 suggesting that protein denaturation is the main cause of hyperthermic cell killing.

More direct evidence that protein denaturation is responsible for thermal killing was provided using DSC.²³ In eukaryotes, the onset for protein denaturation was found to be approximately 40° C. Thermotolerance, cycloheximide, and D₂O all increased the thermostability of proteins and also resulted in increased survival levels after HT.³⁵ These results were further confirmed by studies conducted by Burgman and colleagues³⁶ using electron spin resonance measurements and thermal gel analysis, showing that a correlation existed between protein denaturation and heat killing.

Protein denaturation and subsequent aggregation appear to occur throughout the entire cell, including cytoplasmic, nuclear, and membrane proteins.³⁶ It is unlikely that a single type of protein is the critical target for cell killing by heat. All subcellular structures contain temperature-sensitive proteins that denature and may subsequently aggregate.³⁷ Small changes in temperature can drastically alter the structure of plasma membranes and impair many membrane-related functions that, by themselves or in combination with other protein damage in the cell, could lead to cell death.^38,39 Heat has also been shown to alter the mitotic spindle and centrosome organization, resulting in the formation of multinucleated, nonclonogenic cells,⁴⁰ and to damage mitochondria,⁴¹ with heat-induced inhibition of metabolic functions such as glycolysis and respiration. Heat damages polysomes and microsomes and inhibits the synthesis of proteins, ribonucleic acid (RNA), and DNA in a thermal dose–dependent manner.^39,41 The inhibition of RNA and DNA synthesis appears to be due to heat-induced changes in chromatin structure caused by the denaturation and aggregation of nuclear proteins. Heat-induced changes in the nuclei of eukaryotic cells include the appearance of actin bundles; increased vesiculation; the reduction of intact nucleoli⁴¹; and a dramatic, dose-dependent increase in protein content in isolated chromatin, nuclei, nuclear matrices, and nucleoids.^35,42,43 This appears to be at least partially due to decreased leakage of proteins such as DNA and RNA polymerases from the nucleus during isolation as the heat-denatured proteins become aggregated and thus insoluble. This phenomenon is referred to as heat-induced nuclear protein aggregation. It is likely that nuclear protein aggregation, which occurs largely at the nuclear matrix, interferes with DNA replication, transcription complex interactions, and DNA unwinding needed for DNA replication and DNA repair.⁴⁴

Interestingly, on exposure to heat, a group of proteins called heat shock proteins (HSPs) translocate from the cytoplasm to the nucleus.^41,44,45 It has been shown that HSPs protect against protein aggregation or facilitate disaggregation of heat-induced insoluble protein complexes in cell lines and yeast.^46–48

Regulation of Heat Shock Response

As described earlier, the heat shock response refers to the increased transcription of a set of genes, the heat shock genes, when cells are subjected to heat stress. It is a highly conserved process that occurs in all organisms from bacteria to human. In eukaryotes, the response is mediated by a sequence-specific DNA binding protein termed HSF1.⁷⁰ In its trimeric form, HSF1 binds tightly to multiple copies of a highly conserved sequence motif (nGAAn) in the promoter region of heat shock genes; these HSF-binding sites are termed the heat shock elements (HSEs).^71,72 In unstressed mammalian cells, HSF1 is maintained in a monomeric, non-DNA binding form. Upon heat shock, HSF1 assembles into trimers, binds to HSEs with high affinity, becomes hyperphosphorylated, and activates the transcription of the heat shock genes. Transcriptional activation of these genes, in turn, increases the synthesis of HSPs. During post–heat-shock recovery, the HSF1 trimers dissociate from the DNA, and are eventually converted to the inactive, non-DNA binding monomers. Several other models, similar to the previously described model, have also been proposed.^18,73

Although the importance of HSF1 in the regulation of mammalian heat shock genes is well-established,⁷⁰ data from multiple laboratories obtained using various rodent and human cell lines indicate that the activation of HSF1 by itself is not sufficient for the induction of HSP70 messenger RNA (mRNA) synthesis,^62,74–80 and that other factors may be involved in the regulation of HSP70 transcription. Subsequent studies of Rat-1, hamster HA-1, and HeLa cells provided evidence of the plausible involvement of a negative regulator,^62,81 initially termed constitutive HSE binding factor (CHBF). Upon heat shock, the cellular level of the HSE binding activity of HSF1 rapidly increases (“activation of HSF1”) and that of CHBF rapidly decreases (“inactivation of CHBF”). When HSP70 transcription returns to its pre–heat-shock level during post–heat-shock recovery, the disappearance of the HSF1-HSE binding activity again parallels the reappearance of the CHBF-HSE binding activity.^62,82 Interestingly, in rat, mouse, and human cells treated with sodium salicylate or sodium arsenite, which elicits considerable HSF1-HSE binding activity but shows no reduction in CHBF-HSE binding activity, the induction of HSP70 mRNA synthesis was not observed.^62,77,80,83 Thus, data from studies of a wide variety of rodent and human cells all support the notion of a negative regulatory role of CHBF in HSP70 induction.

Subsequently, CHBF was purified and identified as the Ku protein, a heterodimer of Ku70 and Ku80,⁸¹ allowing more direct testing of the hypothesis that CHBF/Ku acts as a negative regulator in the heat induction of HSP70. To carry out such experiments, Li and colleagues established Rat-1 cells that stably and constitutively express human Ku70 and/or Ku80, and found that overexpression of the human Ku70 by itself, or Ku70 and Ku80 jointly, specifically inhibits heat-induced HSP70 expression without affecting induction of other HSPs or activation of HSF1.^84,85 How does Ku70 specifically suppress HSP70 expression? There are at least four non–mutually exclusive possibilities: (1) Ku70 may bind directly to sites located in the HSP70 promoter and repress transcription; (2) Ku may regulate HSF1 binding to the HSP70 promoter; (3) Ku may affect the interaction among other transcription factors and their binding to the HSP70 promoter; (4) overexpression of Ku70 could result in altered chromatin structure, inhibiting the access of positive factors to their respective regulatory elements in the HSP70 promoter. Related to the second possibility is the recent finding that in vitro translated Ku70/Ku can bind to HSF1 and displace it from HSE.^86,87 Thus the existing evidence strongly supports a role of Ku in HSP70 gene regulation.

Interaction of Heat With Radiation

Heat interacts with radiation in a more than additive way. This synergistic interaction of heat and radiation is interpreted as a heat-induced sensitization of cells to radiation, termed heat radiosensitization, or thermal radiosensitization. Heat radiosensitization can be quantified using a thermal enhancement ratio (TER), defined as the ratio of radiation doses with and without heat to produce the same biologic effect (isoeffect TER). In some cases, TERs are expressed as a ratio of the biologic effects for the same radiation dose, with and without heat (isodose TER), or as a ratio of the D₀’s of the radiation survival curves. Usually, TER increases with increasing heat dose. The maximal interaction is observed when heat and radiation are given simultaneously. The TER decreases with an increasing time interval between heat and radiation (Fig. 74-4). When radiation precedes HT, sensitization is no longer detectable 2 to 3 hours after radiation. When HT precedes radiation, cells can be sensitized for up to several hours.

