International Journal of

Radiation Oncology biology

physics

www.redjournal.org

Physics Contribution

Quantifying the Combined Effect of Radiation Therapy and Hyperthermia in Terms of Equivalent Dose Distributions H. Petra Kok, PhD,* Johannes Crezee, PhD,* Nicolaas A.P. Franken, PhD,*,y Lukas J.A. Stalpers, MD, PhD,* Gerrit W. Barendsen, PhD,y and Arjan Bel, PhD* *Department of Radiation Oncology, and yLaboratory for Experimental Oncology and Radiobiology (LEXOR)/Center for Experimental and Molecular Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Received Jul 26, 2013, and in revised form Nov 8, 2013. Accepted for publication Nov 10, 2013.

Summary Hyperthermia is the clinical application of heat, raising tumor temperatures to 4045 C. This study aimed to develop a method to quantify radiosensitization by hyperthermia. Clinical radiation therapy dose distributions with hyperthermia were converted to equivalent dose distributions without hyperthermia for 15 prostate cancer cases. Results indicate that adding hyperthermia is similar to a dose escalation. After further clinical validation, this model will be useful to estimate the effects of interaction from different cancer treatment modalities.

Purpose: To develop a method to quantify the therapeutic effect of radiosensitization by hyperthermia; to this end, a numerical method was proposed to convert radiation therapy dose distributions with hyperthermia to equivalent dose distributions without hyperthermia. Methods and Materials: Clinical intensity modulated radiation therapy plans were created for 15 prostate cancer cases. To simulate a clinically relevant heterogeneous temperature distribution, hyperthermia treatment planning was performed for heating with the AMC-8 system. The temperature-dependent parameters a (Gy1) and b (Gy2) of the linearequadratic model for prostate cancer were estimated from the literature. No thermal enhancement was assumed for normal tissue. The intensity modulated radiation therapy plans and temperature distributions were exported to our in-house-developed radiation therapy treatment planning system, APlan, and equivalent dose distributions without hyperthermia were calculated voxel by voxel using the linearequadratic model. Results: The planned average tumor temperatures T90, T50, and T10 in the planning target volume were 40.5 C, 41.6 C, and 42.4 C, respectively. The planned minimum, mean, and maximum radiation therapy doses were 62.9 Gy, 76.0 Gy, and 81.0 Gy, respectively. Adding hyperthermia yielded an equivalent dose distribution with an extended 95% isodose level. The equivalent minimum, mean, and maximum doses reflecting the radiosensitization by hyperthermia were 70.3 Gy, 86.3 Gy, and 93.6 Gy, respectively, for a linear increase of a with temperature. This can be considered similar to a dose escalation with a substantial increase in tumor control probability for high-risk prostate carcinoma. Conclusion: A model to quantify the effect of combined radiation therapy and hyperthermia in terms of equivalent dose distributions was presented. This model is particularly instructive to estimate the potential effects of interaction from different treatment modalities.  2014 Elsevier Inc.

Reprint requests to: H. P. Kok, PhD, University of Amsterdam, Academic Medical Center, Department of Radiation Oncology, Meibergdreef 9, 1105

Int J Radiation Oncol Biol Phys, Vol. 88, No. 3, pp. 739e745, 2014 0360-3016/$ - see front matter  2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2013.11.212

AZ Amsterdam, The Netherlands. Tel: (þ31) (0) 20-5664231; E-mail: H.P. [email protected] Conflict of interest: none.

740

Kok et al.

