- Email: [email protected]
Current clinical results of the Tsukuba BNCT trial
ARTICLE IN PRESS
Applied Radiation and Isotopes 61 (2004) 1089–1093
Current clinical results of the Tsukuba BNCT trial T. Yamamotoa, A. Matsumuraa,*, K. Nakaia, Y. Shibataa, K. Endoa, F. Sakuraib, T. Kishib, H. Kumadab, K. Yamamotob, Y. Toriib a
Department of Neurosurgery, Institute of Clinical Medicine, University of Tsukuba, Tenno-dai 1-1-1, Tsukuba City, Ibaraki 305-8575, Japan b Department of Research Reactor, Tokai Research Establishment, Japan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Ibaraki 319-1195, Japan
Abstract Nine high grade gliomas (5 glioblastomas and 4 anaplastic astrocytomas) were treated with BSH-based intaoperative boron neutron capture therapy (IOBNCT). BSH (100 mg/kg body weight) was intravenously injected, followed by single fraction irradiation using the mixed thermal/epithermal beam of Japan Research Reactor 4. The blood boron level at the time of irradiation averaged 29.9 (18.8–39.5) mg/g. The peak thermal neutron flux as determined by postirradiation measurements varied from 1.99 to 2.77 109 n cm 2 s 1. No serious BSH-related toxicity was observed in this series. The interim survival data in this study showed median survival times of 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma, results which are consistent with the current conventional radiotherapy with/ without boost radiation. Of the 4 residual tumors, 2 showed complete response (CR) and 2 showed partial response (PR) within 6 months following BNCT. No linear correlation was proved between the dose and the occurrence of early neurological events. The maximum boron dose of 11.7–12.2 Gy in the brain related to the occurrence of radiation necrosis. The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities on BSH-based IOBNCT proved to be safe and effective in this series. r 2004 Elsevier Ltd. All rights reserved. Keywords: Anaplastic astrocytoma; BNCT; BSH; Clinical; Epithermal; Glioblastoma
1. Introduction Malignant gliomas are refractory to current combined modalities such as surgery and radiochemotherapy. Difficulties in the therapeutic approach to the diffuse microinvasion into the brain parenchyma with low intrinsic radiosensitivity relative to normal brain tolerance result in poor prognosis (Yaes, 1989; Wallner et al., 1989). Theoretically, boron neutron capture therapy (BNCT) would be able to kill these tumors selectively with minimum damage to the circumscribing normal *Corresponding author. Tel.: +81-29-853-3220; fax: +8129-853-3214. E-mail address: [email protected] (A. Matsumura).
tissue due to high LET particles with a short path length ( 10 mm). In the early clinical trials of BNCT, a thermal beam had been combined with intraoperative techniques to overcome the poor penetrability of thermal neutrons (Sweet, 1997). Epithermal neutrons can convert to thermal neutrons in tissue, resulting in an increase in the number of thermal neutrons delivered to deep-seated lesions. The Japan Research Reactor 4 (JRR-4), a light water-moderated and -cooled pool-type reactor had been reconstructed at 3.5 MW; a medical irradiation facility was installed in JRR-4 that had an exclusive operating room for intraoperative BNCT (IOBNCT) in October 1998. At that time JRR-4 allowed variation of the neutron spectrum such as an epithermal beam and a mixed thermal/epithermal beam. A new clinical phase
0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2004.05.010
ARTICLE IN PRESS 1090
T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
I/II trial has been conducted for IOBNCT using a mixed thermal/epithermal beam at JRR-4 in the Japan Atomic Energy Research Institute (JAERI). The treatment protocol had been designed to determine the therapeutic dose of IOBNCT using the mixed thermal/epithermal beam as well as to optimize the JAERI Computational Dose Planning System (JCDS). The procedure of IOBNCT and the current results of the clinical trial will be also described.
