Introduction
Glioblastoma multiforme (GBM) accounts for about 30% of all gliomas and is one of the most malignant neoplasms of the primary intracranial tumours. Commonly used therapeutic modalities for GBM are (1) bulk reductive surgery followed by conventional external beam radiotherapy (EBRT) with 50–60 Gy and (2) chemotherapy using mainly nitrosourea. However, even these multidisciplinary therapies cannot fully control tumour progression, and the 9–14 months of median survival time (MST) achieved in recent reliable reports is not satisfactory.Reference Ramsey and Brand1–Reference Prados, Wara and Sneed10 Since Walker et al.Reference Walker, Strike and Sheline3 reported that dose escalation of local irradiation could increase the tumour-suppressive effect and improve the survival rate of patients with malignant glioma (MG), interstitial brachytherapy (IBT) or intra-operative radiotherapy (IORT) as a local irradiation boost has been performed. The local irradiation boost is a logical radiotherapy, since most GBM cases recur at the primary site, and it actually improves the MST to 16–25 months.Reference Sakai, Yamada, Andoh, Hirata, Nishimura, Miwa, Shimizu and Yanagawa11–Reference Selker, Shapiro and Burger18 Recently, a less invasive and precisely planned local irradiation modality using stereotactic radiosurgery (SRS) for brain tumours has been widely applied. Even in GBM, since the first report in 1992 by Loeffler et al.,Reference Loeffler, Alexander, Shea, Wen, Fine, Kooy and Black19 several clinical results have been published. SRS can be divided generally into the following three modalities: gamma knife (GK), radiosurgery and cyber-knife.
In our institute, IORT was used as an adjunctive local boost in the initial treatment approach of GBM, with conventional EBRT after tumour bulk reductive surgery until 1999. Here, we report clinical results, in a single institute, of GBM treatment using SRS, which has been incorporated into the initial management approach since 2000 in place of IORT, and we compare them with those achieved with IORT. In addition, we assessed whether these local irradiation boost therapies are prognostic factors on the survival analysis of GBM patients.
MATERIALS AND METHODS
Between January 1989 and December 2003, 120 adult patients with newly diagnosed supratentorial GBM were treated at the Department of Neurosurgery, Gifu University Hospital. All patients underwent tumour resection or biopsy, and GBM (World Health Organization Grade IV) was diagnosed histologically. Eighty-seven patients were male, and 33 were female, with ages ranging from 22 to 85 years (mean age, 58.5 years). Pre-operative Karnofsky performance scale (KPS) scoresReference Karnofsky, Burchenal and MacLeod20 ranged from 40 to 100 (mean KPS score, 68.8). Tumours were located in the frontal lobe in 42, the temporal lobe in 36, the parietal lobe in 18, the occipital lobe in 9 and the midline structures/bilateral hemispheres in 15 patients. Tumour size was evaluated as the size at the greatest diameter of the enhanced tumour on Gd-enhanced T1-weighted magnetic resonance (MR) imaging. The number of patients with a tumour size of 5 cm or less was 67, and those with a tumour size of more than 5 cm was 53. In the 120 patients, 31 underwent IORT (IORT group), 29 underwent SRS (SRS group), and the remaining 60 had no local high-dose irradiation boost (non-local boost group). All but 12 died of tumour-related causes before April 2004.
The extent of surgery was evaluated by four neurosurgeons and one neuroradiologist, excluding the chief surgeon in each case, by viewing post-operative Gd-enhanced T1-weighted MR images obtained 1 month after surgery. Gross-total resection of the tumour was defined as the resection with no residual enhancing tumour. Gross-total resection was performed in 56 patients.