FIGURE 74-4 • Survival of Chinese hamster ovary cells irradiated with 5 Gy during or at varying intervals before or after a heat treatment at 42.5° C for 40 minutes. The maximum effect occurs for simultaneous heat and radiation.

(From Hall EJ: Radiobiology for the radiologist. Philadelphia, 1994, JB Lippincott Company, p 272.)

The therapeutic gain factor (TGF) is defined as the ratio of the TER in the tumor to the TER in surrounding normal tissue. Maximal heat radiosensitization is achieved when heat and radiation are applied simultaneously, but a therapeutic gain (more damage to tumor than to normal tissue) may not always be obtained under these conditions, especially if tumor and normal tissue are heated to the same degree.^88,89 In some cases, the TGF may increase when time intervals are introduced between radiation and heating.

The combination of heat and radiation does not result in an enhancement of heat lesions; thus, no radiation-induced heat sensitization takes place and the synergistic effect is most likely due to radiosensitization by heat.⁵⁶ The primary target for radiation is believed to be DNA, whereas protein denaturation and aggregation seem to be involved in thermal cell killing.^42,90 DNA damage (e.g., breaks) can be detected at low lethal doses of x-rays (≤1 Gy), whereas no DNA breaks are found at low lethal heat doses.⁹¹ The combination of radiation and HT generally does not result in more initial DNA breaks than are observed for radiation alone.^90,91 Instead, HT appears to inhibit the rejoining of radiation-induced DNA strand breaks^91–93 and the excision of damaged bases⁹⁴ in a heat-dose–dependent manner.^91,93 This effect may be caused by (1) heat-induced inactivation of DNA repair enzymes^88,95; or (2) alteration of the chromatin structure resulting from protein denaturation and aggregation, which causes decreased accessibility of the damaged sites to the repair machinery.^44,92,94,96 Using cell lines in tissue culture, researchers found a good correlation between the amount of heat-induced nuclear protein aggregates and the extent of thermal radiosensitization.^95,97

The effect of thermotolerance on heat radiosensitization has also been a topic of extensive investigation. Studies using tissue-cultured cell lines have shown that thermoradiosensitization could be affected by thermotolerance development.^98,99 Because it has been reported that HSPs might be involved in the protection against nuclear protein aggregation as well as in facilitated disaggregation,^46,48 the presence of these proteins might indirectly be responsible for the thermotolerance effect on protein aggregation and as such affect DNA repair and cellular radiosensitivity. Rodent cells transfected with human HSP70 showed protection against heat-induced nuclear protein aggregation that was reflected in the extent of thermal radiosensitization.⁹⁵ Transfection with HSP27 resulted in an accelerated nuclear protein disaggregation and was paralleled by an accelerated decline of thermal radiosensitization. These observations point to the importance of HSPs as a regulator in protein denaturation, aggregation, and disaggregation, thereby modulating the heat effects on DNA repair. As such, the expression of HSPs may explain the effect of thermotolerance on the extent and disappearance of thermal radiosensitization and could have a major effect on the choice of sequence and spacing of heat and radiation in clinical practice. Elevated HSP expression should be either avoided, if occurring in the tumor, or exploited, if occurring in the normal tissue at risk during treatment.

Chronic thermotolerance has been found to affect heat radiosensitization in some cell lines,^17,93,99,100 although for other cell lines no effect of thermotolerance was observed at the level of heat radiosensitization.^99,101 For example, chronic thermotolerance development in CHO cells inhibited further sensitization to radiation.¹⁰⁰ This lack of further radiosensitization was also found in HeLa cells when no chronic thermotolerance developed.⁶⁷ Interestingly, long-duration, mild HT (as low as 41° C) given concurrently with low-dose-rate irradiation can cause extensive radiosensitization and eliminate the low-dose-rate irradiation sparing effect.^{69,88,102,103} This thermoradiosensitization to low-dose-rate irradiation occurred for protracted heating times and did not appear to be inhibited for prolonged heating times, supporting the absence of a thermotolerance process in radiosensitization.^69,102 The mechanism underlying this low-temperature radiosensitization may very well be different from the one discussed in the previous section on the mechanism of heat radiosensitization because DNA repair rates are often found to be higher at 41° C^104–106 and heat-induced protein aggregation in the nucleus is only limited (Stege and associates, unpublished results). Effects of heat on cell-cycle progression or changes in fidelity of repair may be possible explanations of radiosensitization by long-duration, low-temperature HT.

Heat and Chemotherapy

Although the combination of HT with radiation has been the focus of more attention, there is an equally strong rationale for combining HT with chemotherapy. Moderate or even mild HT (such as 39° C) enhances cell killing in vivo of a number of chemotherapeutic agents, such as cyclophosphamide, melphalan, mitomycin C, cisplatin, doxorubicin (DOX), bleomycin, and the nitrosoureas. A variety of mechanisms may account for increased chemotherapeutic effect at elevated temperatures, including altered pharmacokinetics or pharmacodynamics, increased DNA damage, decreased DNA repair, reduced oxygen radical detoxification, and increased membrane damage.¹⁰⁷ Concentrations of agents that are not normally toxic at normal body temperature can become cytotoxic at 39° C, and, in some cases, HT may partially overcome some types of drug resistance.

Heat and Liposomes

It is well recognized that when systemic chemotherapy is used for the treatment of solid tumors, significant normal tissue toxicities commonly occur when therapeutic levels of drug at the tumor site are achieved.¹⁰⁸ Liposomes have been identified as one of the promising carriers for therapeutic agents to improve drug delivery in the treatment of cancer.^109,110 Conventional liposomes have been clinically evaluated and approved in a variety of diseases^111–113 and HT has been shown to increase extravasation of conventional liposomes within tumor tissue.¹¹⁴ Early in 1978, Yatvin and colleagues proposed the use of HT with temperature-sensitive liposomes to allow increased control over the release of encapsulated drugs at the diseased site,¹¹⁵ and their efforts have stimulated many additional studies.¹¹⁶ In 2000, Dewhirst and collaborators developed a novel thermosensitive, DOX-containing lipid formulation optimized for rapid drug release under mild hyperthermic temperatures (39° C to 40° C) readily achievable in the clinic (Fig. 74-5).^113,117,118 Their in vitro studies clearly demonstrate advantages that the new lipid composition has compared with existing liposome formulation; the characteristics of attainable and narrow triggering temperature, rapid release of drug, and high accumulative drug release appear to offer potential advantages that could lead to an increased therapeutic effect over traditional chemotherapy and current liposomal and other drug carrier and delivery systems. Furthermore, in vivo studies demonstrated that this novel liposome, in combination with mild HT, was significantly more effective than free-drug or current liposome formulations at reducing tumor growth in a human squamous cell carcinoma xenograft line (FaDu), producing 100% complete regressions in 11 tumors lasting up to 60 days post-treatment.^113,117 To monitor the liposome concentration distribution and drug release in vivo, magnetic resonance imaging (MRI) contrast agent, MnSO(4) and DOX were encapsulated together within this thermosensitive liposomes.¹¹⁹ The MRI clearly showed that the thermally sensitive liposome was only accumulated at the periphery of the tumor, concordant with the release temperature of this formulation (39° C–40° C). Because the liposomal drugs accumulate only in perivascular regions in tumors after intravenous injection, tumor cells in deeper tissue layers cannot be killed. To circumvent this deficiency, Dewhirst and collaborators decided to target tumor microvessels and therefore stop nutrient supply to deeper tumors by constructing a DOX–encapsulated, lysolecithin-containing thermosensitive liposome (LTSL). After DOX-LTSLs in combination with HT treatment, the blood flow of FaDu tumor implanted in dorsal skinfold chamber was significantly decreased, whereas only minor reductions in normal microcirculation in subcutaneous tissues were observed.¹²⁰ Furthermore, a real-time evaluation of therapeutic protocols in association with outcome on an individual subject basis could be achieved by noninvasively monitoring the temporal and spatial patterns of DOX/MnSO(4)-LTSLs delivery with MRI.¹²¹