Introduction Hyperthermia, as used in clinical oncology, is the application of elevated temperatures (40-45 C) to tumor tissue to increase the therapeutic effect of radiation therapy and/or chemotherapy. This sensitizing effect is often referred to as thermal enhancement and quantified as thermal enhancement ratio, that is, the radiation dose required to obtain a given endpoint with radiation alone relative to the radiation dose needed for the same effect with combined hyperthermia and radiation (1). For several tumor sites the efficacy has been demonstrated in clinical studies, with a clear thermal doseeeffect relationship (2-5). The 12-year survival for cervix uteri carcinoma patients almost doubled, from 20% to 37% (PZ.03) by adding hyperthermia to radiation therapy (3). Contrary to most anticancer treatments, hyperthermia has a low risk of side effects when administered properly. To avoid normal tissue complications for combined radiation therapy and hyperthermia treatments, the order of application is important. Overgaard (1) has shown that a favorable therapeutic ratio between tumor and normal tissue can be achieved when hyperthermia is applied after radiation therapy. Hyperthermia is usually given once or twice weekly during the radiation therapy course. Results for once- or twice-weekly hyperthermia schedules have been reported to be similar (6). More frequent sessions will not increase thermal enhancement of the tumor for sequential application (7). Furthermore, an interval of at least 72 hours between the hyperthermia sessions avoids thermal enhancement of normal tissue (7). Deep-seated tumors, such as pelvic carcinomas, are generally heated with phased-array systems operating at a frequency between 60 and 140 MHz. Examples of modern 3-dimensional steered heating systems are the BSD Sigma Eye (BSD Medical Corporation, Salt Lake City, UT) (8) and the AMC-8 system (Academic Medical Center, Amsterdam, The Netherlands) (9). The pursued thermal dose is a homogeneous tumor temperature of 43 C for 1 hour (10), but the clinically achieved temperature distribution is heterogeneous owing to inhomogeneities in tissue properties and blood flow. Despite the clinical evidence for hyperthermia as a very effective radiosensitizer with few side effects, it is not widely applied as a means to improve the outcome of radiation therapy. An important argument to consider hyperthermia is that new treatment delivery techniques in radiation therapy, combining image guidance with intensity modulated (IMRT) beams, result in better dose conformity but inevitably also yield a higher integral dose to organs at risk. Radiosensitization by hyperthermia might be considered as an alternative for dose escalation to improve tumor control for locally advanced prostate cancer patients (11). The potential benefit of hyperthermia combined with radiation therapy for prostate cancer was demonstrated in several phase 1/2 studies, and it was shown to be feasible without significantly increased toxicity (11-14). Hurwitz et al (11) reported for radiation therapy combined with hyperthermia a similar improvement in disease-free survival compared with dose escalation, but without the toxicity associated with dose escalation. On the basis of these encouraging results a multi-institutional phase 3 trial is designed to investigate the effectiveness of hyperthermia for this tumor site (14). The purpose of this study was to develop a generally applicable model and planning software to quantify the combined effect of radiation therapy and hyperthermia in terms of equivalent dose distributions. A numeric method is presented to convert radiation

International Journal of Radiation Oncology  Biology  Physics therapy dose distributions with hyperthermia to equivalent dose distributions without hyperthermia. The patient group considered consists of prostate carcinoma patients, but the developed method and software will be generally applicable.

Methods and Materials The present study focuses on modeling the effect of hyperthermia in combination with radiation therapy in terms of equivalent dose distributions, using the linearequadratic model (LQ model). The potential clinical application will be simulated for 15 prostate cancer cases. Clinically applied IMRT dose distributions were compared with predicted equivalent dose distributions when additional hyperthermia would have been given with the AMC-8 system, applied once or twice weekly after radiation therapy. This system consists of 2 rings of 4 waveguides and can be used both as a single-ring and a double-ring device. A single-ring device was considered in this study.

Treatment planning Radiation therapy treatment planning A planning CT scan with 2.5-mm slice thickness was acquired with the patient in the supine position. Clinical IMRT plans were created with PLATO (Nucletron, Veenendaal, The Netherlands). Patients received a total dose of 70 Gy on the planning target volume (PTV), delivered in 35 fractions of 2 Gy, with an integrated boost delivering 77 Gy to the prostate. These plans were used for the prostate cases in the present study. Treatment plans were exported to our in-housedeveloped research planning system, APlan. APlan uses an improved version of a pencil-beam algorithm in combination with an equivalent tissue-air-ratio (ETAR) correction for inhomogeneities and graphical processing unit (GPU)-accelerated ray-tracing (15). This algorithm was derived from PLATO. An additional module was developed to import 3-dimensional temperature distributions and analyze combined treatments.