A dose of 50–60 Gy of EBRT was administered as an essential adjuvant therapy after the initial surgery in all patients. EBRT was given concurrently with intravenous administration of ACNU [(1-4-amino-2-methyl-5-pyrimidinyl)-methyl-3-nitrosourea] (2 mg/kg) (AR) in 40 patients, with intravenous administration of ACNU (2 mg/kg) and interferon-β (IFN-β) (120,000 IU/kg) (IAR) in 27 patients, and with daily intravenous administration of low-dose cisplatin (5 mg/mReference Walker, Alexander and Hunt2) or carboplatin (15 mg/mReference Walker, Alexander and Hunt2) (R&C/C) in 53 patients.Reference Shinoda, Sakai, Hara, Ueda, Sakai and Nakatani21
Between January 1989 and August 1998, 31 patients received IORT in a prospective clinical study according to the following criteria: patients with tumour size at greatest diameter of 6.5 cm or less, age 75 years or younger, tumour not located in the midline structures/bilateral hemispheres, and who consented to this clinical trial. IORT was performed with craniotomy 1 week after the initial tumour resection using high-energy electrons generated by a microtron (MM-22, Scanditronix AB, Husbyborg, Uppsala, Sweden), which can deliver a dose of 10 Gy/min. The radiation field was determined to include the entire tumour-resected surface and the residual enhancing tumour on computed tomography (CT)/MR imaging by the naked eye and covered using one of seven different diameter collimaters (2, 3, 4, 5, 6, 8 and 10 cm). The electron beam energy was selected so that the 90% isodose line fell at least 1–2 cm below the deepest aspect of the tumour or the tumour-resected surface, where the mean delivered IORT dose was 18.5 Gy (10–25 Gy), and the mean diameter of the collimaters used was 5.2 cm (2–10 cm). EBRT was begun 1 week after IORT. CT/MR imaging was performed routinely 8–12 weeks after IORT and every 3 months thereafter, or more frequently if clinically indicated.
Twenty-four patients were male, and 7 were female, with ages ranging from 22 to 75 years (mean age, 53.1 years), with pre-operative KPS scores ranging from 40 to 90 (mean KPS score, 66.5). Tumours were located in the frontal lobe in 12, the temporal lobe in 7, the parietal lobe in 8, and the occipital lobe in 4 patients. The number of patients with a tumour size of 5 cm or less was 17, and that of more than 5 cm was 14.All but one died of tumour-related causes before April 2004.
Between September 2000 and March 2004, 29 consecutive patients who consented to this clinical trial underwent SRS in a prospective clinical study. SRS was incorporated into the initial management approach in place of IORT after conventional EBRT. SRS was performed using a 10-MV linear accelerator (Mevatron KD2 Primus, Siemens, USA). A Fischer stereotactic head frame (F.L. Fischer, Germany) was used for head fixation during CT/MR imaging planning and treatment. A custom-designed computer system was used to reconstruct three-dimensional models of the tumours and relevant neuroanatomic structures from CT/MR imaging data. With the aid of a beam’s-eye view, a series of arc rotations with one or more isocentres were designed from graphic displays, which permitted optimal distributions of the dose within the treatment volume, but spared the radiosensitive structures nearby. The planning system included software algorithms for calculation of tumour target volume as well as both volumetric and surface dose distributions.
The tumour target volume was defined by the enhancing tumour in cases of residual tumours and by contouring the resected surface in cases with gross total resection on CT/MR imaging, with a mean margin of 2.7 mm (2–3 mm) on each contour. The mean tumour target volume was 43.1 cmReference Walker, Strike and Sheline3 (2.6–195.1 cmReference Walker, Strike and Sheline3). The mean number of isocentres was 3.4 (1–5) and that of arcs per one isocentre was 6.0 (5–6). Dose and normalisation were selected individually, with careful attention to the tumour target volume, the degree of dose homogeneity within the target and the proximity of important radiosensitive structures such as the eyes, optic nerves and chiasm and brainstem. Doses were prescribed to the isodose distribution that covered the target with the margin using a combination of nine different diameter collimators (6–30.6 mm). The radiation dose was normalised to the 85–100% isodose contour, and the mean marginal dose was 20.8 Gy (15–25 Gy). Betamethasone (4 mg/day) was routinely prescribed for 3 days after SRS to prevent acute brain oedema due to high-dose irradiation. MR imaging scans were performed routinely 8–12 weeks after radiosurgery and every 3 months thereafter, or more frequently if clinically indicated.