FIGURE 74-5 • A, Schematic of temperature-sensitive liposome structure. Liposomes are spherical lipid vesicles enclosing an aqueous compartment that contains drug. Amphiphilic drugs can be actively loaded and retained through a pH gradient. The phospholipid bilayer is characterized by a main melting phase transition. Lysolipids can form stable pores at the phase transition temperature. Polymers may be grafted onto lipids to evade immune recognition. Note that only unitary examples of each molecular strategy are shown. B, Schematic of possible mechanisms involved in combination hyperthermia (HT) and liposomal therapy for solid tumors. 1, Liposomes (blue circles) preferentially extravasate from pores in tumor vessel walls. 2, HT increases tumor vessel pore size and thus increases tumor liposome extravasation. 3, HT can trigger drug (yellow) release from liposomes in the tumor vessel. 4, HT can trigger liposomal drug release in the tumor interstitium. 5, HT can be directly cytotoxic to tumor cells.

(From Ponce A, Yujaskovic Z, Yuan F, et al: Hyperthermia mediated liposomal drug delivery, Int J Hyperthermia 22:205–213, 2006.)

Heat Shock Proteins and Cancer Immunotherapy

Although HSPs have been intensively studied at the level of transcription and translation for mammalian cells for more than 2 decades, a novel and physiologically important role for these highly conserved gene products has only become evident in the past several years. In the early 1990s, Srivastava and colleagues found that a tumor rejection antigen, isolated by biochemical fractionation of tumor cells, was an HSP (grp94/gp96).^122–125 Since then, convincing evidence has accumulated that the HSPs derived from a given cancer, including HSP70, HSP90, and grp94/gp96, can elicit protective immunity specific to that particular cancer. HSPs derived from normal tissues do not have such effect against any cancer tested. In recent years, the immunogenicity of HSP preparations from tumors has been seen in different experimental tumor systems of distinct histologic origins, ranging from chemical- or ultraviolet (UV)-induced tumors to spontaneous tumors.^126–128 Immunization of mice and rats with HSP-peptide complexes leads to protective and specific immunity against cancer from which the HSP-peptide complexes are derived.¹²⁴ This phenomenon has been shown in mice¹²⁴ and rats¹²⁹ in a wide array of tumor types, including fibrosarcoma,¹²² hepatocarcinoma,¹²⁹ lung carcinoma, melanoma, colon carcinoma,¹²⁷ squamous carcinoma,¹³⁰ leukemia,¹³¹ and in prophylaxis,^122,129 as well as in therapy of pre-existing diseases.¹²⁷ Two high-molecular weight HSPs, HSP110 and grp170 derived from CT26 and Meth A tumors, were found to induce an antitumor response against these tumors.^132–134 Furthermore, it has been demonstrated that the immunization of mice resulted in the induction of memory T cells.¹³⁰ Taken together, these data suggest that HSPs prepared from tumors can serve as tumor vaccines, and that HSPs may play a central role in the host’s immune system. Currently, one HSP, the endoplasmic reticulum stress-response protein gp96, is undergoing clinical trials for cancer treatment and has yielded promising results, including the induction of antitumor immunity and some benefit for patients when administered as part of a multidose regimen.¹³⁵ A recent phase III comparison of Vitespen, an autologous tumor-derived HSP gp96 peptide complex vaccine, with physician’s choice of treatment for stage IV melanoma, shows that patients with less advanced disease (M1a and M1b) benefit from increasing doses of vaccine treatments.¹³⁶ Future advances in HSP-based immunotherapy will be aided by an understanding of the mechanisms by which HSP-peptide complexes induce innate and adaptive immunity to tumor cells and target the killing of primary and metastatic cancer cells.¹³⁷

Hyperthermia and Gene Therapy

Another new application of HT is its use in gene therapy to provide both spatial and temporal control over gene expression governed by heat shock or other stress-inducible promoters.^138–141 As one example, Huang and colleagues transfected cells with adenovirus vectors containing an HSP70 promoter and genes for green fluorescence, interleukin 12, or tumor necrosis factor alpha.¹⁴² Experiments with heat-induced green fluorescence protein showed detectable expression at 3 hours, reaching a peak at 18 to 24 hours and disappearing by 72 hours after heating. A modest heat shock of 42° C for 30 minutes increased interleukin-12 gene expression by more than 13,000-fold over baseline and increased tumor necrosis alpha gene expression by 6.8 × 10⁵-fold.¹⁴² To optimize the application of heat-activated gene-RT in the clinic, techniques are being investigated to noninvasively monitor the transgene activation. Li and colleagues first used micro–positron emission tomography (microPET) imaging to monitor the improvement in viral vector distribution after a mild heat shock 6 hours after administration of virus (Fig. 74-6).¹⁴³ Rogers and colleagues further used microPET imaging to noninvasively monitor the heat-inducible suicide gene expression in mice bearing head and neck squamous cell carcinoma xenografts.¹⁴⁴ Recently, Chrit and colleagues investigated in vivo spatiotemporal control of transgene activation under the regulation of a heat-inducible promoter using magnetic resonance temperature imaging for image-guided, high-intensity focused ultrasound (US).¹⁴⁵

FIGURE 74-6 • A, Showing transplanted tumor on left and right hind thighs of the mouse. B and C, Micro–positron emission tomography images of unheated and heated mice acquired 48 hours after intratumoral injection of replication-defective adenovirus rAd (CMV-TKeGFP).

(From Li GC, He F, Ling CC: Hyperthermia and gene therapy: potential use of microPET imaging, Int J Hyperthermia 22[3]:215–221, 2006.)