Hyperthermia treatment planning Hyperthermia treatment planning was performed using our inhouse-developed software. A histogram of the Hounsfield units of the planning CT scan was used for tissue segmentation (16). Literature-based dielectric and thermal properties (Table 1) were assigned to muscle, fat, bone, air, and tumor (17, 18). For practical reasons of computation time and computer memory, the patient model was downscaled to 2.5  2.5  5 mm3. Electric field distributions were calculated by solving Maxwell’s equations with the Finite Difference Time Domain method. Antenna settings that yield an optimal steady-state tumor temperature distribution were determined using temperature-based optimization (19). The Pennes model was applied for bio heat transfer computations (20). A temperature of 43 C in the PTV was pursued byX minimizing the objective function: ðmaxð43  Tðx; y; zÞ; 0ÞÞ2 ; ðx; y; zÞ˛ PTV ½1 target

Normal tissue temperatures were constrained to 45 C. An optimized temperature distribution is shown in Figure 1, together with a segmented patient anatomy and the AMC-8 system in clinical use. Tumor temperature distributions were expressed in terms of T10, T50, and T90, that is, the lowest temperature achieved throughout 10%, 50%, and 90% of the tumor volume, respectively.

Volume 88  Number 3  2014 Table 1

741

Values of the dielectric and thermal properties for different tissue types at 70 MHz*

Tissue type Air Bone Fat Muscle Tumor

Dose escalation by adding hyperthermia

s (S m1)

εr (-)

r (kg m3)

c (J kg1  C1)

0 0.05 0.06 0.75 0.74

1 10 10 75 65

1.29 1595 888 1050 1050

1000 1420 2387 3639 3639

k (W m1  C 0.024 0.65 0.22 0.56 0.56

1

)

Wb (kg m3 s1) 0 0.12 1.1 3.6 1.8

* Used in the simulations (17, 18); conductivity (s [S m1]), relative permittivity (εr [-]), density (r [kg m3]), specific heat capacity (c [J kg1  C1]), thermal conductivity (k [W m1  C1]), and perfusion (Wb [kg m3 s1]).

Calculation of the equivalent dose distribution

Values of the LQ parameters a and b

(Table 2). Hyperthermia is a potent radiosensitizer, but the extent of sensitization depends on the achieved temperature and varies per cell line. From Table 2 it can be observed that hyperthermia can change both a and b. As a first estimate, b was kept constant in this study. This is acceptable because in general hyperthermia-induced changes in a are more pronounced (24, 25). This article focuses on the potential efficacy of hyperthermia for prostate cancer. There is some debate about the LQ parameters for prostate tumors (26), but it is generally accepted that the a/b ratio is low. Normothermic values for a and b of a37 Z 0.0391 Gy1 and b37 Z 0.0263 Gy2 were applied (27). Three temperature dependencies of a were evaluated. The most straightforward assumption is a linear dependency: 1:5a37  a37 alin ðTÞZa37 þ ðT  37Þ ½5 41  37 The slope was chosen such that a at 41 C was increased with a factor 1.5 compared with its reference value at 37 C. This is a conservative estimate, both in view of Table 2 and compared with results from Myerson et al (28), who derived a heat-induced radiosensitization from clinical studies for other tumor sites, represented by a change in a of approximately 0.05-0.1 Gy1. The clinical effect of hyperthermia depends nonlinearly on temperature (29), implying a nonlinear temperature dependency of a. A linear function might overestimate the effect of hyperthermia at lower temperatures and underestimate it at higher temperatures. Therefore, an exponential temperature dependency was studied as well: 2 16 aexp ðTÞZa37 eðT37Þ =T0 ; T0 Z alin ð41Þ ½6 ln a37

Radiosensitization by hyperthermia can effectively be expressed by the LQ parameters a and b. However, literature studies rarely focus on deriving a and b values at hyperthermic temperatures. Some studies reported values at a specific temperature (24, 25)

This exponential function models a stronger thermal enhancement for temperatures above 41 C (Fig. 2). The third option was motivated by the fact that calculation of the thermal dose of a hyperthermia treatment is based on the number of equivalent minutes at 43 C (CEM43). Calculation of CEM43 uses