A representative case of GBM treated by SRS after gross total resection and EBRT is presented (Figure 1).
Figure 1. A case of a 65-year-old male with GBM. Pre-operative MRI (A) shows a ring-enhanced tumour in the right temporal lobe. The tumour was gross totally removed followed by external beam radiotherapy (40 Gy), and SRS was performed for targeting the tumour bed surface wall (tumour volume: 50.1 cm3, margin: 3 mm, central dose: 25 Gy, marginal dose: 18 Gy, no. of isocentres: 3, no. of arcs/isocentre: 6) (B).
Nineteen patients were male, and 10 were female, with ages ranging from 29 to 80 years (mean age, 60.6 years), with pre-operative KPS scores ranging from 50 to 100 (mean KPS score, 74.0). Tumours were located in the frontal lobe in 14, the temporal lobe in 8, the parietal lobe in 2, the occipital lobe in 1 and the midline structures/bilateral hemispheres in 4 patients. The number of patients with a tumour size of 5 cm or less was 18, and that with a tumour size of more than 5 cm was 11. Twenty died of tumour-related causes before April 2004.
Sixty patients were those who were not fit for the criteria for IORT or did not consent to the IORT clinical trial between January 1989 and August 1998, who of all treated between September 1998 and August 2000, and who did not consent to the SRS clinical trial between September 2000 and March 2004.
Forty-four patients were male, and 16 were female, with ages ranging from 27 to 85 years (mean age, 60.3 years), with pre-operative KPS scores ranging from 40 to 100 (mean KPS score, 67.5). Tumours were located in the frontal lobe in 16, the temporal lobe in 21, the parietal lobe in 8, the occipital lobe in 4 and the midline structures/bilateral hemispheres in 11 patients. The number of patients with a tumour size of 5 cm or less was 32, and that with a tumour size of more than 5 cm was 28. Fifty-eight died of tumour-related causes before April 2004.
To determine the relevant prognostic factors, the following clinical parameters were analysed in all cases: patient age (<40 years, 40–59 years, or ≥60 years), sex (male or female), pre-operative KPS score (0–60 or 70–100), lesioned lobe (frontal, temporal, parietal, occipital or midline/bilateral), tumor size (≤5 cm or >5 cm), eloquence of the adjacent brain (eloquent area or non-eloquent area), extent of surgery (gross total resection or less than gross total), post-operative adjuvant therapy (AR, IAR or R&C/C), and type of local irradiation boost (IORT, SRS or none). The sensorimotor cortex, language cortex, internal capsule, thalamus, corpus callosum, fornix, hypothalamus and brainstem were regarded as eloquent areas.
The Chi-square test was used to evaluate the significance of the differences in clinical features between IORT and SRS groups and the incidence of local tumour recurrence/progression at the primary site during the first 9 months after the initial surgery among patients who underwent IORT, SRS or neither. Survival was analysed using the method described by Kaplan and Meier,Reference Kaplan and Meier22 and the significance of differences among the survival curves for each parameter was determined using the log-rank test. A multivariate analysis was performed with the Cox proportional hazard regression model.Reference Cox23 All statistical analyses were performed with commercially available software (StatView 4.0 for Macintosh computers; Abacus Concepts, Berkeley, CA). Significance was assigned at a probability value of less than 0.05.
RESULTS
There were no significant differences in patient characteristics among the three groups, except for the extent of surgery and adjuvant therapy (Table 1).