2. Materials and methods Patient selection. Patients from 18 to 70 years of age with a supratentorial unilateral tumor (a tumor no deeper than 6 cm from the brain surface, which has histologic proof of anaplastic astrocytoma or glioblastoma), with a Karnofsky Performance Score (KPS) of 70 or more, who had no previous chemotherapy or radiotherapy, no double cancer and no previous therapy for any other cancers and no allergy to BSH were evaluated for eligibility for the protocol. BSH administration and boron measurement. Prior to neutron irradiation, BSH in 500 ml saline solution was intravenously injected over 1–1.5 h to deliver 100 mg BSH/kg body weight. Blood samples were drawn 0, 1, 3, 6, 9, and 12 h after intravenous injection (or just before irradiation) to estimate the blood boron level based on a clearance curve; a blood sample was also taken immediately following the irradiation to retrospectively correct the mean blood boron level during the irradiations. The blood samples were measured by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) and Prompt Gamma-ray Spectroscopy (PGA). Intraoperative irradiation. Neutron irradiation was performed in a single fraction using a mixed thermal/ epithermal beam of JRR-4. The beam characteristics were described elsewhere (Yamamoto et al., 2003). Briefly, the thermal (o0.53 eV), epithermal (0.53 eV– 10 keV) and fast (>10 keV) neutron flux and g rayabsorbed dose for the beam were 2.0 109, 9.0 108, 3.6 105 n cm 2 s 1, and 3.6 Sv/h, respectively, at a reactor power of 3.5 MW. The peak vascular boron dose of the brain surface, approximately one-third of the boron dose for blood (Rydin et al., 1976), was planned not to exceed 10.8 Gy, which was calculated by estimating the mean blood boron level and directly measuring thermal neutron flux using gold wires withdrawn from the patient’s brain. Irradiation time was adjusted to approximately 1–2 h at a reactor power ranging from 1.5 to 2 MW. Following the open craniotomy, a lithium thermoplastic cover was applied around the craniotomy to eliminate the neutron exposure of the normal structure. The patient’s head position was simulated in the
operating room using the setting simulator, which has a geometrical replica of the neutron irradiation port and wall. The head position in relation to the beam port is precisely delineated using laser beams of the setting simulator and the 3D digitizer to calculate the 3D dose distribution by JCDS. BNCT was carried out under open craniotomy to irradiate the post-surgical cavity directly. Neutron irradiation was performed under general anesthesia, then the gold wires and TLD(s) were withdrawn from the patient’s brain, and the closure operation was performed. Patient follow-up and data analyses. A local hospital supplied pre- and post-BNCT patient care for a week. All of the patients were followed-up clinically and by MRI every 1–3 months. Based on pre-irradiation CT and MRI data, JCDS was optimized by comparing the dose calculation to the direct measurements of the remaining gold wires that were removed from the brain after irradiation. Neurological events and MRI findings were compared to the dose distributions of the individual dose component. The tumorto-blood 10B concentration ratio of 1.0 was used to estimate the boron dose delivered to the tumor tissue.
3. Results Nine patients with histopathologically diagnosed malignant gliomas (5 glioblastoma, 4 anaplastic astrocytoma) were included in this study lasting from October 1999 to July 2002 (Table 1). This series consisted of 6 females and 3 men with an average age of 51.6 years (range 20–66). In 7 of 9 patients, a gross total resection was performed in the first operation. Open biopsy (Case 6) and partial resection (Case 1) were carried out in the remaining 2 patients. The performance status at irradiation was good with a median KPS of 90. One patient had suffered from disseminated intravascular coagulation (DIC) immediately following the first operation and scored a KPS of 40 at the time of irradiation. No serious BSH-related toxicity was observed in this series. Mild and transient adverse effects such as erythema, low grade fever, itching and vascular pain were detected following or during infusion, which were interpreted as probable BSH-related toxicity. Mild peripheral vasocontruction was observed when general anesthesia was maintained, but it was relieved by intravenous infusion and retaining warmth in all cases. Three patients experienced acute peripheral oculomotor palsy with swelling of orbital muscles within a couple of days after BNCT. One patient recovered spontaneously, and the others improved as result of steroid administration. Three seizures, 1 dysarthria and 1 case of blurred vision were recorded; however, no apparent
ARTICLE IN PRESS T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
1091
Table 1 Clinical summary of the patients treated with IOBNCT No.
Age, sex
Pathologya
Surgeryb
Follow-up (month)
Responsec
Recurrenced
10
B level (mg/g)
max. nth-flux (n cm 2)
Cause of death
1 2 3 4 5 6 7 8 9
55, 46, 66, 20, 38, 64, 58, 66, 51,
GM GM AA AA GM GM AA AA GM
P T T T T Biopsy T T T
15.4 33.8 16.8 44.9 28.7 8.8 35.0A 3.8 23.2A
PR CR NT CR NT CR NT NT NT
No No Local Distant Distant Subarachnoid No Subarachnoid Distant
18.8 27.0 25.5 32.1 39.5 19.7 38.8 31.3 36.0
2.8 1013 2.5 1013 2.0 1013 2.6 1013 2.6 1013 2.3 1013 2.5 1013 1.6 1013 2.1 1013
Pneumonia Sepsis Pancytopenia Tumor progression Tumor progression Subarachnoid seeding — Subarachnoid seeding —
F F M F F F F M M
A: alive. a All tumors were diagnosed histopathologically as glioblastoma (GM) or anaplastic astrocytoma (AA). b The extent of surgical resection was categorized macroscopically as gross total removal (T), partial removal (P) and biopsy. c The maximum response on MRI (or CT) at 6 months after IOBNCT: complete response (CR), partial response (PR), no residual tumor (NT). No progression was recorded. d Recurrent tumors were categorized as being a local recurrence in or adjacent to the primary site, a distant and isolated recurrent intraparenchymal tumor, or a subarachnoid (intraventricle and/or meningeal) seeding tumor.
