- Email: [email protected]
lipiodol-BNCT) for treatment of multiple liver tumors
Int. J. Radiation Oncology Biol. Phys., Vol. 58, No. 3, pp. 892– 896, 2004 Copyright © 2004 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/04/$–see front matter
doi:10.1016/j.ijrobp.2003.09.084
PHYSICS CONTRIBUTION
DOSIMETRIC STUDY OF BORON NEUTRON CAPTURE THERAPY WITH BOROCAPTATE SODIUM (BSH)/LIPIODOL EMULSION (BSH/LIPIODOLBNCT) FOR TREATMENT OF MULTIPLE LIVER TUMORS MINORU SUZUKI, M.D., PH.D.,* YOSHINORI SAKURAI, PH.D.,† SHINICHIRO MASUNAGA, M.D., PH.D.,* YUKO KINASHI, M.D., PH.D.,* KENJI NAGATA, M.D., PH.D.,* AND KOJI ONO, M.D., PH.D.* *Radiation Oncology Research Laboratory and †Division of Radiation Life Science, Research Reactor Institute, Kyoto University, Osaka, Japan Purpose: We performed a computational study to investigate the feasibility of borocaptate sodium (BSH)/ lipiodol-boron neutron capture therapy (BSH/lipiodol-BNCT) for multiple liver tumors using Simulation Environment for Radiotherapy Applications (SERA), a currently available BNCT treatment planning system. Methods and Materials: Three treatment plans for BSH/lipiodol-BNCT using two or three epithermal neutron beams in one fraction were generated for 4 patients with multiple liver tumors using the SERA system. The 10B concentrations in the tumor and the liver assumed in the study were 197.3 and 15.3 ppm, respectively; and were obtained from experimental studies in animals. The therapeutic gain factors for the liver tumors, defined as the minimum dose to the tumor/maximum dose to the liver, and the inhomogeneity index of the thermal neutron fluence for the whole of the liver, defined as the maximum neutron fluence ⴚ minimum neutron fluence/mean neutron fluence, were evaluated in each plan. Results: Three epithermal neutron beams incident on the anterior, posterior, and right side of the patient can deliver the most homogeneous distribution of thermal neutron fluence to the whole of the liver and provide the greatest therapeutic gain factors for tumors in the right lobe and approximately equal therapeutic gain factors for tumors in the left lobe, compared with the two opposed (anterior-posterior) and two orthogonal (anteriorright) beams. Conclusions: From a dosimetric viewpoint, the BSH/lipiodol-BNCT treatment plan using three epithermal neutron beams is the most suitable for the treatment of multiple liver tumors. © 2004 Elsevier Inc. BNCT, Multiple liver tumors, BSH.
INTRODUCTION
conventional photon therapy, and success rates are low. BNCT using high-linear energy transfer (LET) particles has been applied in an attempt to improve treatment outcomes. We have investigated the possibility of the application of BNCT to malignant liver tumors (7). Using a rat liver tumor model, we have successfully selectively accumulated high 10 B concentrations in experimental liver tumors by the intra-arterial administration of borocaptate sodium (BSH)/ lipiodol emulsion. The 10B concentration in liver tumors and the tumor/liver (T/L) 10B concentration ratio at 6 h after the administration of BSH/lipiodol emulsion were 197.3 ppm and 14.9, respectively (8). These results encourage us to extend the application of BNCT using an intra-arterial administration of BSH/lipiodol emulsion (BSH/lipiodolBNCT). To investigate the possibility of the use of BSH/lipiodolBNCT in a clinical environment, we constructed treatment plans for 4 patients with multiple liver tumors using Simu-
Boron neutron capture therapy (BNCT) is based on a nuclear reaction; nonradioactive isotope 10B atoms that have absorbed low energy (⬍0.5 eV) neutrons disintegrate into alpha (4He) particles and recoiled lithium nuclei (7Li). These particles deposit high energies along their very short paths (less than 10 m) (1). Malignant cells with 10B are thus destroyed after thermal neutron irradiation. If a sufficient number of 10B atoms accumulate in the tumor cells and the gradient of the amount of 10B atoms between the tumor and the surrounding normal tissues is large, then selective boron neutron capture irradiation will be successfully delivered to the tumor. Malignant gliomas and malignant melanomas have been treated with BNCT in Japan, Europe, and the United States since the 1970s (2– 6). Malignant gliomas and malignant melanomas are both very difficult to treat effectively using Reprint requests to: Minoru Suzuki, M.D., Ph.D., Radiation Oncology Research Laboratory, Research Reactor Institute, Kyoto University, Noda, Kumatori-cho, Sennan-gun, Osaka 590-0494, Japan. Tel: (⫹81) 724-51-2407; Fax: (⫹81) 724-51-2627;
E-mail: [email protected] Received Apr 25, 2003, and in revised form Sep 17, 2003. Accepted for publication Sep 22, 2003. 892
BNCT treatment planning for liver tumors
Table 1. RBE and CBE factors used for conversion of physical dose (Gy) to photon-equivalent dose (Gy-Eq) BNCT dose component B (n, ␣) 7Li N (n, p) 14C Fast neutron 10 14
Tumor
Liver
2.29 (CBE)* 3.20† 3.20†
0.94 (CBE)* 1.67* 2.40‡
Abbreviations: BNCT ⫽ boron neutron capture therapy; CBE ⫽ compound biologic effectiveness; RBE ⫽ relative biologic effectiveness. * Data from Suzuki et al. (7). † Data from Coderre et al. (13). ‡ Data from Ono et al (14).
lation Environment for Radiotherapy Applications (SERA), a currently available BNCT treatment planning system. The SERA system has been developed by the Idaho National Engineering and Environmental Laboratory in collaboration with Montana State University (9 –11). The SERA system was used in clinical trials for glioblastoma at Brookhaven National Laboratory (12). The present study aimed to investigate the feasibility of BSH/lipiodol-BNCT for multiple liver tumors from a dosimetric viewpoint using SERA to perform the necessary calculations. METHODS AND MATERIALS Overview for BNCT treatment planning with SERA For the present study, the computed tomographic (CT) images of 4 patients with multiple liver tumors were used to define the geometric models in the SERA system. Two patients had multiple liver metastases from colon cancer, and the other 2 patients had multiple primary hepatocellular carcinomas. The SERA system requires the entry of several userdefined parameters. These parameters include the 10B concentrations in the liver tumors and the normal liver, the nuclear composition of the tissues, the relative biologic effectiveness (RBE) of each component of the beam, and the compound biologic effectiveness (CBE) factors of the boron compounds. It is also necessary to specify the incident reactor source. The CBE factors were used as an alternative RBE in evaluating the biologically absorbed dose by BNCT because different or the same boron compounds might yield variable effects on different tissues due to variations in their microdistribution of the boron compounds and the morphologic character of the target cells (1). The 10B concentrations in the tumor and the liver assumed in the present study were 197.3 and 15.3 ppm, respectively; the levels were obtained from experimental studies in animals (8). The RBE or CBE values, which were reported previously, including ours, were used in the present study (7, 13, 14). Table 1 summarizes the parameters used in the present study. The treatment planning included the following steps: first, the CT images of each patient with multiple liver tumors were input to the SERA system. The volumes of the
● M. SUZUKI et al.