Table 1. Clinical features of GBM patients in IORT, SRS and non-local boost groups
AR = external beam radiotherapy (EBRT) with ACNU; GBM = glioblastoma multiforme; IAR = EBRT with ACNU and interferon-β; IORT = intra-operative radiotherapy; KPS = Karnofsky performance scale; R&C/C = EBRT with low-dose cisplatin or carboplatin; SRS = stereotactic radiosurgery.
*Significance set at p < 0.05 (Chi-square test).
The 1-year survival rate, 2-year survival rate and MST of patients in the IORT group were 67.7%, 12.9% and 15 months, respectively; those of patients in the SRS group were 76.7%, 25.2% and 19 months, respectively; those of patients in the non-local boost group were 40.0%, 10.6% and 11 months, respectively (Figure 2).
Figure 2. Graph of Kaplan–Meier survival curves of 120 patients with newly diagnosed GBMs, demonstrating a comparison among those of IORT (n = 31), SRS (n = 29) and non-local boost (n = 60) groups.
In univariate analysis, KPS score (p = 0.0097), extent of surgery (p = 0.0003) and local irradiation boost (p = 0.0099) were significantly predictive of survival. The multivariate analysis showed that gross total resection (p = 0.0044) and SRS (p = 0.005) were independent positive prognostic factors for survival (Table 2).
Table 2. Survival analysis of clinical parameters in 120 patients with GBM
AR = external beam radiotherapy (EBRT) with ACNU; GBM = glioblastoma multiforme; IAR = EBRT with ACNU and interferon-β; IORT = intra-operative radiotherapy; KPS = Karnofsky performance scale; R&C/C = EBRT with low-dose cisplatin or carboplatin; SLRS = stereotactic linac radiosurgery.
*Log-rank test.
†Cox proportional hazard regression analysis.
‡Significance set at p < 0.05.
Local tumour recurrence/progression at the primary site during the first 9 months after the initial surgery was observed on MR imaging in 24 (77.4%) of 31 patients in the IORT group, 13 (50.0%) of 26 patients in the SRS group (excluding three patients whose follow-up period could not reach 9 months after the initial surgery, in spite of no local tumour recurrence/progression at the primary site) and 52 (86.7%) of 60 patients in the non-local boost group. There were significant differences in the local tumour recurrence/progression rates among the three groups (p = 0.0012) (Table 3).
Table 3. Local tumour recurrence/progression at the primary site up to 9 months after initial surgery in patients with GBM
GBM=glioblastoma multiforme; IORT=intra-operative radiotherapy; SLRS=stereotactic linac radiosurgery.
†There were significant differences in local tumour recurrence/progression rates at the primary site up to 9 months after the initial surgery among the three groups (p = 0.0012, Chi-square test).
‡Three patients whose follow-up period could not reach 9 months after the initial surgery, in spite of no local tumour recurrence/progression in the primary site, were excluded from all the 29 patients in the SLRS group.
Thirty patients died of local tumour recurrence/progression at the primary site before April 2004. In 24 patients, tumour recurrence/progression emerged during the 9 months after the initial surgery. The remaining one patient is living without any tumour recurrence/progression at both primary and remote sites as of April 2004.
Of the 20 patients who died before April 2004, 15 died of local tumour recurrence/progression at the primary site. In 11 of the 15 patients, tumour recurrence/progression emerged until 9 months after the initial surgery. The remaining five patients died of remote tumour progression without local tumour recurrence/progression at the primary site.
Of the nine living patients, four were suffering from local tumour recurrence/progression, in which the recurrence/progression occurred until 9 months after the initial surgery in two and after 9 months after the initial surgery in the other. The remaining five patients, including three patients whose follow-up period could not reach 9 months after the initial surgery, have no tumour recurrence/progression at both the primary and remote sites as of April 2004.