relation of the neurological events to the dose of IOBNCT was found. One seizure accompanied with post-ictal brain swelling required surgical intervention. There have been 3 cases with delayed radiation-induced necrosis, of which 2 underwent surgery to remove necrotic tissue. The blood boron level at the time of irradiation averaged 29.9 (18.8–39.5) mg/g. The peak thermal neutron flux according to the post-irradiation measurements varied from 1.99 to 2.77 109 n cm 2 s 1. The minimum boron dose for the tumor and target volume averaged 15.9 Gy (range 7.5–24.6 Gy) and 7.3 Gy (range 3.7–11.9 Gy), respectively. At present 7 (4 glioblastoma and 3 anaplastic astrocytoma) of the 9 patients have died. The four tumor-related deaths consisted of 2 subarachnoid seeding and 2 distant, isolated intraparenchymal recurrences. One patient (Case 3) recurred locally 10 months after BNCT, and re-operation and gamma-knife therapy against the recurrence site were performed. The patient died from drug-induced pancytopenia. A distant recurrence was seen in Case 4 in the deep temporal region 14 months after BNCT. Following the re-operation and additional conventional radiotherapy, this patient returned to college and graduated. She died from a distant intraparenchymal tumor recurrence at 49.8 months of follow-up. In Case 6, once the tumor had completely disappeared on contrast-enhanced CT scan, the patient expired from subarachnoid seeding at 8.8 months of follow-up. In Case 8, the patient’s general status considerably deteriorated due to systemic disease, immediately after which a fatal subarachnoid seeding occurred at 3.8 months of follow-up. The mean followup time of the 9 patients was 24.3 months (3.8–49.8). At
present, the median survival time is 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma. Four possible residual tumors showed complete response (CR) in 2 patients and partial response (PR) in 2 patients in the serial MRI and/or CT scans within 6 months after BNCT.
4. Discussion The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities on BSH-based IOBNCT proved to be safe. No serious BSH-related toxicity was observed in this series. Although vasocontruction has not been previously reported as a BSH-related toxicity, the vascular pain in patients implies a similar vascular irritability to BSH. The maximum dose of the BSHbased IOBNCT has to be assessed by the moderate and mild adverse events in the early or delayed phase in the specific cases. Radiation-induced lesions on the oculomotor nerves running in the cavernous sinus or orbit are possible cause of the oculomotor palsy. Although the dose in the unilateral cavernous sinus, orbit and normal structure tended to be high in the specific patients, a linear correlation was not proved between the dose and the occurrence of the event. In the other patients with thickened lithium cover around the orbit, no oculomotor palsy occurred, except in one patient with transient and mild palsy. In our series, radiation necrosis developed within or adjacent to the target volume that included 2 cm beyond the surgical margin or residual tumor. One area of radiation necrosis appeared to be in a ‘‘wedge shape’’, implying the influence of neutron distribution and boron
ARTICLE IN PRESS 1092
T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
accumulation in the periutumoral edematous brain and vasculature. The boron concentration in the normal brain is thought to be quite low; however, that in the peritumoral edematous brain has been contraversial. When the tumor-to-blood-boron-concentration ratio was assumed to be 1.0, the edematous brain-to-blood ratio of 0.14–0.32 (Haselsberger et al., 2002). The maximum boron dose in the edematous brain of more than 12.0 Gy was related to the occurrence of radiation necrosis when edematous brain-to-blood ratio of 0.32 was employed in the JCDS calculations. In 3 cases of radiation necrosis, the maximum boron dose in the edematous brain averaged 12.2 Gy (range 12.1–12.2 Gy). In 6 patients without radiation necrosis, the maximum boron dose averaged 10.7 Gy (range 8.9–11.9 Gy). However, two of the 6 patients without radiation necrosis died of subarachnoid seeding within a year, i.e., the maximum boron dose of 11.7 and 11.9 Gy seems to have a potential risk of radiation necrosis. The dose of 11.7–12.2 Gy as a predictor of radiation necrosis is lower than the reported boron vascular dose in the brain of 2178.1 Gy for 19 cases of radiation necrosis in 159 post-BNCT patients (Nakagawa et al., 2002), even after standardization by vascular