Hyperthermia and Tumor Hypoxia

Tumor hypoxia is not only one of the major factors contributing to RT resistance of solid tumors that leads to adverse treatment outcome, but also is related to disease progression and metastasis.^146–149 Mild HT (<42° C) is believed to dilate tumor blood vessels and increase tumor oxygenation, and so enhance tumor radiosensitivity.¹⁵⁰ Clinical and experimental studies showed that mild HT could increase the overall tumor pO₂ level in rodent,¹⁵¹ canine,¹⁵² and human tumors,^153,154 and the reoxygenation can even last for 1 to 2 days after heating in some tumor types.¹⁵⁰ Recently, the change in tumor hypoxia induced by mild HT was investigated at the microscopic level by dual hypoxia marker immunohistochemistry technique in a human colorectal adenocarcinoma xenograft model (HT29).¹⁵⁵ It was found that although the overall hypoxic fraction was significantly decreased during heating, mild HT induced both increases and decreases in tumor hypoxia in different regions of the tumor (Fig. 74-7).¹⁵⁵ Specifically, mild HT decreased hypoxia in the regions with relatively well-perfused blood vessels, but increased hypoxia in regions that were poorly perfused. These data implied that the changes in tumor oxygenation depend on the vasodilation ability of tumor microvessels in response to HT.

FIGURE 74-7 • Comparison of binary image detail from the two different regions of the same heated tumor. Images from left to right are pimonidazole, EF5, CD31 and Hoechst 33342 together, and the composite binary image showing all markers, respectively. The upper panel shows an example region where the hypoxic foci decrease during heating; the lower panel shows a region where the hypoxic foci increase during heating. Original magnification 50×, Scale bar = 200 µm.

(From Sun XR, Li XF, Russell J, et al: Changes in tumor hypoxia induced by mild temperature hyperthermia as assessed by dual-tracer immunohistochemistry, Radiother Oncol, 88[2]:269, 2008.)

Mechanisms of Heating

Heating of biologic tissue may be accomplished by one of three primary mechanisms: (1) thermal conduction of heat away from a source at higher temperature; (2) a combination of resistive and dielectric losses in tissue from an applied electromagnetic (EM) field; or (3) mechanical losses from molecular oscillations caused by an ultrasonic pressure wave. The underlying physical principles of each of these three modalities are reviewed briefly.

Electromagnetic Power Deposition

All living tissues contain electrical charges, some free and some bound, that are available for interaction with externally applied EM fields. Tissues with high water content and blood circulation such as muscle, skin, brain, and internal body organs have a high percentage of polar water molecules and thus conduct electric current well. Tissues with low water content, such as fat and bone, have mostly bound charges that do not conduct current well. Across the frequency range of interest for HT, all tissues may be considered as lossy dielectrics—somewhere between poor electric conductors and poor insulators. The electric properties of human tissues vary considerably and may be characterized by the relative dielectric constant (ε_r) and electrical conductivity (σ). The primary mechanism of EM heating of tissue also varies with frequency. In the radiofrequency (RF) range less than approximately 10 MHz, the alternating field induces a net movement of free electrons, and power deposition results from resistive (ohmic) losses caused by the electric currents in lossy tissue. At microwave (MW) frequencies higher than approximately 100 MHz, the dielectric losses in tissue predominate and heating results primarily from friction caused by interactions between polar water molecules that rotate and oscillate to maintain alignment with the time-varying EM field. In either case, the time-averaged APD is:

in watts per cubic meter, in which J = σE is the induced current density. This APD is closely related to the more commonly used power deposition rate or specific absorption rate (SAR) by the formula:

in which ρ is the mass density of tissue in kilograms per cubic meter.

For both RF and MW radiation, APD and SAR decrease with depth in tissue. The APD for the simplest case of plane-wave radiation is plotted as a function of penetration depth into muscle tissue in Fig. 74-8. At higher frequencies, power is deposited more superficially, whereas lower frequencies provide deeper penetration. However, for practical HT applicators, the tissue penetration may be much less than this because of coupling problems at the applicator-tissue interface, tissue heterogeneities, and nonuniform beam profile in the near-field region of HT applicators that are small compared with the wavelength.

FIGURE 74-8 • Electromagnetic plane wave penetration in muscle tissue as a function of frequency, showing deeper penetration with lower frequency.

For frequencies used in HT, the wavelength (λ) of EM energy in soft tissue varies from approximately 3.4 cm at 10 MHz to less than 4 cm at 1000 MHz. The minimum spot size for focusing power deposition is approximately λ/2, or approximately 20 cm at the lowest RF frequencies typically used (100 MHz) and approximately 2 cm at the higher MW frequencies. Thus, lower RF frequencies penetrate well but affect large regions of the body, whereas higher MW frequencies may be localized effectively in tumor-sized tissue volumes. Compilations of electrical conductivity, dielectric constants, wavelengths, and penetration depths of EM waves in air, muscle, and fat are readily available.^161,162

Ultrasound Power Deposition

Unlike EM radiation, US energy propagates through tissue as a traveling pressure wave. Variations in pressure from the alternating compressive and expansive forces cause tissue molecules to vibrate and collide, producing heat from mechanical friction within the molecular structure. As a result of these losses, US energy within the pressure wave is attenuated exponentially as it propagates through homogeneous tissue such that the acoustic pressure P = P₀e^−αx, in which α is the amplitude attenuation coefficient, x is the distance into tissue, and P₀ is the initial pressure at the surface. The acoustic intensity I (the rate of energy flow through a unit area) equals P²/(2ρc_s) in which ρ is the tissue-mass density and c_s is the acoustic wave velocity, or speed of sound in tissue. Thus, the acoustic power deposition, or APD is:

in watts per cubic meter, and again SAR is APD/ρ in watts per kilogram.

Fig. 74-9 shows the penetration depth characteristics of US in homogenous muscle tissue as a function of frequency, assuming well-behaved plane wave radiation. Higher frequencies result in more superficial localization of power and lower frequencies result in deeper penetration. For practical HT transducers, the target volume usually appears in the transducer near field, where there are marked fluctuations in beam intensity both longitudinally in the direction of propagation as well as across the wavefront, as opposed to much smoother inverse-square-law longitudinal decay of intensity in the far field. Thus, reflections, scattering, and thermal conduction within the tissue are relied on to smooth the temperature distribution that results from the peaky SAR. Nonfocused applicators operating at approximately 3.5 MHz can heat tumors up to 4 to 6 cm in depth.¹⁶³ Nonfocused applicators operating at lower frequencies generally produce increased pain because of unavoidable absorption in underlying bone. Deeper penetration is made possible by directing the US beams to avoid bone and air interfaces, using a larger acoustic window at the surface, and focusing the US beams to increase the power density at depth relative to that at the skin surface and critical normal tissues outside the focal volume.

FIGURE 74-9 • Ultrasound plane wave penetration in muscle tissue as a function of frequency, showing deeper penetration with lower frequency.

Because the speed of sound (c_s) is orders of magnitude lower than the speed of light, the wavelength (λ) of US radiation λ = c_s/f in tissue higher than the frequency (f) range of interest (0.5–10 MHz) is 0.1 to 3 mm. With such small wavelengths, US beams are easily focused into very small volumes at depth. With care, treatment of larger volumes may be accomplished by using mechanically or electrically scanned focal spots or by using multiple beams aimed at different locations within the target volume.