893
tumors and normal liver were delineated on the CT images using the visualization tools available with the SERA system. This process required 1–2 h for each patient. By this process, a three-dimensional geometric description of the upper abdomen was constructed. The incident neutron beam direction and the collimator size were then defined and oriented relative to the anatomy. After boron concentrations in the tumor or the liver and the values for the RBE or CBE were entered, the SERA system was run for the dose calculation. The treatment planning cord was run on Sun Blade 1000, Model 2750 (Sun Microsystems, Inc., Santa Clara, CA). The SERA system simulates about 4,000,000 source neutrons and 4,000,000 source photons, and required a processing time of 15–30 min for each field. BNCT treatment planning with multi-beam Three treatment plans using two or three epithermal neutron beams in one fraction were generated for each patient. The first treatment plan was constructed with two opposed anterior-posterior beams (AP-beams). The anterior field collimator size was determined to encompass the whole of the liver with 2-cm margins in all directions. The posterior field collimator size was determined to encompass the right lobe with a 2-cm margin and to not include the spinal cord. The second plan was constructed using two orthogonal anteriorright beams (AR-beams). The anterior beam collimator size was determined in the same manner as in the first plan. The right beam collimator size was determined to encompass the whole of the liver with a 2-cm margin in all directions. The third plan was constructed using three beams, that is, anterior-right-posterior beams (ARP-beams). Each collimation size was determined in the same manner as in the first and second plans. In the present study, each beam was equally weighted. Comparison of the treatment plans The SERA system can provide dose–volume histogram (DVH) data for the volume of each of the tumors or for the liver as a whole. For comparison, all treatment plans were normalized to deliver mean doses of 5 Gray-equivalent (Gy-Eq) to the whole of the liver. The maximum, mean, and minimum doses to the tumors and the liver were assessed for each plan. The success of treatment of multiple liver tumors with radiotherapy depends on the dose gradient between the tumors and the normal liver. In the present study, the therapeutic gain factor was defined as the ratio of the minimum dose to the tumor (Dmin for the tumor) to the maximum dose to the liver (Dmax for the liver). To achieve large therapeutic gain factors for all liver tumors, a homogeneous distribution of the thermal neutron fluence across the whole of the liver is preferable. To evaluate the distribution of the thermal neutron fluence, the inhomogeneity index of thermal neutron fluence for the whole of the liver was defined as the ratio of the difference between the maximum neutron fluence (⌽ max) and the minimum neutron fluence (⌽ min) relative to the mean neutron fluence (⌽ mean).
894
I. J. Radiation Oncology
● Biology ● Physics
Table 2. Summary of dose–volume histogram data showing averages (with range) for the tumors in the right lobe Beam directions
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
AP beam
Maximum dose (Gy-Eq)
36.5 45.6 61.5 (20.6–65.5) (27.7–86.0) (43.8–107.2) AR beam 46.1 60.5 76.1 (9.3–65.5) (21.5–88.8) (46.8–112.6) ARP beam 50.3 65.1 86.1 (17.6–69.0) (27.3–80.6) (42.5–108.0)
Therapeutic gain factor 3.8 (2.5–8.0) 4.5 (1.0–7.2) 6.1 (2.1–8.3)
Volume 58, Number 3, 2004
Table 3. Summary of dose–volume histogram data showing averages (with range) for the tumors in the left lobe Beam directions
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
Maximum dose (Gy-Eq)
Therapeutic gain factor
AP beam
36.5 53.3 81.1 3.8 (15.2–59.1) (35.1–82.8) (47.1–104.7) (1.5–6.7) AR beam 30.6 45.1 67.9 3.5 (8.8–47.3) (27.8–66.1) (38.7–83.6) (1.0–5.9) ARP beam 27.4 39.9 58.3 3.5 (11.7–40.8) (26.1–56.6) (34.6–71.5) (1.41–5.2)
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior.
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior.
RESULTS
greater than those by AP- or AR-beams (6.1 ⫾ 2.1 vs. 3.8 ⫾ 1.8, p ⫽ 0.0079; and 6.1 ⫾ 2.1 vs. 4.5 ⫾ 2.1, p ⫽ 0.0023). The average of mean doses delivered to the tumors in the left lobe by the AP-beam was 53.3 ⫾ 23.4 Gy-Eq, which was higher than 45.1 ⫾ 19.4 Gy-Eq by the AR-beams or 39.9 ⫾ 15.5 Gy-Eq by the ARP-beams, but not significantly so (p ⫽ 0.0678 and p ⫽ 0.0563, respectively). The average of therapeutic gain factors for the tumors in the left lobe were 3.8 ⫾ 2.4 (AP), 3.5 ⫾ 2.2 (AR), and 3.5 ⫾ 1.8 (ARP), respectively.