dose estimation and by the brain-to-blood ratio. The majority of the reported patients had been treated with thermal beams, resulting in different levels of relative biological effectiveness (RBE) and doses of the other radiation components. Moreover, radiation necrosis developed in the cases with large target volumes, i.e. relatively large tumors before surgery and brain edema. The average target volumes in the cases with and without radiation necrosis were 168.7 and 127.3 cm3, respectively (p=0.025 by Mann–Whitney’s U test). The interim survival data in this study showed a median survival time of 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma, which was consistent with the current conventional radiotherapy with or without boost radiation. However, this study did have limitations. First, only a small number of patients with relatively good performance status were entered in this series. Immediately pre- and post- BNCT deterioration due to BNCT-unrelated, extracranial problems (Case 6 and Case 8) may have affected and diminished the survival time. Greater penetrability of epithermal neutrons, as well as the advantage of intraoperative irradiation (Harrison et al., 1998) by the irradiation supplied by the reflected skin method and the craniotomy brought about better thermal neutron distribution, especially in deep-seated tumors. Insertion of an airfilled balloon as a ‘‘void’’ into the cavity was also utilized as part of the IOBNCT procedure. The measurements of the two-dimensional thermal neutron distribution experimentally proved the effectiveness of the ‘‘void’’ method in IOBNCT. When an air balloon is placed in a surgical defect of the brain, an increase of
thermal neutron flux is observed not only in the direction of the beam axis but also in the vertical and horizontal directions (Yamamoto et al., 2002). The interpretation of the failure pattern in relation to boron or total dose distribution was made difficult by the limited number of cases in this study. The minimum target boron dose in the patients without recurrence averaged 6.9 Gy (range 6.3–7.7 Gy), while local recurrence occurred when a dose of 6.8 Gy had been administrated.
5. Conclusions The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities in BSH-based IOBNCT proved to be safe and effective in this series. After optimization of JCDS and careful evaluation of the clinical data, a new protocol can be planned to introduce BSH-based IOBNCT using a epithermal neutron beam in which JCDS-based dose planning and calculation are utilized.
Acknowledgements This project was supported by the Fund-in Trust for Cancer Research from the Governor of Ibaraki Prefecture, Japan.
References Harrison, L.B., Minsky, B.D., Enker, W.E., et al., 1998. High dose rate intraoperative radiation therapy (HDR-IORT) as part of the management strategy for locally advanced primary and recurrent rectal cancer. Int. J. Radiat. Oncol. Biol. Phys. 42, 325–330. Haselsberger, K., Radner, H., Pendl, G., 2002. Tissue sampling for BNCT of Glioblastoma. In: Sauerwein, W., Moss, R.L., Wittig, A. (Eds.), Research and Development in Neutron Capture Therapy. Monduzzi Editore, Bologna, Italy, pp. 1013–1116. Nakagawa, Y., Pooh, K., Kageji, T., et al., 2002. Clinical review of boron neutron capture therapy and a proposed strategy using epithermal neuron beams in Japan. In: Sauerwein, W., Moss, R.L., Wittig, A. (eds.), Research and Development in Neutron Capture Therapy. Monduzzi Editore, Bologna, Italy, pp. 1113–1116. Rydin, R.A., Deutsch, O.L., Murray, B.W., 1976. The effect of geometry on capillary wall dose for boron neutron capture therapy. Phys. Med. Biol. 21, 134–138. Sweet, W.H., 1997. Early history of development of boron neutron capture therapy of tumors. J. Neurooncol. 33, 19–26. Yaes, R.J., 1989. Tumor heterogeneity, tumor size, and radioresistance. Int. J. Radiat. Oncol. Biol. Phys. 17, 993–1005.
ARTICLE IN PRESS T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093 Yamamoto, T., Matsumura, A., Yamamoto, K., et al., 2002. In-phantom two-dimensional thermal neutron distribution for intraoperative boron neutron capture therapy of brain tumours. Phys. Med. Biol. 47, 2387–2396. Yamamoto, T., Matsumura, A., Yamamoto, K., et al., 2003. Characterization of neutron beams for boron neutron
1093
capture therapy: in-air radiobiological dosimetry. Radiat. Res. 160, 70–76. Wallner, K.E., Galicich, J.H., Krol, G., et al., 1989. Patterns of failure following treatment for glioblastoma multiforme and anaplastic astrocytoma. J. Radiat. Oncol. Biol. Phys. 16, 1405–1409.