The LQ model was applied to calculate equivalent dose distributions for combined radiation therapyehyperthermia treatments (21). This model expresses the survival fraction SF(n, d, a, b) after delivering n fractions with fraction dose d (Gy) by 2 SFðn; d; a; bÞZenðadþbd Þ ½2 In this model, the number of lethal lesions is the sum of the lethal lesions produced from a single radiation track (which are linearly related to dose, ad ) and the lethal lesions produced from 2 radiation tracks (which are quadratically related to dose, bd2). When N0 is the number of malignant cells, the probability to eradicate all tumor cells can be calculated by the tumor control probability TCP: TCPZeN0 SFðn;d;a;bÞ ½3 This Poisson TCP model is frequently used in radiation therapy literature (22, 23). Because hyperthermia is a radiosensitizer, the parameters a and b are temperature dependent, and thus TCP is a function of the temperature T. To calculate the equivalent fraction dose with hyperthermia dHT that would yield a similar therapeutic effect as the conventional dose, the TCP with hyperthermia has to be equal to the TCP without hyperthermia, yielding an equivalent fraction dose pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi a þ a2 þ 4bdðaðTÞ þ d bðTÞÞ ½4 dHT Z 2b This conversion is independent of the number of fractions.

Fig. 1. Left: The AMC-8 system in clinical use. The system consists of 2 rings of 4 waveguides. The amplitude and phase of each waveguide can be adjusted separately for 3-dimensional power steering in the patient. Right: Example of segmented patient anatomy and an optimized temperature distribution for heating with a single ring of the AMC-8 system.

742

International Journal of Radiation Oncology  Biology  Physics

Kok et al.

Table 2 Effect of hyperthermia on the average values of a and b, compared with their values with radiation therapy (RT) alone for various tumor cell lines Cell line

Treatment

Lung (SW-1573)

Colon (RKO) Prostate (PC-3) Prostate (DU-145)

RT RT þ 1 h RT þ 1 h RT RT þ 1 h RT RT þ 1 h RT RT þ 1 h

41 C 43 C 41 C 44 C 44 C

a (Gy1)

b (Gy2)

0.21 0.06 0.49 0.55 0.93 0.18 0.43 0.13 0.1

0.06 0.11 0.12 0.02 0.05 0.37 0.25 0.056 0.1

See references 24 and 25.

a transition temperature of 42.5 C (29). Thus, a linear variation was modeled with different slopes below and above 42.5 C.

alin2 ðTÞZalin ðTÞ T  42:5 C alin2 ðTÞZalin ð42:5Þþ

aexp ð45Þ alin ð42:5Þ 4542:5

ðT 42:5Þ

½7a T >42:5 C ½7b

Because clinically achieved tumor temperatures usually vary around 41 C, Equations 5-7a were chosen to yield the same enhancement at 41 C for a fair comparison. Equations 6 and 7b yield the same value at 45 C. Their behavior within the hyperthermic temperature range is different.

Dose distributions for prostate cancer cases Equivalent dose distributions with and without hyperthermia were evaluated for 15 prostate cancer cases. For this purpose the temperature distributions were resampled to the original CT resolution using trilinear interpolation. Dose conversion was applied on a voxel-by-voxel basis, and the equivalent minimum, mean, and maximum dose in the PTV were compared. The equivalent dose in normal tissue remains the same, because thermal enhancement of normal tissue was neglected. This is justified because sequential application of hyperthermia after radiation therapy, with 1 or 2 weekly sessions is assumed, which yields no significant effect

Fig. 2. Plot of the different temperature dependencies of a examined in this study.

on normal tissue complications (1, 14). The evaluation was performed for the 3 different temperature dependency models of a (Eqs. 5-7). Furthermore, the coverage of the 95% isodose levels (ie, 95% of the dose in the isocenter) was compared with and without radiosensitization by hyperthermia.