DVH analysis for the liver tumors The position in which the tumors were located was expected to affect the dose distribution in the tumors. Thus, the DVH analyses for the tumors were summarized separately for tumors in the left and right lobes. The total number of tumors in the right and left lobes in all cases were 9 and 4, respectively. Tables 2 and 3 show the DVH analysis for the tumors including the maximum, mean, and minimum dose (Gy-Eq) and the therapeutic gain factors for the tumors in the right and left lobes. For all of the tumors in the right or left lobe, all of the treatment plans accomplished therapeutic gain factors of greater than 1. Under the condition that the mean dose to the whole of the liver was normalized to 5 Gy-Eq, the average of the mean doses delivered to the tumors in the right lobe by ARP-beams was significantly higher than that by AP-beams (65.1 ⫾ 19.5 vs. 45.6 ⫾ 19.1 Gy-Eq, p ⫽ 0.0255). Steep decrease in thermal neutron fluence within the body yielded wide distribution of the mean doses among the tumors. The average of therapeutic gain factors for the tumors in the right lobe by ARP-beams was significantly
DVH analysis for the normal liver Table 4 summarizes the DVH analysis for the liver including the maximum and minimum dose (Gy-Eq) and the inhomogeneity index by each plan. Under the condition that the mean dose to the whole liver was normalized to 5 Gy-Eq, the average maximum and minimum dose delivered to the whole liver in each plan ranged from 8.0 to 10.1 Gy-Eq and from 1.1 to 1.2 Gy-Eq, respectively. The average inhomogeneity index of the thermal neutron fluence by ARP-beams was 1.74, which was the lowest of the three treatment plans. Figure 1 compares the distribution of the
Fig. 1. Thermal neutron fluence distribution delivered by an AP-beam (a), AR-beam (b), and ARP-beam (c). The contour of the liver is outlined as a solid white line. The distributions of the thermal neutron fluence are shown with color wash display as follows; ⬎100% Blue, 90 –100% Green, 80 –90% Red, 70 – 80% White, 60 –70% Yellow, 50 – 60% Magenta, 40 –50% Cyan, 30 – 40% Blue, 20 –30% Green, 10 –20% Red, and 0 –10% white.
BNCT treatment planning for liver tumors
Table 4. Summary of dose–volume histogram data showing averages (with range) for the liver Beam directions AP beam AR beam ARP beam
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
Maximum dose (Gy-Eq)
Inhomogeneity index
1.2 (0.8–1.3) 1.2 (0.8–1.3) 1.1 (0.8–1.3)
5.0*
10.1 (8.8–11.3) 8.8 (8.1–9.1) 8.0 (7.6–8.4)
2.39 (1.82–2.73) 1.94 (1.69–2.05) 1.74 (1.61–1.87)
5.0* 5.0*
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior. * The mean dose to the liver normalized to 5.0 Gy-Eq.
thermal neutron fluence delivered by the AP-, AR-, and ARP-beams. The ARP-beam can be seen to provide a more homogeneous distribution of thermal neutrons compared with the AP- or AR-beams. DISCUSSION The present study using the SERA system, a currently available BNCT treatment planning system, shows that BSH/lipiodol-BNCT has potential as a new treatment option for multiple liver tumors. In contrast with multimodal treatment options for solitary liver tumors, that is, surgical resections, thermal ablations, percutaneous ethanol injections, cryoablations, heavy charged particle irradiation, and stereotactic irradiation with photons (15–20), the treatment options for multiple liver tumors have been limited. Although the intra-arterial injection of chemotherapeutic agents with or without embolization is effective for multiple liver tumors, recurrent liver tumors, which are refractory to chemotherapeutic agents, develop after the repetition of this treatment (21, 22). Our novel method, BSH/lipiodol-BNCT, may offer a more promising option. Two or three liver tumors visible on CT or magnetic resonance imaging (MRI) may be treated successfully using three-dimensional conformal radiotherapy or intensity-modulated radiotherapy with large dose-gradients between the tumors and the surrounding normal liver (23, 24). However, any type of external radiotherapy cannot treat occult metastatic liver tumors undetectable by CT or MRI. Because lipiodol has been used clinically as a contrast agent for detecting occult liver metastases or hepatocellular carcinomas, BSH accompanied by lipiodol is expected to accumulate in the occult micrometastatic tumors as well as obvious tumors (25, 26). Thus, BSH/lipiodol-BNCT has potential in the treatment of both visible and occult micrometastatic tumors. For multiple liver tumors to be successfully treated with radiation therapy including BNCT, the therapeutic gain factors defined as Dmin for the tumor/Dmax for the liver should be greater than 1 for all liver tumors. One of the most difficult problems in applying BNCT to deep-seated tumors such as
● M. SUZUKI et al.
895
liver tumors is the poor penetration of thermal neutrons. To improve the depth distribution of thermal neutron fluence, an epithermal neutron beam is preferable compared with a thermal neutron beam. At Kyoto University Research Reactor, the Heavy Water Neutron Irradiation Facility was reconstructed from November 1995 to March 1996 to improve neutron capture therapy (27). At the updated facility, almost pure thermal to epithermal neutron beams are available. In the present study, we applied the beam data on the epithermal neutron beam of the Kyoto University Research Reactor to the dose calculation using the SERA system. To obtain large therapeutic gain factors for all obvious or occult liver tumors, a homogeneous distribution of thermal neutron fluence and sufficient neutron fluence to the deep portion in the liver is a prerequisite. Pinelli et al. (28) reported a novel method for delivering sufficient and homogenous thermal neutron fluence to the whole of the liver. In that work, a patient suffering multiple liver metastases from colon cancer was treated with BNCT. The liver was surgically extracted and the isolated liver was then moved to a reactor for irradiation with thermal neutrons. The irradiation facility was designed to irradiate the whole of the liver homogeneously in all directions. After irradiation, the liver was autotransplanted back to the patient. The patient was in good general condition and the multiple liver metastases had disappeared 7 months after BNCT. Their success encourages future work in the field of BNCT. In contrast to the treatment protocol by Pinelli et al., the procedure for treating multiple liver tumors with BSH/lipiodol-BNCT is almost noninvasive except for intra-arterial administration of BSH/lipiodol emulsion. In our protocol, after receiving intra-arterial administration of BSH/lipiodol emulsion, the patient is irradiated with epithermal neutron beams collimated to the liver from two or three directions in one fraction. Provided that the 10B concentration (200 ppm) is high with an acceptable T/L 10B concentration ratio (⬎13), the BSH/lipiodol-BNCT protocol using two or three epithermal neutron beams can offer therapeutic gain factors of greater than 1 for all of the liver tumors analyzed in the present study. Compared with AP- or AR- beams, the ARP-beams can deliver the most homogeneous distribution of thermal neutron fluence to the whole of the liver, and provide the greatest therapeutic gain factors for tumors in the right lobe, along with approximately equal therapeutic gain factors for tumors in the left lobe. CONCLUSIONS BSH/lipiodol-BNCT using two or three epithermal neutron beams has potential in the treatment of multiple liver tumors with curative intent, if high 10B concentrations in the liver tumor (200 ppm) and a high T/L 10B concentration ratio (⬎13) are achieved. From a dosimetric viewpoint, BSH/lipiodol-BNCT using three epithermal neutron beams (ARP-beams) is the most suitable for the treatment of multiple liver tumors.