The characteristic acoustic impedance Z = ρc_s is an important parameter in determining the behavior of US at tissue interfaces. The Z’s of most soft tissues are quite similar, so there is little reflective loss during the propagation of US from one soft tissue to another. However, the Z values for bone, air, and lung are considerably different from those of soft tissues, causing almost complete reflection at soft tissue–gas interfaces and both reflection and rapid absorption of the transmitted portion of the wave at soft tissue–bone interfaces.¹⁶⁴ A tabulation of acoustic properties of tissues and more in-depth coverage of US interactions with tissue are available.^161,162,165

Because of the problem of reflection of US energy at air interfaces, it is critical to carefully couple US energy into the body through degassed water and a thin layer of US coupling gel, with elimination of all air pockets and bubbles between the source and tumor. For US frequencies less than approximately 1.5 MHz, a second water bag coupled with gel to the skin at the exit beam site may be necessary to avoid skin burns. US reflection at air interfaces may be turned to an advantage to block out critical normal tissues located close to the desired target (e.g., the mandible adjacent to a neck mass) by intentionally drying off the skin surface and placing a 3- to 5-mm layer of closed cell foam (e.g., common plastic packing material) on the skin surface under a portion of the applicator.

Heating Equipment and Techniques

During the past 2 decades, HT equipment has evolved remarkably, with a consistent trend toward more complex multiple-applicator systems with improved control of SAR pattern. Many unique devices have resulted and may be classified as local (external-superficial, interstitial, or intracavitary), regional (external-deep), or whole-body heating techniques. Selected representative devices are mentioned in the following text.

Local Heating—External Sources

A large development effort has gone into producing local heating devices with improved control of power deposition pattern. EM techniques are generally used for superficial tumors less than approximately 3 cm in depth, such as chest-wall recurrence of breast carcinoma, whereas US beams are useful for somewhat deeper tumors up to approximately 6 cm deep.^{161,166–169}

Electromagnetic Techniques for Superficial Heating

The simplest, most reliable applicator used in clinical HT has been the single-aperture MW waveguide. A MW radiator mounted within a waveguide (a tubular or rectangular structure for guiding EM waves) radiates EM waves out the waveguide opening (aperture). Commercial applicators have been available in a variety of sizes from 7.5 to 24 cm on a side, normally operated at 915 MHz or approximately 430 MHz as these are the Industrial, Scientific, and Medical band frequencies allowed for use with no power restrictions in the United States, and Europe and Asia respectively. The effective heating area, or the area under the applicator producing 50% or greater of the maximal SAR, typically covers only 30% to 60% of the applicator face, normally with high-SAR “hot spots” located centrally, surrounded by a “cool” periphery.^170,171 These single-aperture waveguide devices are useful for only small, superficial tumors up to approximately 3 cm in diameter.⁸ One group has placed variable absorption bolus pads in front of the aperture to reduce SAR centrally and achieve a larger area of effective heating.¹⁷² The HTS-100 Lens Applicator has movable metal plate “vanes” within the concave opening of a single waveguide horn aperture to form convergent 430-MHz beams for heating small tumors up to 3 to 5 cm deep.¹⁷³

A variety of approaches have been taken to build applicators capable of heating larger areas, such as computer-controlled motorized scanning of one or two spiral microstrip MW antennas repetitively over the target surface^174–176 with thermal feedback control systems. Alternatively, several groups have developed multiaperture array waveguide applicators for large-area superficial disease such as the Microtherm 1000 (Labthermics Technologies, Urbana, IL)—a 16-element, 915-MHz waveguide planar array,¹⁷⁷ and a somewhat deeper heating array of 433-MHz Lucite Cone Applicators.^178,179 Other more flexible tissue-conforming, large-area MW array applicators have been constructed and used clinically, including a 433-MHz contact-flexible microwave applicator^180,181; a hinged array of 433-MHz current sheet applicators¹⁸²; a 915-MHz, 25-element spiral microstrip conformal array¹⁸³; and subsequent commercial implementations (Fig. 74-10)¹⁸⁴; and the largest 915-MHz conformal microwave array (CMA) applicator.^168,185 Fig. 74-11 shows a 6-mm thick water bolus vest and flexible printed circuit board CMA applicator being set up on a patient with chest-wall recurrence of breast carcinoma extending across the left upper back and shoulder. The patient is shown at right with an elastic overgarment to secure the applicator in place, allowing the patient freedom to stand, sit, or lie down during heat treatment while providing effective temperature control of the skin surface.¹⁸⁶ This applicator is essentially transparent to photon and electron radiation and thus may be used to apply heat simultaneously with radiation for potentially much higher TER.^187–189 Recent developments are beginning to produce EM superficial heating applicators with the capability of noninvasive thermal monitoring of surface¹⁹⁰ and subsurface^191–194 temperature distributions for feedback control of multiple antenna arrays.

FIGURE 74-10 • The BSD 500 microwave hyperthermia system with 24 antenna conformal array applicator treating chestwall recurrence. The applicator is coupled to the patient through a temperature-controlled water bolus.

(Photo courtesy BSD Medical, Inc.)

FIGURE 74-11 • Custom conformal microwave array (CMA) applicator treatment setup showing: 6-mm thick 42.5° C circulating water bolus vest (A), 12-aperture printed circuit board CMA applicator over large area back lesion (B), and patient undergoing hyperthermia treatment with CMA applicator as monitored with fiberoptic sensors mapped through catheters lying on the skin surface under the antenna array (C).

Ultrasound Techniques

Early efforts to apply US for superficial heating applications used single-round disk “piston” style transducers operating at 0.5 to 3.5 MHz.¹⁹⁵ These devices were capable of heating only small, superficial lesions less than 3 cm in diameter.⁸ Later, multitransducer arrays were constructed to increase the lateral extent and control of heating. The Sonotherm 1000 (Labthermics Technologies, Urbana, IL) is one such device with 16 individual transducers arranged in a 15-cm square 4 × 4 element planar array coupled to the patient with an attached, degassed water bolus bag.¹⁶³ This device has been used successfully to heat tumors up to 4 to 6 cm in depth.

Another approach uses arrays of focused transducers. The scanned focused US system uses a small number of low-frequency transducers (≤2 MHz) that produce a small focal spot at depth that is mechanically scanned rapidly around the tumor volume.¹⁹⁶ The addition of less-penetrating 4-MHz transducers and a patient pain feedback button to turn off power for short segments of the scan path improved the heating uniformity of this scanning US technique and yielded higher average tumor temperatures.¹⁹⁷ Another innovative US system has a computer-controlled scanning reflector for distributing US energy at two different frequencies (e.g., 1 and 5 MHz) over large surface areas.¹⁹⁸ With only water and plexiglas over the tumor surface, this system allows simultaneous HT and RT for potentially much higher TER.^199–201 Development of high-intensity focused US for heating small- to medium-size regions at depth continues, with most commercial development supporting thermal ablation applications.^202,203

Local Heating—Interstitial Sources

For applications requiring precise localization of heat in regions that are too difficult to access externally, there are at least nine distinctly different interstitial heating technologies. Excellent reviews are available.^{166,204–206}