Applied Radiation and Isotopes 61 (2004) 1089–1093
Current clinical results of the Tsukuba BNCT trial T. Yamamotoa, A. Matsumuraa,*, K. Nakaia, Y. Shibataa, K. Endoa, F. Sakuraib, T. Kishib, H. Kumadab, K. Yamamotob, Y. Toriib a
Department of Neurosurgery, Institute of Clinical Medicine, University of Tsukuba, Tenno-dai 1-1-1, Tsukuba City, Ibaraki 305-8575, Japan b Department of Research Reactor, Tokai Research Establishment, Japan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Ibaraki 319-1195, Japan
Abstract Nine high grade gliomas (5 glioblastomas and 4 anaplastic astrocytomas) were treated with BSH-based intaoperative boron neutron capture therapy (IOBNCT). BSH (100 mg/kg body weight) was intravenously injected, followed by single fraction irradiation using the mixed thermal/epithermal beam of Japan Research Reactor 4. The blood boron level at the time of irradiation averaged 29.9 (18.8–39.5) mg/g. The peak thermal neutron flux as determined by postirradiation measurements varied from 1.99 to 2.77 109 n cm 2 s 1. No serious BSH-related toxicity was observed in this series. The interim survival data in this study showed median survival times of 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma, results which are consistent with the current conventional radiotherapy with/ without boost radiation. Of the 4 residual tumors, 2 showed complete response (CR) and 2 showed partial response (PR) within 6 months following BNCT. No linear correlation was proved between the dose and the occurrence of early neurological events. The maximum boron dose of 11.7–12.2 Gy in the brain related to the occurrence of radiation necrosis. The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities on BSH-based IOBNCT proved to be safe and effective in this series. r 2004 Elsevier Ltd. All rights reserved. Keywords: Anaplastic astrocytoma; BNCT; BSH; Clinical; Epithermal; Glioblastoma
1. Introduction Malignant gliomas are refractory to current combined modalities such as surgery and radiochemotherapy. Difficulties in the therapeutic approach to the diffuse microinvasion into the brain parenchyma with low intrinsic radiosensitivity relative to normal brain tolerance result in poor prognosis (Yaes, 1989; Wallner et al., 1989). Theoretically, boron neutron capture therapy (BNCT) would be able to kill these tumors selectively with minimum damage to the circumscribing normal *Corresponding author. Tel.: +81-29-853-3220; fax: +8129-853-3214. E-mail address: [email protected] (A. Matsumura).
tissue due to high LET particles with a short path length ( 10 mm). In the early clinical trials of BNCT, a thermal beam had been combined with intraoperative techniques to overcome the poor penetrability of thermal neutrons (Sweet, 1997). Epithermal neutrons can convert to thermal neutrons in tissue, resulting in an increase in the number of thermal neutrons delivered to deep-seated lesions. The Japan Research Reactor 4 (JRR-4), a light water-moderated and -cooled pool-type reactor had been reconstructed at 3.5 MW; a medical irradiation facility was installed in JRR-4 that had an exclusive operating room for intraoperative BNCT (IOBNCT) in October 1998. At that time JRR-4 allowed variation of the neutron spectrum such as an epithermal beam and a mixed thermal/epithermal beam. A new clinical phase
0969-8043/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.apradiso.2004.05.010
ARTICLE IN PRESS 1090
T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
I/II trial has been conducted for IOBNCT using a mixed thermal/epithermal beam at JRR-4 in the Japan Atomic Energy Research Institute (JAERI). The treatment protocol had been designed to determine the therapeutic dose of IOBNCT using the mixed thermal/epithermal beam as well as to optimize the JAERI Computational Dose Planning System (JCDS). The procedure of IOBNCT and the current results of the clinical trial will be also described.