896
I. J. Radiation Oncology
● Biology ● Physics
Volume 58, Number 3, 2004
REFERENCES 1. Coderre JA, Morris GM. The radiation biology of boron neutron capture therapy. Radiat Res 1999;151:1–18. 2. Hatanaka H, Nakagawa Y. Clinical results of long-surviving brain tumor patients who underwent boron neutron capture therapy. Int J Radiat Oncol Biol Phys 1994;28:1061–1066. 3. Nakagawa Y, Hatanaka H. Boron neutron capture therapy, clinical brain tumor studies. J Neurooncol 1997;33:105–115. 4. Takagaki M, Oda Y, Miyatake S, et al. Boron neutron capture therapy: Preliminary study of BNCT with sodium borocaptate (Na2B12H11SH) on glioblastoma. J Neurooncol 1997;35:177– 185. 5. Sauerwein W, Moss R, Rassow J, et al. Organization and management of the first clinical trial of BNCT in Europe (EORTC Protocol 11961). EORTC BNCT study group. Strahlenther Onkol 1999;175:108–111. 6. Kageji T, Nagahiro S, Kitamura K, et al. Optimal timing of neutron irradiation for boron neutron capture therapy after intravenous infusion of sodium borocaptate in patients with glioblastoma. Int J Radiat Oncol Biol Phys 2001;51:120–130. 7. Suzuki M, Masunaga S, Kinashi Y, et al. The effects of boron neutron capture therapy on liver tumors and normal hepatocytes in mice. Jpn J Cancer Res 2000;91:1058–1064. 8. Suzuki M, Masunaga S, Kinashi Y, et al. Intra-arterial administration of sodium borocaptate (BSH)/lipiodol emulsion delivers B-10 to liver tumors highly selectively for boron neutron capture therapy: Experimental studies in the rat liver model. Int J Radiat Oncol Biol Phys; in press. 9. Nigg DW, Wemple CA, Wessol DE, et al. SERA—An advanced treatment planning system for neutron capture therapy and BNCT. Trans Am Nucl Soc 1999;80:66–68. 10. Wojnecki C, Green S. A preliminary comparative study of two treatment planning systems developed for boron neutron capture therapy: MacNCTPlan and SERA. Med Phys 2002;29: 1710–1715. 11. Nichols TL, Kabalka GW, Miller LF, et al. Improved treatment planning for boron neutron capture therapy for glioblastoma multiforme using fluorine-18 labeled boronophenylalanine and positron emission tomography. Med Phys 2002;29: 2351–2358. 12. Chadha M, Capala J, Coderre JA, et al. Boron neutron capture therapy (BNCT) for glioblastoma multiforme (GBM) using the epithermal neutron beam at the Brookhaven National Laboratory. Int J Radiat Oncol Biol Phys 1998;40:829–834. 13. Coderre JA, Makar MS, Micca PL, et al. Derivations of relative biological effectiveness for the high-LET radiations produced during boron neutron capture irradiations of the 9L rat gliosarcoma in vitro and in vivo. Int J Radiat Oncol Biol Phy 1993;27:1121–1129. 14. Ono K, Nagata Y, Akuta K, et al. Frequency of micronuclei in hepatocytes following X and fast-neutron irradiations: An analysis by a linear-quadratic model. Radiat Res 1990;123: 345–347.
15. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: Analysis of 1001 consecutive cases. Ann Surg 1999; 230:309–321. 16. Lau WY, Thomas WT, Leung MD, et al. Percutaneous local ablative therapy for hepatocellular carcinoma. Ann Surg 2003; 237:171–179. 17. Teratani T, Ishikawa T, Shiratori Y, et al. Hepatocellular carcinoma in elderly patients: Beneficial therapeutic efficacy using percutaneous ethanol injection therapy. Cancer 2002; 95:816–823. 18. Mala T, Edwin B, Tillung T, et al. Percutaneous cryoablation of colorectal liver metastases: Potentiated by two consecutive freeze-thaw cycles. Cryobiology 2003;46:99–102. 19. Ohara K, Okumura T, Thujii H, et al. Clearance of parenchymal tumors following radiotherapy: Analysis of hepatocellular carcinomas treated by proton beams. Radiother Oncol 1996; 41:233–236. 20. Sato M, Uematsu M, Shioda A, et al. Feasibility of frameless stereotactic high-dose radiation therapy for primary or metastatic liver cancer. J Radiosurg 1998;1:233–238. 21. Camma´ C, Schepis F, Orlando A, et al. Transarterial chemoembolization for unresectable hepatocellular carcinoma: Meta-analysis of randomized controlled trials. Radiology 2002;224:47–54. 22. Maeda S, Fujiyama S, Tanaka M, et al. Survival and local recurrence rates of hepatocellular carcinoma patients treated by transarterial chemolipiodolization with and without embolization. Hepatol Res 2002;23:202–210. 23. Park HC, Seong J, Han KH, et al. Dose-response relationship in local radiotherapy for hepatocellular carcinoma. Int J Radiat Oncol Biol Phys 2002;54:150–155. 24. Hugo GD, Agazaryan N, Solberg TD, et al. An evaluation of gating window size, delivery method, and composite field dosimetry of respiratory-gated IMRT. Med Phys 2002;29: 2517–2525. 25. Yumoto Y, Jinno K, Tokuyama K, et al. Hepatocellular carcinoma detected by iodized oil. Radiology 1985;154:19–24. 26. Nakakuma K, Tashiro S, Hiraoka T, et al. Hepatocellular carcinoma and metastatic cancer detected by iodized oil. Radiology 1985;154:15–17. 27. Sakurai Y, Kobayashi T. The medical-irradiation characteristics for neutron capture therapy at the Heavy Water Neutron Irradiation Facility of Kyoto University Research Reactor. Med Phys 2002;29:2328–2337. 28. Pinelli T, Zonta A, Altieri S, et al. TAOeMINA: From the first idea to the application to the human liver. In: Sauerwein W, Moss R, Wittig A, editors. Research and development in neutron capture therapy. Bologna: Monduzzui Ediotore, 2002: p. 1065–1072.