Microwave Antennas

Interstitial MW antennas are usually constructed from flexible or semirigid coaxial cable less than 1.5 mm in diameter that fits snugly inside afterloading plastic implant catheters. The most common types are dipole antennas, which can be used in phased arrays to enlarge the heated volume,^207–209 and helical coil antennas, which tend to restrict heating more effectively to the region immediately surrounding the active coil length of the antenna.^210,211 Frequencies between 433 and 2450 MHz are normally used for interstitial MW heating. Radial penetration of heating may be extended by several millimeters with air or water cooling of the implanted antenna surface, but generally multiple antennas required for heating larger tumors up to 5 cm across should be placed within 1.5 to 2 cm from each other.²¹² Implantable MW antennas have been most useful in high-perfusion tissues or when nonparallel or flexible catheter implants are required. Recent developments have produced interstitial MW antennas with integrated MW radiometry for real-time monitoring of tissue temperature around each antenna.^213,214

Radiofrequency Electrodes

RF electrode systems heat tissue with resistively or capacitively coupled electric currents between implanted metal electrodes. Initial work used arrays of implant needles up to 20 cm long spaced 1 to 1.25 cm apart connected to a single RF power amplifier operating at less than 1 MHz. Subsequent systems used time-sequenced computer controlled localized current field (LCF) sources that distributed power to multiple electrode pairs around the implant array.²⁰⁴ Systems with segmented RF electrodes,^215,216 printed circuit board implant templates to simplify power connections, and computer-control programs allow simultaneous long-duration heat and brachytherapy.^204,217 At frequencies above 10 MHz, current may be coupled capacitively through a thin layer of insulating plastic, resulting in a more uniform distribution of heating current, as in the 64-channel, 27.12-MHz capacitively coupled RF (CC-RF) electrode system.^216,218,219 For both LCF and CC-RF systems, the radial falloff of power deposition is extremely steep so that implant spacing is limited to less than 1.25 to 1.5 cm for HT applications. Methods of cooling the electrode surface and driving multiple electrode sets have proven useful for extending the effective heating volume, especially for thermal ablation applications in which higher temperatures can be tolerated near the sources.²¹²

Hot Source Techniques

Because the penetration of thermally conducted heat energy is limited, hot source techniques are appropriate only for tissues with low to moderate perfusion that can be implanted with closely spaced arrays of implants—generally no more than 1.0 to 1.25 cm apart. However, heating length is almost unlimited. Hot source techniques include hot water perfusion through implanted catheters,^204,217 resistive wire implants heated by computer-controlled direct-current voltage sources,²²⁰ and ferromagnetic implants (FMIs) heated by an externally applied magnetic field at less than 1 MHz.^221–223 An advantage of the FMI approach is that the implants can be made of special high-permeability alloys that operate near their Curie point transition from magnetic to nonmagnetic such that they automatically thermoregulate over a narrow range of temperature.^224,225 The ferromagnetic material providing this controlled temperature can be a needle or rod, 5- to 10-mm long cylindrical seeds interspersed with 5-mm long or less radioactive brachytherapy sources for simultaneous thermobrachytherapy, or ferromagnetic fluids or nanoparticles that are either inserted directly in the tumor in one or more needle tracks or are delivered systemically.^226,227 The FMI techniques simplify reheating of permanently implanted seeds,²²⁸ rods,²²⁹ or nanoparticles²³⁰ because no externalized connections are required for either power or thermal monitoring of the thermoregulating implant material.

Ultrasound Radiators

Interstitial US applicators appear close to revolutionizing interstitial HT (IHT) delivery for many sites. Linear arrays of 0.5- to 1-cm long tubular transducers have been constructed either with water cooling²³¹ or direct coupled sources with an open central lumen to accommodate brachytherapy sources for simultaneous thermobrachytherapy.²³² Implant spacing up to 2.5 cm apart should be possible for the water-cooled version. These applicators allow precise power control along the implant length as well as radially directional heating capability. Fig. 74-12 shows an example of a clinical treatment of prostate with an array of different-length US array applicators that fit inside 13-g brachytherapy catheters for highly steerable power deposition. Implantable US arrays have been used for treatment with real-time, noninvasive MRI monitoring of both temperature and tissue property changes during heating.^233–235

FIGURE 74-12 • Interstitial ultrasound applicators designed to deliver dynamic three-dimensional control of power deposition from within plastic high-dose-rate brachytherapy catheters (top). Treatment plan of peripheral implant of prostate demonstrating conformal distribution of heating within the implant volume, with energy directed away from rectum. Computed tomography and patient treatment setup for directional heating of prostate (bottom).

(Courtesy C.J. Diederich.)

Regional Heating

The most common regional heating approaches involve the use of radiated EM fields in the frequency range of 10 to 140 MHz in which the wavelength in tissue is large compared with the height of the human body, so that focusing of heat to a small region within the torso is generally not possible. These techniques are intended to heat a large region at depth and usually produce temperatures of approximately 39° C to 42° C throughout the tumor and surrounding region, limited by power deposition in critical normal tissues.

Early regional heating devices included the single large 27-MHz ridged waveguide aperture²⁴² and the single 82-MHz torso-sized helix coil.²⁴³ Using a 70-MHz coaxial transverse electromagnetic (TEM) applicator, the entire patient is positioned inside a hollow 60-cm diameter chamber filled with coupling water, and partial steering of SAR is possible by shifting the patient’s position within the chamber.^244,245 Another single-power-source, deep-heating technique uses 8 to 27 MHz RF capacitive heating, such as the Thermotron RF-8.^246,247 This device creates an electric field between two saline bolus–coupled metal plate electrodes on opposite sides of the patient. Power deposition is maximum in the high-resistance fat layer, but this problem can be overcome with aggressive skin cooling in patients with a fat layer less than 1 to 2 cm thick.²⁴⁸ Limited steering of power deposition is possible through adjustment of water temperature, electrode placement, and by using a smaller electrode on one side to concentrate heating.

It is possible to localize heating in a region at depth by using an array of phase-focused radiating antennas. The first commercial device built for this approach was the Annular Phased Array (BSD Medical Corp, Salt Lake City UT), which consisted of an array of four equal-power, equal-phase twin dipole radiators mounted on the surface of a large cylinder coupled tightly around the body torso with a refillable distilled water bolus.²⁴⁹ Subsequent development of the 75- to 120-MHz Sigma 30, 60, and Ellipse applicators improved the patient interface and increased control flexibility via four independent relative phase and amplitude controls for the eight radiators.^250,251 A similar device, the Matched Phased Array system, has four large 70-MHz waveguide sources with flexible positioning as well as phase and amplitude steering of power deposition within the body.²⁵² Treatment planning systems have been developed to take advantage of this control flexibility and improve the quality of regional heating.^253,254 Although these regional heating devices are still in clinical use, newer systems are now available, such as the 100-MHz SigmaEye applicator (BSD Medical Corp, Salt Lake City UT), with 12 independently-controlled twin dipole radiators mounted in three rings to allow phase and amplitude steering both longitudinally and radially within the body.²⁵⁵ More recently, the SigmaEye system has been integrated with both MRI and treatment planning systems. These hybrid SigmaEye/MRI systems (Fig. 74-13) allow MRI-based preplanning as well as noninvasive monitoring of temperature rise during HT to facilitate beam steering into the tumor.^256,257 Smaller versions of the Sigma series annular phased array applicators have been implemented for MRI guided HT of tumors in the extremities.^258–261 Application of MRI-guided heating of breast is already available for thermal surgery applications^262,263and should be available for deep-breast HT soon.²⁶⁰

FIGURE 74-13 • The “SigmaEye/MR” deep regional heating device with 12 twin dipole antennas mounted in three rings around a water bolus–filled elliptical form and driven at approximately 100 MHz for deep heating in the pelvis. Phase and amplitude can be adjusted independently for each dipole pair of each of the three rings to steer power deposition laterally and longitudinally, based on control feedback from magnetic resonance images of temperature rise during treatment.