2. Materials and methods Patient selection. Patients from 18 to 70 years of age with a supratentorial unilateral tumor (a tumor no deeper than 6 cm from the brain surface, which has histologic proof of anaplastic astrocytoma or glioblastoma), with a Karnofsky Performance Score (KPS) of 70 or more, who had no previous chemotherapy or radiotherapy, no double cancer and no previous therapy for any other cancers and no allergy to BSH were evaluated for eligibility for the protocol. BSH administration and boron measurement. Prior to neutron irradiation, BSH in 500 ml saline solution was intravenously injected over 1–1.5 h to deliver 100 mg BSH/kg body weight. Blood samples were drawn 0, 1, 3, 6, 9, and 12 h after intravenous injection (or just before irradiation) to estimate the blood boron level based on a clearance curve; a blood sample was also taken immediately following the irradiation to retrospectively correct the mean blood boron level during the irradiations. The blood samples were measured by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) and Prompt Gamma-ray Spectroscopy (PGA). Intraoperative irradiation. Neutron irradiation was performed in a single fraction using a mixed thermal/ epithermal beam of JRR-4. The beam characteristics were described elsewhere (Yamamoto et al., 2003). Briefly, the thermal (o0.53 eV), epithermal (0.53 eV– 10 keV) and fast (>10 keV) neutron flux and g rayabsorbed dose for the beam were 2.0 109, 9.0 108, 3.6 105 n cm 2 s 1, and 3.6 Sv/h, respectively, at a reactor power of 3.5 MW. The peak vascular boron dose of the brain surface, approximately one-third of the boron dose for blood (Rydin et al., 1976), was planned not to exceed 10.8 Gy, which was calculated by estimating the mean blood boron level and directly measuring thermal neutron flux using gold wires withdrawn from the patient’s brain. Irradiation time was adjusted to approximately 1–2 h at a reactor power ranging from 1.5 to 2 MW. Following the open craniotomy, a lithium thermoplastic cover was applied around the craniotomy to eliminate the neutron exposure of the normal structure. The patient’s head position was simulated in the
operating room using the setting simulator, which has a geometrical replica of the neutron irradiation port and wall. The head position in relation to the beam port is precisely delineated using laser beams of the setting simulator and the 3D digitizer to calculate the 3D dose distribution by JCDS. BNCT was carried out under open craniotomy to irradiate the post-surgical cavity directly. Neutron irradiation was performed under general anesthesia, then the gold wires and TLD(s) were withdrawn from the patient’s brain, and the closure operation was performed. Patient follow-up and data analyses. A local hospital supplied pre- and post-BNCT patient care for a week. All of the patients were followed-up clinically and by MRI every 1–3 months. Based on pre-irradiation CT and MRI data, JCDS was optimized by comparing the dose calculation to the direct measurements of the remaining gold wires that were removed from the brain after irradiation. Neurological events and MRI findings were compared to the dose distributions of the individual dose component. The tumorto-blood 10B concentration ratio of 1.0 was used to estimate the boron dose delivered to the tumor tissue.
3. Results Nine patients with histopathologically diagnosed malignant gliomas (5 glioblastoma, 4 anaplastic astrocytoma) were included in this study lasting from October 1999 to July 2002 (Table 1). This series consisted of 6 females and 3 men with an average age of 51.6 years (range 20–66). In 7 of 9 patients, a gross total resection was performed in the first operation. Open biopsy (Case 6) and partial resection (Case 1) were carried out in the remaining 2 patients. The performance status at irradiation was good with a median KPS of 90. One patient had suffered from disseminated intravascular coagulation (DIC) immediately following the first operation and scored a KPS of 40 at the time of irradiation. No serious BSH-related toxicity was observed in this series. Mild and transient adverse effects such as erythema, low grade fever, itching and vascular pain were detected following or during infusion, which were interpreted as probable BSH-related toxicity. Mild peripheral vasocontruction was observed when general anesthesia was maintained, but it was relieved by intravenous infusion and retaining warmth in all cases. Three patients experienced acute peripheral oculomotor palsy with swelling of orbital muscles within a couple of days after BNCT. One patient recovered spontaneously, and the others improved as result of steroid administration. Three seizures, 1 dysarthria and 1 case of blurred vision were recorded; however, no apparent
ARTICLE IN PRESS T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
1091
Table 1 Clinical summary of the patients treated with IOBNCT No.
Age, sex
Pathologya
Surgeryb
Follow-up (month)
Responsec
Recurrenced
10
B level (mg/g)
max. nth-flux (n cm 2)
Cause of death
1 2 3 4 5 6 7 8 9
55, 46, 66, 20, 38, 64, 58, 66, 51,
GM GM AA AA GM GM AA AA GM
P T T T T Biopsy T T T
15.4 33.8 16.8 44.9 28.7 8.8 35.0A 3.8 23.2A
PR CR NT CR NT CR NT NT NT
No No Local Distant Distant Subarachnoid No Subarachnoid Distant
18.8 27.0 25.5 32.1 39.5 19.7 38.8 31.3 36.0
2.8 1013 2.5 1013 2.0 1013 2.6 1013 2.6 1013 2.3 1013 2.5 1013 1.6 1013 2.1 1013
Pneumonia Sepsis Pancytopenia Tumor progression Tumor progression Subarachnoid seeding — Subarachnoid seeding —
F F M F F F F M M
A: alive. a All tumors were diagnosed histopathologically as glioblastoma (GM) or anaplastic astrocytoma (AA). b The extent of surgical resection was categorized macroscopically as gross total removal (T), partial removal (P) and biopsy. c The maximum response on MRI (or CT) at 6 months after IOBNCT: complete response (CR), partial response (PR), no residual tumor (NT). No progression was recorded. d Recurrent tumors were categorized as being a local recurrence in or adjacent to the primary site, a distant and isolated recurrent intraparenchymal tumor, or a subarachnoid (intraventricle and/or meningeal) seeding tumor.