doi:10.1016/j.ijrobp.2003.09.084
PHYSICS CONTRIBUTION
DOSIMETRIC STUDY OF BORON NEUTRON CAPTURE THERAPY WITH BOROCAPTATE SODIUM (BSH)/LIPIODOL EMULSION (BSH/LIPIODOLBNCT) FOR TREATMENT OF MULTIPLE LIVER TUMORS MINORU SUZUKI, M.D., PH.D.,* YOSHINORI SAKURAI, PH.D.,† SHINICHIRO MASUNAGA, M.D., PH.D.,* YUKO KINASHI, M.D., PH.D.,* KENJI NAGATA, M.D., PH.D.,* AND KOJI ONO, M.D., PH.D.* *Radiation Oncology Research Laboratory and †Division of Radiation Life Science, Research Reactor Institute, Kyoto University, Osaka, Japan Purpose: We performed a computational study to investigate the feasibility of borocaptate sodium (BSH)/ lipiodol-boron neutron capture therapy (BSH/lipiodol-BNCT) for multiple liver tumors using Simulation Environment for Radiotherapy Applications (SERA), a currently available BNCT treatment planning system. Methods and Materials: Three treatment plans for BSH/lipiodol-BNCT using two or three epithermal neutron beams in one fraction were generated for 4 patients with multiple liver tumors using the SERA system. The 10B concentrations in the tumor and the liver assumed in the study were 197.3 and 15.3 ppm, respectively; and were obtained from experimental studies in animals. The therapeutic gain factors for the liver tumors, defined as the minimum dose to the tumor/maximum dose to the liver, and the inhomogeneity index of the thermal neutron fluence for the whole of the liver, defined as the maximum neutron fluence ⴚ minimum neutron fluence/mean neutron fluence, were evaluated in each plan. Results: Three epithermal neutron beams incident on the anterior, posterior, and right side of the patient can deliver the most homogeneous distribution of thermal neutron fluence to the whole of the liver and provide the greatest therapeutic gain factors for tumors in the right lobe and approximately equal therapeutic gain factors for tumors in the left lobe, compared with the two opposed (anterior-posterior) and two orthogonal (anteriorright) beams. Conclusions: From a dosimetric viewpoint, the BSH/lipiodol-BNCT treatment plan using three epithermal neutron beams is the most suitable for the treatment of multiple liver tumors. © 2004 Elsevier Inc. BNCT, Multiple liver tumors, BSH.
INTRODUCTION
conventional photon therapy, and success rates are low. BNCT using high-linear energy transfer (LET) particles has been applied in an attempt to improve treatment outcomes. We have investigated the possibility of the application of BNCT to malignant liver tumors (7). Using a rat liver tumor model, we have successfully selectively accumulated high 10 B concentrations in experimental liver tumors by the intra-arterial administration of borocaptate sodium (BSH)/ lipiodol emulsion. The 10B concentration in liver tumors and the tumor/liver (T/L) 10B concentration ratio at 6 h after the administration of BSH/lipiodol emulsion were 197.3 ppm and 14.9, respectively (8). These results encourage us to extend the application of BNCT using an intra-arterial administration of BSH/lipiodol emulsion (BSH/lipiodolBNCT). To investigate the possibility of the use of BSH/lipiodolBNCT in a clinical environment, we constructed treatment plans for 4 patients with multiple liver tumors using Simu-
Boron neutron capture therapy (BNCT) is based on a nuclear reaction; nonradioactive isotope 10B atoms that have absorbed low energy (⬍0.5 eV) neutrons disintegrate into alpha (4He) particles and recoiled lithium nuclei (7Li). These particles deposit high energies along their very short paths (less than 10 m) (1). Malignant cells with 10B are thus destroyed after thermal neutron irradiation. If a sufficient number of 10B atoms accumulate in the tumor cells and the gradient of the amount of 10B atoms between the tumor and the surrounding normal tissues is large, then selective boron neutron capture irradiation will be successfully delivered to the tumor. Malignant gliomas and malignant melanomas have been treated with BNCT in Japan, Europe, and the United States since the 1970s (2– 6). Malignant gliomas and malignant melanomas are both very difficult to treat effectively using Reprint requests to: Minoru Suzuki, M.D., Ph.D., Radiation Oncology Research Laboratory, Research Reactor Institute, Kyoto University, Noda, Kumatori-cho, Sennan-gun, Osaka 590-0494, Japan. Tel: (⫹81) 724-51-2407; Fax: (⫹81) 724-51-2627;
E-mail: [email protected] Received Apr 25, 2003, and in revised form Sep 17, 2003. Accepted for publication Sep 22, 2003. 892
BNCT treatment planning for liver tumors
Table 1. RBE and CBE factors used for conversion of physical dose (Gy) to photon-equivalent dose (Gy-Eq) BNCT dose component B (n, ␣) 7Li N (n, p) 14C Fast neutron 10 14
Tumor
Liver
2.29 (CBE)* 3.20† 3.20†
0.94 (CBE)* 1.67* 2.40‡
Abbreviations: BNCT ⫽ boron neutron capture therapy; CBE ⫽ compound biologic effectiveness; RBE ⫽ relative biologic effectiveness. * Data from Suzuki et al. (7). † Data from Coderre et al. (13). ‡ Data from Ono et al (14).
lation Environment for Radiotherapy Applications (SERA), a currently available BNCT treatment planning system. The SERA system has been developed by the Idaho National Engineering and Environmental Laboratory in collaboration with Montana State University (9 –11). The SERA system was used in clinical trials for glioblastoma at Brookhaven National Laboratory (12). The present study aimed to investigate the feasibility of BSH/lipiodol-BNCT for multiple liver tumors from a dosimetric viewpoint using SERA to perform the necessary calculations. METHODS AND MATERIALS Overview for BNCT treatment planning with SERA For the present study, the computed tomographic (CT) images of 4 patients with multiple liver tumors were used to define the geometric models in the SERA system. Two patients had multiple liver metastases from colon cancer, and the other 2 patients had multiple primary hepatocellular carcinomas. The SERA system requires the entry of several userdefined parameters. These parameters include the 10B concentrations in the liver tumors and the normal liver, the nuclear composition of the tissues, the relative biologic effectiveness (RBE) of each component of the beam, and the compound biologic effectiveness (CBE) factors of the boron compounds. It is also necessary to specify the incident reactor source. The CBE factors were used as an alternative RBE in evaluating the biologically absorbed dose by BNCT because different or the same boron compounds might yield variable effects on different tissues due to variations in their microdistribution of the boron compounds and the morphologic character of the target cells (1). The 10B concentrations in the tumor and the liver assumed in the present study were 197.3 and 15.3 ppm, respectively; the levels were obtained from experimental studies in animals (8). The RBE or CBE values, which were reported previously, including ours, were used in the present study (7, 13, 14). Table 1 summarizes the parameters used in the present study. The treatment planning included the following steps: first, the CT images of each patient with multiple liver tumors were input to the SERA system. The volumes of the
● M. SUZUKI et al.