Results The optimized temperature distributions for the 15 prostate cancer cases yielded on average a T90, T50, and T10 of 40.5 C  1.1 C, 41.6 C  1.3 C, and 42.4 C  1.5 C in the PTV, respectively (cf. Fig. 3). These temperatures show a realistic heterogeneity consistent with clinically realized temperatures for pelvic tumors. Equivalent dose distributions for radiation therapy combined with hyperthermia were evaluated. Figure 3 shows a histogram of the minimum, mean, and maximum dose for radiation therapy alone and radiation therapy combined with hyperthermia for the 3 different temperature dependencies of a. Results averaged over the 15 cases show that addition of hyperthermia effectively escalates the radiation dose. The linear model uses the most conservative estimate of the enhancement of a with temperature, and the minimum, mean, and maximum radiation doses increased from 62.9, 76.0, and 81.0 Gy, respectively, to equivalent doses of 70.3, 86.3, and 93.6 Gy, respectively, by adding hyperthermia. This implies that the equivalent mean delivered dose is roughly 10 Gy higher compared with radiation therapy alone. An assumed exponential or dual linear behavior of a yields a higher mean and maximum equivalent dose compared with the linear model. Additionally, the standard deviations of the mean and maximum equivalent dose become higher, because the temperature dependency becomes more pronounced. TCP curves based on clinical data show that an increase in equivalent dose from 76 to 86 Gy yields an increase in local control from approximately 90% and approximately 85% to almost 100% for low- and intermediate-risk prostate carcinoma, respectively (30). For highrisk prostate carcinoma the local control increases substantially. According to a doseeresponse curve based on clinical data, the local control can increase from approximately 50% to approximately 90% (30), but it should be noted that androgen deprivation is a confounding factor for high-risk disease. Adding hyperthermia yields an extension of the equivalent high-dose region around the PTV. Figure 4 shows examples for 2 cases with different tumor temperatures and equivalent dose distributions. The equivalent dose was calculated using alin, but aexp and alin2 also yield an extension of the high-dose region. For case 1 the minimum, mean, and maximum doses were 58.5, 75.8, and 82.3 Gy, respectively. This increased to 64.9, 81.7, and 88.7 Gy, respectively, when adding mild hyperthermia, with T90, T50, and T10 of 38.6 C, 39.5 C, and 40.4 C, respectively. As observed from Figure 4, adding hyperthermia extends the 95% isodose level. For case 2 a better temperature distribution could be realized without overheating normal tissue. The T90, T50, and T10 were 41.5 C, 42.6 C, and 43.6 C, respectively. Radiation therapy alone yielded minimum, mean, and maximum doses of 64.7, 75.1, and 80.8 Gy, respectively. This increased to 74.9, 87.8, and 95.8 Gy, respectively, with hyperthermia. For case 2 the equivalent 95% isodose level extends into the rectum, but the evaluation of the equivalent dose is only meaningful for the PTV because no radiosensitizing effect of hyperthermia on normal tissue was assumed. This example shows that the extent of dose escalation depends on the radiation therapy dose and the tumor temperature distribution.

Volume 88  Number 3  2014

Dose escalation by adding hyperthermia

743

Fig. 3. Left: Minimum, mean, and maximum dose (SD) for radiation therapy alone and radiation therapy combined with hyperthermia for the 3 different temperature dependencies of a. Right: Average calculated temperatures (SD) in the planning target volume (PTV). Results averaged over the 15 cases.

Discussion This article presents a new method to estimate the effect of hyperthermia on radiation therapy dose distributions. By calculation of the equivalent fraction dose, the potential of hyperthermia to improve radiation therapy treatment outcome can be quantified. A well-known problem in radiation therapy dose (IMRT) optimization is the trade-off that has to be made close to organs at risk. For prostate irradiation the margin around the target defined to ensure sufficient target coverage will yield a rectum dose that often leads to complications. On the other hand, the use of very tight margins required with dose escalation might not be desirable for locally advanced prostate cancer patients. Additional hyperthermia could be an interesting option to overcome these problems. Several phase 1/2 studies combining radiation therapy and hyperthermia demonstrated feasibility without significantly increased toxicity (11-14). The clinical benefit of hyperthermia for prostate cancer will be investigated in a phase 3 trial (14). A study that applied hyperthermia with transrectal ultrasound before radiation therapy reported significant acute rectal toxicity,

limited to grade 2 (31). Although these symptoms disappeared within 1 month after completion of the radiation therapy, and no significant effect on late toxicity was observed (11, 32), this indicates that it is desirable to expand this planning software with tumor control probability and normal tissue complication probability models for a reliable clinical application. For this purpose, thermal enhancement of normal tissues should be investigated in more detail, discriminating between early and late effects. The working mechanism of hyperthermia is complex, and several factors (eg, sequence, heating technique, heterogeneity of the temperature distribution) might influence thermal enhancement. A better insight in thermal enhancement for different heating techniques and sequences will allow a general clinical application of the proposed model. Hyperthermia can be considered a promising alternative for straightforward dose escalation, and the presented model provides a tool to quantify this. With adequate knowledge of the effect of hyperthermia on the LQ parameters, the choice between a higher dose, including possible toxicity to organs at risk, or additional hyperthermia will become obvious. In the present study a realistic