relation of the neurological events to the dose of IOBNCT was found. One seizure accompanied with post-ictal brain swelling required surgical intervention. There have been 3 cases with delayed radiation-induced necrosis, of which 2 underwent surgery to remove necrotic tissue. The blood boron level at the time of irradiation averaged 29.9 (18.8–39.5) mg/g. The peak thermal neutron flux according to the post-irradiation measurements varied from 1.99 to 2.77 109 n cm 2 s 1. The minimum boron dose for the tumor and target volume averaged 15.9 Gy (range 7.5–24.6 Gy) and 7.3 Gy (range 3.7–11.9 Gy), respectively. At present 7 (4 glioblastoma and 3 anaplastic astrocytoma) of the 9 patients have died. The four tumor-related deaths consisted of 2 subarachnoid seeding and 2 distant, isolated intraparenchymal recurrences. One patient (Case 3) recurred locally 10 months after BNCT, and re-operation and gamma-knife therapy against the recurrence site were performed. The patient died from drug-induced pancytopenia. A distant recurrence was seen in Case 4 in the deep temporal region 14 months after BNCT. Following the re-operation and additional conventional radiotherapy, this patient returned to college and graduated. She died from a distant intraparenchymal tumor recurrence at 49.8 months of follow-up. In Case 6, once the tumor had completely disappeared on contrast-enhanced CT scan, the patient expired from subarachnoid seeding at 8.8 months of follow-up. In Case 8, the patient’s general status considerably deteriorated due to systemic disease, immediately after which a fatal subarachnoid seeding occurred at 3.8 months of follow-up. The mean followup time of the 9 patients was 24.3 months (3.8–49.8). At
present, the median survival time is 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma. Four possible residual tumors showed complete response (CR) in 2 patients and partial response (PR) in 2 patients in the serial MRI and/or CT scans within 6 months after BNCT.
4. Discussion The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities on BSH-based IOBNCT proved to be safe. No serious BSH-related toxicity was observed in this series. Although vasocontruction has not been previously reported as a BSH-related toxicity, the vascular pain in patients implies a similar vascular irritability to BSH. The maximum dose of the BSHbased IOBNCT has to be assessed by the moderate and mild adverse events in the early or delayed phase in the specific cases. Radiation-induced lesions on the oculomotor nerves running in the cavernous sinus or orbit are possible cause of the oculomotor palsy. Although the dose in the unilateral cavernous sinus, orbit and normal structure tended to be high in the specific patients, a linear correlation was not proved between the dose and the occurrence of the event. In the other patients with thickened lithium cover around the orbit, no oculomotor palsy occurred, except in one patient with transient and mild palsy. In our series, radiation necrosis developed within or adjacent to the target volume that included 2 cm beyond the surgical margin or residual tumor. One area of radiation necrosis appeared to be in a ‘‘wedge shape’’, implying the influence of neutron distribution and boron
ARTICLE IN PRESS 1092
T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093
accumulation in the periutumoral edematous brain and vasculature. The boron concentration in the normal brain is thought to be quite low; however, that in the peritumoral edematous brain has been contraversial. When the tumor-to-blood-boron-concentration ratio was assumed to be 1.0, the edematous brain-to-blood ratio of 0.14–0.32 (Haselsberger et al., 2002). The maximum boron dose in the edematous brain of more than 12.0 Gy was related to the occurrence of radiation necrosis when edematous brain-to-blood ratio of 0.32 was employed in the JCDS calculations. In 3 cases of radiation necrosis, the maximum boron dose in the edematous brain averaged 12.2 Gy (range 12.1–12.2 Gy). In 6 patients without radiation necrosis, the maximum boron dose averaged 10.7 Gy (range 8.9–11.9 Gy). However, two of the 6 patients without radiation necrosis died of subarachnoid seeding within a year, i.e., the maximum boron dose of 11.7 and 11.9 Gy seems to have a potential risk of radiation necrosis. The dose of 11.7–12.2 Gy as a predictor of radiation necrosis is lower than the reported boron vascular dose in the brain of 2178.1 Gy for 19 cases of radiation necrosis in 159 post-BNCT patients (Nakagawa et al., 2002), even after standardization by vascular dose estimation and by