893
tumors and normal liver were delineated on the CT images using the visualization tools available with the SERA system. This process required 1–2 h for each patient. By this process, a three-dimensional geometric description of the upper abdomen was constructed. The incident neutron beam direction and the collimator size were then defined and oriented relative to the anatomy. After boron concentrations in the tumor or the liver and the values for the RBE or CBE were entered, the SERA system was run for the dose calculation. The treatment planning cord was run on Sun Blade 1000, Model 2750 (Sun Microsystems, Inc., Santa Clara, CA). The SERA system simulates about 4,000,000 source neutrons and 4,000,000 source photons, and required a processing time of 15–30 min for each field. BNCT treatment planning with multi-beam Three treatment plans using two or three epithermal neutron beams in one fraction were generated for each patient. The first treatment plan was constructed with two opposed anterior-posterior beams (AP-beams). The anterior field collimator size was determined to encompass the whole of the liver with 2-cm margins in all directions. The posterior field collimator size was determined to encompass the right lobe with a 2-cm margin and to not include the spinal cord. The second plan was constructed using two orthogonal anteriorright beams (AR-beams). The anterior beam collimator size was determined in the same manner as in the first plan. The right beam collimator size was determined to encompass the whole of the liver with a 2-cm margin in all directions. The third plan was constructed using three beams, that is, anterior-right-posterior beams (ARP-beams). Each collimation size was determined in the same manner as in the first and second plans. In the present study, each beam was equally weighted. Comparison of the treatment plans The SERA system can provide dose–volume histogram (DVH) data for the volume of each of the tumors or for the liver as a whole. For comparison, all treatment plans were normalized to deliver mean doses of 5 Gray-equivalent (Gy-Eq) to the whole of the liver. The maximum, mean, and minimum doses to the tumors and the liver were assessed for each plan. The success of treatment of multiple liver tumors with radiotherapy depends on the dose gradient between the tumors and the normal liver. In the present study, the therapeutic gain factor was defined as the ratio of the minimum dose to the tumor (Dmin for the tumor) to the maximum dose to the liver (Dmax for the liver). To achieve large therapeutic gain factors for all liver tumors, a homogeneous distribution of the thermal neutron fluence across the whole of the liver is preferable. To evaluate the distribution of the thermal neutron fluence, the inhomogeneity index of thermal neutron fluence for the whole of the liver was defined as the ratio of the difference between the maximum neutron fluence (⌽ max) and the minimum neutron fluence (⌽ min) relative to the mean neutron fluence (⌽ mean).
894
I. J. Radiation Oncology
● Biology ● Physics
Table 2. Summary of dose–volume histogram data showing averages (with range) for the tumors in the right lobe Beam directions
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
AP beam
Maximum dose (Gy-Eq)
36.5 45.6 61.5 (20.6–65.5) (27.7–86.0) (43.8–107.2) AR beam 46.1 60.5 76.1 (9.3–65.5) (21.5–88.8) (46.8–112.6) ARP beam 50.3 65.1 86.1 (17.6–69.0) (27.3–80.6) (42.5–108.0)
Therapeutic gain factor 3.8 (2.5–8.0) 4.5 (1.0–7.2) 6.1 (2.1–8.3)
Volume 58, Number 3, 2004
Table 3. Summary of dose–volume histogram data showing averages (with range) for the tumors in the left lobe Beam directions
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
Maximum dose (Gy-Eq)
Therapeutic gain factor
AP beam
36.5 53.3 81.1 3.8 (15.2–59.1) (35.1–82.8) (47.1–104.7) (1.5–6.7) AR beam 30.6 45.1 67.9 3.5 (8.8–47.3) (27.8–66.1) (38.7–83.6) (1.0–5.9) ARP beam 27.4 39.9 58.3 3.5 (11.7–40.8) (26.1–56.6) (34.6–71.5) (1.41–5.2)
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior.
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior.
RESULTS
greater than those by AP- or AR-beams (6.1 ⫾ 2.1 vs. 3.8 ⫾ 1.8, p ⫽ 0.0079; and 6.1 ⫾ 2.1 vs. 4.5 ⫾ 2.1, p ⫽ 0.0023). The average of mean doses delivered to the tumors in the left lobe by the AP-beam was 53.3 ⫾ 23.4 Gy-Eq, which was higher than 45.1 ⫾ 19.4 Gy-Eq by the AR-beams or 39.9 ⫾ 15.5 Gy-Eq by the ARP-beams, but not significantly so (p ⫽ 0.0678 and p ⫽ 0.0563, respectively). The average of therapeutic gain factors for the tumors in the left lobe were 3.8 ⫾ 2.4 (AP), 3.5 ⫾ 2.2 (AR), and 3.5 ⫾ 1.8 (ARP), respectively.