Fig. 4. Example of a treatment planning with dose distributions for radiation therapy alone and equivalent dose distributions for combined radiation therapy and hyperthermia for 2 different cases with very different dose and temperature distributions. Adding hyperthermia yields an extension of the high-dose region. In this example, the 95% isodose level (ie, 95% of the dose in the isocenter) is shown.

744

Kok et al.

range of a values was considered. All 3 temperature dependency models for a(T ) (ie, alin, aexp, and alin2) showed a dose escalation when adding hyperthermia, and even when a conservative estimate of a(T ) was applied (ie, alin) a clinically relevant increase in average equivalent dose of approximately 10 Gy was still observed. However, more precise knowledge about the temperature dependency of the LQ parameters is required for a more accurate clinical application of the model. This should be obtained from detailed preclinical experiments at different temperature levels covering the complete hyperthermic range. Additionally, LQ parameters can be derived from data from clinical studies. Although the spread in clinical temperatures is more limited compared with preclinical experiments, values of a and b based on real clinical data are indispensable. Deriving LQ parameters under hyperthermic circumstances is beyond the scope of this article, and therefore a realistic estimate was made, but further research is presently conducted at our department. This model is a first step to quantify the clinical effect of hyperthermia in terms of equivalent dose. A more sophisticated model would also take into account separate biological mechanisms, such as repair, repopulation, and reoxygenation. Additionally, the dominant working mechanism of hyperthermia depends on the temperature level. Reoxygenation is probably more important at lower temperatures around 41 C, whereas inhibition of DNA damage repair is more significant at higher temperatures around 43 C. Furthermore, at higher temperatures in the hyperthermic range the vasculature might be compromised, yielding a decrease in blood flow and thereby also in oxygenation. Because clinical hyperthermia temperature distributions are heterogeneous, including these mechanisms would further improve the reliability of the model. In the present study only the effect of radiosensitization was taken into account. Furthermore, the LQ parameters at a specific temperature were considered homogeneous in this study. In reality these values are heterogeneous, which is usually modeled using a normal distribution with appropriate standard deviation. These improvements and incorporation of tumor control probability and normal tissue complication probability models are the subjects of further research. In this study the effect of locoregional hyperthermia after radiation therapy was modeled for prostate cancer, assuming 1 or 2 sessions per week. The developed model is general and can also be applied to other tumor sites, heating techniques, and/or sequences by including appropriate values of the LQ parameters. For example, simultaneous hyperthermia and radiation therapy shows a much stronger increase in a for the tumor, depending on the number of hyperthermia sessions (7, 28). When modeling other sequences thermal radiosensitization of normal tissue should also be taken into account (1). This does not require adaptations to the model, because equivalent dose distributions are calculated voxel by voxel. Hyperthermia treatment planning is currently a research tool, rather than clinical software. The plans will become more quantitatively reliable once uncertainties in dielectric and thermal properties of tissues are reduced (33, 34). Presently, literature values are used for dielectric and thermal parameters, which show some spread (17). Achieving patient-specific tissue properties and improving the reliability of treatment planning is the subject of current research at our department. Monte Carlo simulations can be used to take uncertainties in calculated temperatures into account. Furthermore, during optimization constraints were set to 45 C for all normal tissue, but for clinical application a lower constraint might be necessary for specific tissues (eg, 44 C for the

International Journal of Radiation Oncology  Biology  Physics bladder [35]). Nevertheless, the simulated temperature levels and heterogeneity of the temperature distributions in the tumor were realistic and comparable to clinically realized temperatures for pelvic tumors; therefore, this does not affect the results in this study. Thus, uncertainties in biological parameters and hyperthermia treatment planning hamper a strict quantitative interpretation of the results in this article. Nevertheless, the tumor temperature distributions were in the clinically relevant range, with very realistic inhomogeneities. Furthermore, realistic biological parameters were used, and enhancement factors of a chosen were somewhat conservative to prevent a strong overestimation of the effect of hyperthermia. Results demonstrated that a significant increase in equivalent dose can be realised by adding hyperthermia, which makes the study a fair “proof of principle.” Future work will be aimed at reducing uncertainties to enable a clinically meaningful application of the model presented in this article.