(Courtesy BSD Medical Corp.)

Thermometry and Quality Assurance

Appropriate monitoring of tissue temperature during HT is an essential component of therapy, and successful analyses correlating tumor response with thermal parameters require detailed knowledge of tumor temperature distributions. Excellent reviews of thermometry techniques for treatment monitoring and control are available.^{160,161,166,168,268}

Thermal Dose and Applicator Adequacy Standards in Clinical Hyperthermia

Clinical HT, as distinguished from high temperature thermal ablation, targets volumes that include dose-limiting normal tissue. The clinical objective is to achieve tumor regression in a controlled fashion, with normal tissue regenerating and replacing the space left by receding tumor. Just as similar clinical constraints lead to restrictions on radiation dose per fraction, there are limits on target temperature range, as well as the number and duration of HT sessions. The typical tumor time/temperature objective is usually approximately 45 to 60 minutes at approximately 41° C to 45° C, with normal tissue temperatures limited to less than approximately 45° C. The need to limit the temperature of normal tissue within or close to the target volume means that one may not arbitrarily increase the total applied power to ensure adequate minimum tumor temperature. This means that the distribution of applied power (SAR) within the target volume must be sufficiently uniform to keep temperature within range and still avoid areas of tumor underdosage. Applicator adequacy is discussed in the next paragraph. In addition to using an appropriate applicator, it is important to have adequate thermometry. This is because perfusion at a given location can vary during an HT session and can change from session to session. Thus, unlike RT, it is not possible to predict in advance of a treatment the delivery of a biologically meaningful dose. The variability of heat dissipation is also a factor in determining the number of HT treatments. Tumor locations that receive adequate SAR but achieve suboptimal temperature in one session may do better in another session if the local perfusion is less on another day (whether because of cumulative effects of thermoradiotherapy or random fluctuation). Implications for the number of HT sessions will also be discussed in the following text.

Results With Hyperthermia Alone

Early clinical studies using HT alone for the treatment of small, superficial neoplasms showed that HT alone (using multiple 30-60 minute HT sessions) has very limited efficacy. Overgaard³¹⁴ summarized the results of 18 such studies including 380 tumors treated with 2 to 22 heat sessions with a CR rate of 13% and a partial response (PR) rate of 36%. The duration of HT responses was generally brief.

Hyperthermia and Radiation

Summary of Randomized Human Studies of Radiotherapy and Hyperthermia

Table 74-2 summarizes major published prospective, randomized human HT trials of RT with or without HT.^{8–12,335,362,367–371} Considering all trials, more than 1700 patients were treated in 14 trials with a CR rate of 40% for RT alone versus 52% for RT plus HT. Overgaard³⁷² performed a meta-analysis of 22 known trials, comparing the risk of failure for patients treated with RT plus HT versus RT alone. Although a few of the individual trials were inconclusive, the combined data indicated a significantly reduced risk of failure in patients who received adjuvant HT, with an odds ratio of 1.77 within 0.20 (95% confidence interval) and a P-value of less than 0.00001. These data are grouped by disease site and presented in graphical form in Fig. 74-14, showing clear evidence for a benefit of adjuvant HT for melanomas and head and neck, chest wall, cervical, rectal, and bladder cancers, but no benefit for prostate or intact breast cancers. Results of the thermal dosimetry studies suggest that improved heating of breast and prostate tissue possible with developing equipment should improve clinical response.

FIGURE 74-14 • Graphical representation of a meta analysis of risk of failure in 22 prospective, randomized trials comparing radiotherapy alone to radiation plus hyperthermia. The number of patients included in the trials is plotted versus the odds ratio with the 95% confidence interval, grouped by disease site. There is clear evidence for a benefit of adjuvant hyperthermia for melanomas and head and neck, chest wall, cervix, rectum, and bladder cancers.

(Redrawn from Jens Overgaard, unpublished.)

Hyperthermia and Chemotherapy

Multiple, mostly small, phase II trials have been performed combining local, endoluminal, regional, or WBH with chemotherapy for treating esophageal cancer preoperatively, stomach cancer, pancreatic cancer, metastatic or high-risk soft-tissue sarcomas, recurrent cervical cancer, advanced rectal cancer, miscellaneous advanced or metastatic cancers,^373,374 small cell lung cancer,³⁷⁵ disseminated melanoma,³⁷⁶ and ovarian cancer.^377,378 A phase III randomized trial comparing preoperative chemotherapy to preoperative chemotherapy plus HT was performed in 40 esophageal cancer patients showing higher rates of subjective and radiographic improvement (70% versus 40% and 50% versus 25%, respectively) and a higher proportion of necrosed cancer cells on histologic examination in the combined therapy arm.³⁷⁹ Following up on promising phase II chemothermotherapy trials in patients with high-risk soft-tissue sarcomas, ESHO and the European Organization for Research and Treatment of Cancer are conducting a phase III trial of neoadjuvant chemotherapy (etoposide, ifosfamide, and DOX) with or without regional hyperthermia (RHT) prior to resection and RT.³⁸⁰

Preoperative hyperthermic isolated limb perfusion (HILP) with melphalan, tumor necrosis factor, or other agents has been used extensively in patients with high-risk or recurrent extremity melanomas. One prospective, randomized trial was closed early because of significantly higher local control and survival in patients treated with preoperative melphalan HILP compared with those who underwent surgery alone.³⁸¹ A large phase III, multicenter, multicontinent randomized trial in 832 assessable patients showed a trend toward longer disease-free interval but no benefit in terms of survival or time-to-distant metastasis for HILP prior to surgery compared with surgery alone.³⁸²

HT has also been used in combination with drug-containing liposomes to improve drug delivery. Kong and Dewhirst have provided an excellent review of preclinical data¹¹⁶ and Park has reviewed early clinical data and rationale for liposome-based breast cancer treatment.³⁸³ A phase I and II trial was performed in 25 patients with chest-wall recurrence of breast cancer using superficial HT immediately before intravenous administration of DOX-containing liposomes (Doxil) at 45 mg/m² every 28 days for 4 to 6 cycles.^384,385 The response rate in heated fields was 57% with 1 CR and 12 PRs. Furthermore, HT was associated with significantly improved response in 18 patients with both heated and nonheated disease (P < 0.04). Another phase I and II trial was performed in 15 patients with advanced, chemotherapy-resistant cancers using fever-range WBH with DOX-containing liposomes, 5-fluorouracil, and low-dose metronymic interferon-α repeated every 28 days for one to nine cycles (median, three cycles).³⁸⁶ Six patients had objective responses, five had stable disease for more than 12 weeks, and four had progressive disease.