DVH analysis for the liver tumors The position in which the tumors were located was expected to affect the dose distribution in the tumors. Thus, the DVH analyses for the tumors were summarized separately for tumors in the left and right lobes. The total number of tumors in the right and left lobes in all cases were 9 and 4, respectively. Tables 2 and 3 show the DVH analysis for the tumors including the maximum, mean, and minimum dose (Gy-Eq) and the therapeutic gain factors for the tumors in the right and left lobes. For all of the tumors in the right or left lobe, all of the treatment plans accomplished therapeutic gain factors of greater than 1. Under the condition that the mean dose to the whole of the liver was normalized to 5 Gy-Eq, the average of the mean doses delivered to the tumors in the right lobe by ARP-beams was significantly higher than that by AP-beams (65.1 ⫾ 19.5 vs. 45.6 ⫾ 19.1 Gy-Eq, p ⫽ 0.0255). Steep decrease in thermal neutron fluence within the body yielded wide distribution of the mean doses among the tumors. The average of therapeutic gain factors for the tumors in the right lobe by ARP-beams was significantly
DVH analysis for the normal liver Table 4 summarizes the DVH analysis for the liver including the maximum and minimum dose (Gy-Eq) and the inhomogeneity index by each plan. Under the condition that the mean dose to the whole liver was normalized to 5 Gy-Eq, the average maximum and minimum dose delivered to the whole liver in each plan ranged from 8.0 to 10.1 Gy-Eq and from 1.1 to 1.2 Gy-Eq, respectively. The average inhomogeneity index of the thermal neutron fluence by ARP-beams was 1.74, which was the lowest of the three treatment plans. Figure 1 compares the distribution of the
Fig. 1. Thermal neutron fluence distribution delivered by an AP-beam (a), AR-beam (b), and ARP-beam (c). The contour of the liver is outlined as a solid white line. The distributions of the thermal neutron fluence are shown with color wash display as follows; ⬎100% Blue, 90 –100% Green, 80 –90% Red, 70 – 80% White, 60 –70% Yellow, 50 – 60% Magenta, 40 –50% Cyan, 30 – 40% Blue, 20 –30% Green, 10 –20% Red, and 0 –10% white.
BNCT treatment planning for liver tumors
Table 4. Summary of dose–volume histogram data showing averages (with range) for the liver Beam directions AP beam AR beam ARP beam
Minimum dose (Gy-Eq)
Mean dose (Gy-Eq)
Maximum dose (Gy-Eq)
Inhomogeneity index
1.2 (0.8–1.3) 1.2 (0.8–1.3) 1.1 (0.8–1.3)
5.0*
10.1 (8.8–11.3) 8.8 (8.1–9.1) 8.0 (7.6–8.4)
2.39 (1.82–2.73) 1.94 (1.69–2.05) 1.74 (1.61–1.87)
5.0* 5.0*
Abbreviations: AP ⫽ anterior-posterior; AR ⫽ anterior-right; ARP ⫽ anterior-right-posterior. * The mean dose to the liver normalized to 5.0 Gy-Eq.
thermal neutron fluence delivered by the AP-, AR-, and ARP-beams. The ARP-beam can be seen to provide a more homogeneous distribution of thermal neutrons compared with the AP- or AR-beams. DISCUSSION The present study using the SERA system, a currently available BNCT treatment planning system, shows that BSH/lipiodol-BNCT has potential as a new treatment option for multiple liver tumors. In contrast with multimodal treatment options for solitary liver tumors, that is, surgical resections, thermal ablations, percutaneous ethanol injections, cryoablations, heavy charged particle irradiation, and stereotactic irradiation with photons (15–20), the treatment options for multiple liver tumors have been limited. Although the intra-arterial injection of chemotherapeutic agents with or without embolization is effective for multiple liver tumors, recurrent liver tumors, which are refractory to chemotherapeutic agents, develop after the repetition of this treatment (21, 22). Our novel method, BSH/lipiodol-BNCT, may offer a more promising option. Two or three liver tumors visible on CT or magnetic resonance imaging (MRI) may be treated successfully using three-dimensional conformal radiotherapy or intensity-modulated radiotherapy with large dose-gradients between the tumors and the surrounding normal liver (23, 24). However, any type of external radiotherapy cannot treat occult metastatic liver tumors undetectable by CT or MRI. Because lipiodol has been used clinically as a contrast agent for detecting occult liver metastases or hepatocellular carcinomas, BSH accompanied by lipiodol is expected to accumulate in the occult micrometastatic tumors as well as obvious tumors (25, 26). Thus, BSH/lipiodol-BNCT has potential in the treatment of both visible and occult micrometastatic tumors. For multiple liver tumors to be successfully treated with radiation therapy including BNCT, the therapeutic gain factors defined as Dmin for the tumor/Dmax for the liver should be greater than 1 for all liver tumors. One of the most difficult problems in applying BNCT to deep-seated tumors such as
● M. SUZUKI et al.
895
liver tumors is the poor penetration of thermal neutrons. To improve the depth distribution of thermal neutron fluence, an epithermal neutron beam is preferable compared with a thermal neutron beam. At Kyoto University Research Reactor, the Heavy Water Neutron Irradiation Facility was reconstructed from November 1995 to March 1996 to improve neutron capture therapy (27). At the updated facility, almost pure thermal to epithermal neutron beams are available. In the present study, we applied the beam data on the epithermal neutron beam of the Kyoto University Research Reactor to the dose calculation using the SERA system. To obtain large therapeutic gain factors for all obvious or occult liver tumors, a homogeneous distribution of thermal neutron fluence and sufficient neutron fluence to the deep portion in the liver is a prerequisite. Pinelli et al. (28) reported a novel method for delivering sufficient and homogenous thermal neutron fluence to the whole of the liver. In that work, a patient suffering multiple liver metastases from colon cancer was treated with BNCT. The liver was surgically extracted and the isolated liver was then moved to a reactor for irradiation with thermal neutrons. The irradiation facility was designed to irradiate the whole of the liver homogeneously in all directions. After irradiation, the liver was autotransplanted back to the patient. The patient was in good general condition and the multiple liver metastases had disappeared 7 months after BNCT. Their success encourages future work in the field of BNCT. In contrast to the treatment protocol by Pinelli et al., the procedure for treating multiple liver tumors with BSH/lipiodol-BNCT is almost noninvasive except for intra-arterial administration of BSH/lipiodol emulsion. In our protocol, after receiving intra-arterial administration of BSH/lipiodol emulsion, the patient is irradiated with epithermal neutron beams collimated to the liver from two or three directions in one fraction. Provided that the 10B concentration (200 ppm) is high with an acceptable T/L 10B concentration ratio (⬎13), the BSH/lipiodol-BNCT protocol using two or three epithermal neutron beams can offer therapeutic gain factors of greater than 1 for all of the liver tumors analyzed in the present study. Compared with AP- or AR- beams, the ARP-beams can deliver the most homogeneous distribution of thermal neutron fluence to the whole of the liver, and provide the greatest therapeutic gain factors for tumors in the right lobe, along with approximately equal therapeutic gain factors for tumors in the left lobe. CONCLUSIONS BSH/lipiodol-BNCT using two or three epithermal neutron beams has potential in the treatment of multiple liver tumors with curative intent, if high 10B concentrations in the liver tumor (200 ppm) and a high T/L 10B concentration ratio (⬎13) are achieved. From a dosimetric viewpoint, BSH/lipiodol-BNCT using three epithermal neutron beams (ARP-beams) is the most suitable for the treatment of multiple liver tumors.