Conclusion This study showed that the therapeutic effect of radiosensitization by hyperthermia is equivalent to radiation-dose escalation with improved TCP. To quantify the effect of hyperthermia on radiation therapy dose distributions a theoretical framework and software tools were developed to calculate equivalent dose distributions. With reliable information about the temperature dependency of the LQ parameters, this model allows a detailed analysis of combined radiation therapy and hyperthermia treatments. This is very useful to analyze the clinical effect of multimodality treatments and guide further clinical studies on dose escalation.

References 1. Overgaard J. Simultaneous and sequential hyperthermia and radiation treatment of an experimental tumor and its surrounding normal tissue in vivo. Int J Radiat Oncol Biol Phys 1980;6:1507-1517. 2. Vernon CC, Hand JW, Field SB, et al. Radiotherapy with or without hyperthermia in the treatment of superficial localized breast cancer: Results from five randomized controlled trials. International Collaborative Hyperthermia Group. Int J Radiat Oncol Biol Phys 1996;35: 731-744. 3. Franckena M, Stalpers LJ, Koper PC, et al. Long-term improvement in treatment outcome after radiotherapy and hyperthermia in locoregionally advanced cervix cancer: An update of the Dutch Deep Hyperthermia Trial. Int J Radiat Oncol Biol Phys 2008;70:1176-1182. 4. Overgaard J, Gonza´lez Gonza´lez D, Hulshof MCCM, et al. Randomised trial of hyperthermia as adjuvant to radiotherapy for recurrent or metastatic malignant melanoma. European Society for Hyperthermic Oncology. Lancet 1995;345:540-543. 5. Huilgol NG, Gupta S, Sridhar CR. Hyperthermia with radiation in the treatment of locally advanced head and neck cancer: A report of randomized trial. J Cancer Res Ther 2010;6:492-496. 6. Engin K, Tupchong L, Moylan DJ, et al. Randomized trial of one versus two adjuvant hyperthermia treatments per week in patients with superficial tumours. Int J Hyperthermia 1993;9:327-340. 7. Overgaard J. Fractionated radiation and hyperthermia: Experimental and clinical studies. Cancer 1981;48:1116-1123. 8. Wust P, Beck R, Berger J, et al. Electric field distributions in a phasedarray applicator with 12 channels: Measurements and numerical simulations. Med Phys 2000;27:2565-2579. 9. Crezee J, Van Haaren PMA, Westendorp H, et al. Improving locoregional hyperthermia delivery using the 3-D controlled AMC-8 phased

Volume 88  Number 3  2014

10. 11.

12.

13.

14.

15. 16.

17. 18.

19. 20. 21. 22.

array hyperthermia system: A preclinical study. Int J Hyperthermia 2009;25:581-592. Sapareto SA, Dewey WC. Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys 1984;10:787-800. Hurwitz MD, Hansen JL, Prokopios-Davos S, et al. Hyperthermia combined with radiation for the treatment of locally advanced prostate cancer: Long-term results from Dana-Farber Cancer Institute study 94153. Cancer 2011;117:510-516. Anscher MS, Samulski TV, Dodge R, et al. Combined external beam irradiation and external regional hyperthermia for locally advanced adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys 1997;37: 1059-1065. Tilly W, Gellermann J, Graf R, et al. Regional hyperthermia in conjunction with definitive radiotherapy against recurrent or locally advanced prostate cancer T3 pN0 M0. Strahlenther Onkol 2005;181: 35-41. Maluta S, Dall’oglio S, Romano M, et al. Conformal radiotherapy plus local hyperthermia in patients affected by locally advanced high risk prostate cancer: Preliminary results of a prospective phase II study. Int J Hyperthermia 2007;23:451-456. De Greef M, Crezee J, van Eijk JC, et al. Accelerated ray tracing for radiotherapy dose calculations on a GPU. Med Phys 2009;36:4095-4102. Hornsleth SN, Mella O, Dahl O. A new segmentation algorithm for finite difference based treatment planning systems. In: Franconi C, Arcangeli G, Cavaliere R, editors. Hyperthermic OncologyVol. 2. Rome: Tor Vergata; 1996. p. 521-523. Gabriel C, Gabriel S, Corthout E. The dielectric properties of biological tissues: I. Literature survey. Phys Med Biol 1996;41:2231-2249. ESHO Taskgroup Committee. Treatment Planning and Modelling in Hyperthermia, a Task Group Report of the European Society for Hyperthermic Oncology. Rome: Tor Vergata; 1992. Das SK, Clegg ST, Samulski TV. Computational techniques for fast hyperthermia temperature optimization. Med Phys 1999;26:319-328. Pennes HH. Analysis of tissue and arterial blood temperatures in the resting human forearm. J Appl Physiol 1948;1948(1):93-122. Fowler JF. The linear-quadratic formula and progress in fractionated radiotherapy. Br J Radiol 1989;62:679-694. O’Rourke SF, McAneney H, Hillen T. Linear quadratic and tumour control probability modelling in external beam radiotherapy. J Math Biol 2009;58:799-817.