Hyperthermia, Radiation, and Chemotherapy

There is also growing literature on trimodality studies combining HT, RT, and chemotherapy for head and neck, breast, esophageal, and rectal cancers, including promising phase III studies for resectable or unresectable esophageal cancer.³⁷³ In one trial, 66 patients with resectable squamous cell carcinoma of the esophagus were randomized to preoperative chemoradiotherapy or thermochemoradiotherapy using twice weekly RF capacitive HT with an endoluminal applicator, cisplatin or bleomycin, and 30 Gy in 15 fractions for 3 weeks.³⁸⁷ Operative specimens showed absence of viable cancer cells in 25% of the trimodality patients versus 6% of the chemoradiotherapy patients, and 3-year survival rates were 50% versus 24%, respectively. Kouloulias and colleagues³⁸⁸ have reported a promising initial study of trimodality approach in recurrent breast cancer. Colombo and colleagues³⁸⁹ compared intravesical mitomycin C chemotherapy alone versus chemotherapy plus transurethral MW HT for 52 patients with superficial bladder tumors. There was a significant improvement in response and disease-free survival. To date, experience with thermochemotherapy and thermochemoradiotherapy is too limited to draw conclusions, although preliminary results are encouraging.

Heat Shock Proteins and Immunotherapy

The fact that vaccines against tumor HSPs are applicable to the treatment of cancer has opened up a new perspective in cancer immunotherapy. A number of preliminary clinical trials have been carried out to test the use of HSPs derived from autologous human tumors in immunotherapy of human cancer: Janetzki and colleagues prepared individual gp96 vaccines derived from the tumors of 16 patients with advanced malignancies³⁹⁰; Amato and colleagues have treated patients with renal cell carcinoma with gp96 vectors.³⁹¹ Eton and colleagues also used gp96 vaccines in metastatic melanoma using a similar dosing regimen as described for renal cancers.³⁹² However, all of these clinical trials are nonrandomized and the clinical data generated thus far can only be suggestive. To the extent that one may draw any inferences from them, the results so far are consistent with the murine data that showed that the autologous gp96 vaccine was effective in the adjuvant setting in which most treated animals remained disease-free for their entire life span, and less effective in animals with progressive tumors, in which only disease stabilization was achieved.¹²⁷ Most recently, double-blind phase III clinical trials using gp96-peptide complex for the treatment of stage III renal cell carcinoma in the adjuvant setting are being conducted at numerous centers worldwide.³⁹³

The advantages and disadvantages of various vaccination approaches are discussed in a recent review article by Manjili and colleagues.³⁹⁴ In addition, several other excellent reviews of this field emphasize the development, analysis, and application of HSP-related vaccination approaches.^{127,395–399}

Randomized Animal Studies

In comparison with human trials, studies in pet animals often allow for better acceptance of novel therapies and randomization, more detailed thermometry, better control over the heat treatment, and better tolerance resulting from the use of general anesthesia, resulting in higher minimum tumor temperatures. Four prospective, randomized trials performed in pet animals with spontaneous tumors all demonstrated significant benefit for HT in combination with RT over RT alone.^29,400–402 A University of Arizona trial included 236 dogs and cats with a variety of spontaneous cancers stratified by histologic type and randomized to receive RT alone (36.8 Gy in eight fractions over 4 weeks) or RT plus once-weekly HT for 30 minutes. For 217 evaluable animals, the CR rate was 35% in the RT arm and 59% in the RT plus HT arm, with no increase in acute or late effects.²⁹

Denman and colleagues⁴⁰⁰ performed a phase III randomized trial in spontaneous canine malignancies stratified by histologic type and anatomic site and randomized to receive RT (49.0 Gy in 14 fractions given on Mondays, Wednesdays, and Fridays) with or without once-weekly HT for 30 minutes. HT significantly increased the CR rate for sarcomas (33% for RT alone versus 73% for RT plus HT), tumors of the trunk and extremities (45% versus 83%), and tumors less than 10 cm³ in volume (50% versus 84%).

Gillette and associates⁴⁰¹ randomized oral carcinomas in 38 dogs to graded radiation doses (25-45 Gy in 10 fractions for 22 days) with or without seven 30-minute HT treatments to determine the dose that would result in a 50% probability of tumor control (TCD₅₀) and then to determine the TER as the ratio of the TCD₅₀ for RT alone to the TCD₅₀ for RT plus HT. To limit normal tissue temperatures, the tumor margins did not exceed 41.5° C. The TER was 1.15 and the slope of the dose-response curve was much steeper for RT and HT than for RT alone. Late complications were not increased in the animals receiving heat.

A similar study was performed in 64 canines with spontaneous soft tissue sarcomas.⁴⁰² The animals were stratified by tumor volume and randomized to receive graded doses of RT (35.0-55.0 Gy in 10 fractions on a Monday-Wednesday-Friday schedule) with or without US HT for 30 minutes 3 hours after RT twice weekly on a Monday-Friday schedule. The HT goal was a minimum tumor temperature of 42° C. Freedom from local tumor progression was significantly increased in animals treated with RT plus HT (Wilcoxon P = 0.046; log rank P = 0.064). There was no significant difference between the two study arms in terms of late normal-tissue complications or distant-metastatic rate.

A fifth phase III study of RT (36 Gy at 9 Gy per fraction) with or without local HT for spontaneous head and neck tumors in 145 dogs showed no difference between the treatment arms in terms of CR rate (47% for RT versus 43% for RT and HT)⁴⁰³ or progression-free survival (median, 26 weeks for RT versus 27 weeks for RT and HT). The authors cited inadequate thermal dose as a probable cause for the negative results.

Thus all but one of these well-controlled, randomized studies show convincingly that adjuvant HT enhances the response and control rates of spontaneous malignancies in anesthetized animals, in which adequate minimum thermal doses are more often achievable than in nonanesthetized human patients.

WBH has also been investigated in combination RT or chemotherapy in dogs. A phase III trial of RT with or without WBH performed in 45 dogs with spontaneous brain tumors resulted in 1-year survival probabilities of 44% for RT versus 40% for RT and WBH, and six dogs had severe toxicity from WBH.⁴⁰⁴ In another study, 64 dogs with spontaneous soft-tissue sarcomas were randomized to RT with localized HT with or without WBH to improve T_min and T₉₀ temperatures. The addition of WBH significantly increased CEM 43° C T₉₀, but also resulted in more toxicity and significantly shorter time to metastasis (P = 0.02).⁴⁰⁵ However, WBH was associated with a trend toward improved disease-free survival among 46 dogs with high-grade lymphoma achieving a CR with DOX and then randomized to five additional DOX treatments with or without WBH.^406,407