896
I. J. Radiation Oncology
● Biology ● Physics
Volume 58, Number 3, 2004
REFERENCES 1. Coderre JA, Morris GM. The radiation biology of boron neutron capture therapy. Radiat Res 1999;151:1–18. 2. Hatanaka H, Nakagawa Y. Clinical results of long-surviving brain tumor patients who underwent boron neutron capture therapy. Int J Radiat Oncol Biol Phys 1994;28:1061–1066. 3. Nakagawa Y, Hatanaka H. Boron neutron capture therapy, clinical brain tumor studies. J Neurooncol 1997;33:105–115. 4. Takagaki M, Oda Y, Miyatake S, et al. Boron neutron capture therapy: Preliminary study of BNCT with sodium borocaptate (Na2B12H11SH) on glioblastoma. J Neurooncol 1997;35:177– 185. 5. Sauerwein W, Moss R, Rassow J, et al. Organization and management of the first clinical trial of BNCT in Europe (EORTC Protocol 11961). EORTC BNCT study group. Strahlenther Onkol 1999;175:108–111. 6. Kageji T, Nagahiro S, Kitamura K, et al. Optimal timing of neutron irradiation for boron neutron capture therapy after intravenous infusion of sodium borocaptate in patients with glioblastoma. Int J Radiat Oncol Biol Phys 2001;51:120–130. 7. Suzuki M, Masunaga S, Kinashi Y, et al. The effects of boron neutron capture therapy on liver tumors and normal hepatocytes in mice. Jpn J Cancer Res 2000;91:1058–1064. 8. Suzuki M, Masunaga S, Kinashi Y, et al. Intra-arterial administration of sodium borocaptate (BSH)/lipiodol emulsion delivers B-10 to liver tumors highly selectively for boron neutron capture therapy: Experimental studies in the rat liver model. Int J Radiat Oncol Biol Phys; in press. 9. Nigg DW, Wemple CA, Wessol DE, et al. SERA—An advanced treatment planning system for neutron capture therapy and BNCT. Trans Am Nucl Soc 1999;80:66–68. 10. Wojnecki C, Green S. A preliminary comparative study of two treatment planning systems developed for boron neutron capture therapy: MacNCTPlan and SERA. Med Phys 2002;29: 1710–1715. 11. Nichols TL, Kabalka GW, Miller LF, et al. Improved treatment planning for boron neutron capture therapy for glioblastoma multiforme using fluorine-18 labeled boronophenylalanine and positron emission tomography. Med Phys 2002;29: 2351–2358. 12. Chadha M, Capala J, Coderre JA, et al. Boron neutron capture therapy (BNCT) for glioblastoma multiforme (GBM) using the epithermal neutron beam at the Brookhaven National Laboratory. Int J Radiat Oncol Biol Phys 1998;40:829–834. 13. Coderre JA, Makar MS, Micca PL, et al. Derivations of relative biological effectiveness for the high-LET radiations produced during boron neutron capture irradiations of the 9L rat gliosarcoma in vitro and in vivo. Int J Radiat Oncol Biol Phy 1993;27:1121–1129. 14. Ono K, Nagata Y, Akuta K, et al. Frequency of micronuclei in hepatocytes following X and fast-neutron irradiations: An analysis by a linear-quadratic model. Radiat Res 1990;123: 345–347.
15. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: Analysis of 1001 consecutive cases. Ann Surg 1999; 230:309–321. 16. Lau WY, Thomas WT, Leung MD, et al. Percutaneous local ablative therapy for hepatocellular carcinoma. Ann Surg 2003; 237:171–179. 17. Teratani T, Ishikawa T, Shiratori Y, et al. Hepatocellular carcinoma in elderly patients: Beneficial therapeutic efficacy using percutaneous ethanol injection therapy. Cancer 2002; 95:816–823. 18. Mala T, Edwin B, Tillung T, et al. Percutaneous cryoablation of colorectal liver metastases: Potentiated by two consecutive freeze-thaw cycles. Cryobiology 2003;46:99–102. 19. Ohara K, Okumura T, Thujii H, et al. Clearance of parenchymal tumors following radiotherapy: Analysis of hepatocellular carcinomas treated by proton beams. Radiother Oncol 1996; 41:233–236. 20. Sato M, Uematsu M, Shioda A, et al. Feasibility of frameless stereotactic high-dose radiation therapy for primary or metastatic liver cancer. J Radiosurg 1998;1:233–238. 21. Camma´ C, Schepis F, Orlando A, et al. Transarterial chemoembolization for unresectable hepatocellular carcinoma: Meta-analysis of randomized controlled trials. Radiology 2002;224:47–54. 22. Maeda S, Fujiyama S, Tanaka M, et al. Survival and local recurrence rates of hepatocellular carcinoma patients treated by transarterial chemolipiodolization with and without embolization. Hepatol Res 2002;23:202–210. 23. Park HC, Seong J, Han KH, et al. Dose-response relationship in local radiotherapy for hepatocellular carcinoma. Int J Radiat Oncol Biol Phys 2002;54:150–155. 24. Hugo GD, Agazaryan N, Solberg TD, et al. An evaluation of gating window size, delivery method, and composite field dosimetry of respiratory-gated IMRT. Med Phys 2002;29: 2517–2525. 25. Yumoto Y, Jinno K, Tokuyama K, et al. Hepatocellular carcinoma detected by iodized oil. Radiology 1985;154:19–24. 26. Nakakuma K, Tashiro S, Hiraoka T, et al. Hepatocellular carcinoma and metastatic cancer detected by iodized oil. Radiology 1985;154:15–17. 27. Sakurai Y, Kobayashi T. The medical-irradiation characteristics for neutron capture therapy at the Heavy Water Neutron Irradiation Facility of Kyoto University Research Reactor. Med Phys 2002;29:2328–2337. 28. Pinelli T, Zonta A, Altieri S, et al. TAOeMINA: From the first idea to the application to the human liver. In: Sauerwein W, Moss R, Wittig A, editors. Research and development in neutron capture therapy. Bologna: Monduzzui Ediotore, 2002: p. 1065–1072.