Dose escalation by adding hyperthermia

745

23. Warkentin B, Stavrev P, Stavreva N, et al. A TCP-NTCP estimation module using DVHs and known radiobiological models and parameter sets. J Appl Clin Med Phys 2004;5:50-63. 24. Pajonk F, Van Ophoven A, McBride WH. Hyperthermia-induced proteasome inhibition and loss of androgen receptor expression in human prostate cancer cells. Cancer Res 2005;65:4836-4843. 25. Franken NA, Oei AL, Kok HP, et al. Cell survival and radiosensitisation: Modulation of the linear and quadratic parameters of the LQ model (review). Int J Oncol 2013;42:1501-1515. 26. Oliveira SM, Teixeira NJ, Fernandes L. What do we know about the alpha/beta for prostate cancer? Med Phys 2012;39:3189-3201. 27. Fowler J, Chappell R, Ritter M. Is alpha/beta for prostate tumors really low? Int J Radiat Oncol Biol Phys 2001;50:1021-1031. 28. Myerson RJ, Roti Roti JL, Moros EG, et al. Modelling heat-induced radiosensitization: clinical implications. Int J Hyperthermia 2004;20: 201-212. 29. Hand JW, Lagendijk JJ, Bach Andersen J, et al. Quality assurance guidelines for ESHO protocols. Int J Hyperthermia 1989;5:421-428. 30. Levegrun S, Jackson A, Zelefsky MJ, et al. Fitting tumor control probability models to biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer: Pitfalls in deducing radiobiologic parameters for tumors from clinical data. Int J Radiat Oncol Biol Phys 2001;51:1064-1080. 31. Hurwitz MD, Kaplan ID, Hansen JL, et al. Association of rectal toxicity with thermal dose parameters in treatment of locally advanced prostate cancer with radiation and hyperthermia. Int J Radiat Oncol Biol Phys 2002;53:913-918. 32. Hurwitz MD, Kaplan ID, Hansen JL, et al. Hyperthermia combined with radiation in treatment of locally advanced prostate cancer is associated with a favourable toxicity profile. Int J Hyperthermia 2005; 21:649-656. 33. De Greef M, Kok HP, Correia D, et al. Optimization in hyperthermia treatment planning: The impact of tissue perfusion uncertainty. Med Phys 2010;37:4540-4550. 34. Canters RA, Paulides MM, Franckena M, et al. Benefit of replacing the Sigma-60 by the Sigma-Eye applicator. A Monte Carlo-based uncertainty analysis. Strahlenther Onkol 2013;189:74-80. 35. Haveman J, Smals OA, Rodermond HM. Effects of hyperthermia on the rat bladder: A pre-clinical study on thermometry and functional damage after treatment. Int J Hyperthermia 2003;19:45-57.

Quantifying the combined effect of radiation therapy and hyperthermia in terms of equivalent dose distributions.

To develop a method to quantify the therapeutic effect of radiosensitization by hyperthermia; to this end, a numerical method was proposed to convert ...
778KB Sizes 0 Downloads 0 Views