the brain-to-blood ratio. The majority of the reported patients had been treated with thermal beams, resulting in different levels of relative biological effectiveness (RBE) and doses of the other radiation components. Moreover, radiation necrosis developed in the cases with large target volumes, i.e. relatively large tumors before surgery and brain edema. The average target volumes in the cases with and without radiation necrosis were 168.7 and 127.3 cm3, respectively (p=0.025 by Mann–Whitney’s U test). The interim survival data in this study showed a median survival time of 23.2 months for glioblastoma and 25.9 months for anaplastic astrocytoma, which was consistent with the current conventional radiotherapy with or without boost radiation. However, this study did have limitations. First, only a small number of patients with relatively good performance status were entered in this series. Immediately pre- and post- BNCT deterioration due to BNCT-unrelated, extracranial problems (Case 6 and Case 8) may have affected and diminished the survival time. Greater penetrability of epithermal neutrons, as well as the advantage of intraoperative irradiation (Harrison et al., 1998) by the irradiation supplied by the reflected skin method and the craniotomy brought about better thermal neutron distribution, especially in deep-seated tumors. Insertion of an airfilled balloon as a ‘‘void’’ into the cavity was also utilized as part of the IOBNCT procedure. The measurements of the two-dimensional thermal neutron distribution experimentally proved the effectiveness of the ‘‘void’’ method in IOBNCT. When an air balloon is placed in a surgical defect of the brain, an increase of
thermal neutron flux is observed not only in the direction of the beam axis but also in the vertical and horizontal directions (Yamamoto et al., 2002). The interpretation of the failure pattern in relation to boron or total dose distribution was made difficult by the limited number of cases in this study. The minimum target boron dose in the patients without recurrence averaged 6.9 Gy (range 6.3–7.7 Gy), while local recurrence occurred when a dose of 6.8 Gy had been administrated.
5. Conclusions The clinical application of a mixed thermal/epithermal beam and JRR-4 facilities in BSH-based IOBNCT proved to be safe and effective in this series. After optimization of JCDS and careful evaluation of the clinical data, a new protocol can be planned to introduce BSH-based IOBNCT using a epithermal neutron beam in which JCDS-based dose planning and calculation are utilized.
Acknowledgements This project was supported by the Fund-in Trust for Cancer Research from the Governor of Ibaraki Prefecture, Japan.
References Harrison, L.B., Minsky, B.D., Enker, W.E., et al., 1998. High dose rate intraoperative radiation therapy (HDR-IORT) as part of the management strategy for locally advanced primary and recurrent rectal cancer. Int. J. Radiat. Oncol. Biol. Phys. 42, 325–330. Haselsberger, K., Radner, H., Pendl, G., 2002. Tissue sampling for BNCT of Glioblastoma. In: Sauerwein, W., Moss, R.L., Wittig, A. (Eds.), Research and Development in Neutron Capture Therapy. Monduzzi Editore, Bologna, Italy, pp. 1013–1116. Nakagawa, Y., Pooh, K., Kageji, T., et al., 2002. Clinical review of boron neutron capture therapy and a proposed strategy using epithermal neuron beams in Japan. In: Sauerwein, W., Moss, R.L., Wittig, A. (eds.), Research and Development in Neutron Capture Therapy. Monduzzi Editore, Bologna, Italy, pp. 1113–1116. Rydin, R.A., Deutsch, O.L., Murray, B.W., 1976. The effect of geometry on capillary wall dose for boron neutron capture therapy. Phys. Med. Biol. 21, 134–138. Sweet, W.H., 1997. Early history of development of boron neutron capture therapy of tumors. J. Neurooncol. 33, 19–26. Yaes, R.J., 1989. Tumor heterogeneity, tumor size, and radioresistance. Int. J. Radiat. Oncol. Biol. Phys. 17, 993–1005.
ARTICLE IN PRESS T. Yamamoto et al. / Applied Radiation and Isotopes 61 (2004) 1089–1093 Yamamoto, T., Matsumura, A., Yamamoto, K., et al., 2002. In-phantom two-dimensional thermal neutron distribution for intraoperative boron neutron capture therapy of brain tumours. Phys. Med. Biol. 47, 2387–2396. Yamamoto, T., Matsumura, A., Yamamoto, K., et al., 2003. Characterization of neutron beams for boron neutron
1093
capture therapy: in-air radiobiological dosimetry. Radiat. Res. 160, 70–76. Wallner, K.E., Galicich, J.H., Krol, G., et al., 1989. Patterns of failure following treatment for glioblastoma multiforme and anaplastic astrocytoma. J. Radiat. Oncol. Biol. Phys. 16, 1405–1409.