DISCUSSION
SRS is a sophisticated and minimally invasive therapeutic modality of a precisely planned local irradiation boost for the treatment of brain tumours, and it has been incorporated into the initial management approach for GBM with the widespread use of the stereotactic radiosurgical system. To the best of our knowledge, as for the clinical results of newly diagnosed GBM treated using SRS incorporated into the initial management approach, ten reports of case seriesReference Buatti, Friedman, Bova and Mendenhall24–Reference Shrieve, Alexander, Black, Wen, Fine, Kooy and Loeffler32 and two case–control studiesReference Nwokedi, DiBiase, Jabbour, Herman, Amin and Chin33, Reference Prisco, Weltman, de Hanriot and Brandt34 have been published since the first clinical trial report published in 1992 by Loeffler et al.Reference Loeffler, Alexander, Shea, Wen, Fine, Kooy and Black19
The largest clinical study using SRS (the study of the Joint Center for Radiation Therapy, Boston, MA) was published in 1999 by Shrieve et al.Reference Shrieve, Alexander, Black, Wen, Fine, Kooy and Loeffler32 They treated 78 post-operative patients using SRS after conventional EBRT, and reported an MST of 19.9 months and a 1-year survival rate of 88.5%. These were encouraging results, although rather favourable patients were selected under the criteria of tumour size of 4 cm or less and KPS score of 70 or more.
The first clinical trial using GK was reported in 1996 by Larson et al.Reference Larson, Gutin and McDermott27 They divided newly diagnosed GBM patients into two groups. One group included patients with unifocal tumours, tumours located in the supratentorial subcortical hemisphere, age less than 70 years, KPS score of more than 60 and tumour maximum diameter less than 5 cm, and the other group included patients without at least one of the above criteria. They reported MSTs of 86 weeks in the former group and 40 weeks in the latter. The largest clinical study using GK was that by Kondziolka et al.,Reference Kondziolka, Flickinger, Bissonette, Bozik and Lunsford28 in 1997. They enrolled 64 GBM patients, reported a promising MST of 26 months, and concluded that SRS was safe and that its effectiveness as an adjuvant therapy deserved a properly stratified randomised trial.
In the first case–control study of patients with newly diagnosed GBM, Nwokedi et al. Reference Nwokedi, DiBiase, Jabbour, Herman, Amin and Chin33 showed a significantly longer MST of 25 months in the group of 31 patients who received both EBRT and GK, compared with that of 13 months in the group of 33 patients who received EBRT alone. Also, in another case–control study recently reported by Prisco et al.,Reference Prisco, Weltman, de Hanriot and Brandt34 the group of 15 patients who received EBRT and SRS showed a longer MST of 21.4 months compared with that of 11.6 months in the group of patients who received EBRT alone. In these two clinical studies, a local irradiation boost using SRS was reported as a significant, positive prognostic factor for survival also in multivariate analysis.Reference Nwokedi, DiBiase, Jabbour, Herman, Amin and Chin33, Reference Prisco, Weltman, de Hanriot and Brandt34
To summarise these previous clinical results, the MST was 9.5–26 months, and the 1-year survival rate was 19–88.5% for patients with newly diagnosed GBM treated with SRS used after post-operative EBRT, although there was some variation in patient selection and instruments used for SRS.Reference Buatti, Friedman, Bova and Mendenhall24–Reference Prisco, Weltman, de Hanriot and Brandt34 These results are better than those reported in GBM patients who underwent EBRT alone, and are approximately equal to those in GBM patients who underwent EBRT with IORT or IBT as a local irradiation boost.Reference Sakai, Yamada, Andoh, Hirata, Nishimura, Miwa, Shimizu and Yanagawa11–Reference Selker, Shapiro and Burger18
IORT is a therapeutic modality that can deliver a large, single irradiation dose with a high-energy electron beam during the operation. An advantage of IORT is that it improves the therapeutic effect by the direct application of a large single dose to the surgically exposed tumour, while excluding most normal tissue from the path of the radiation beam and producing greater local tumour control. The study by Matsutani et al. Reference Matsutani, Nakamura, Nagashima, Asai, Fujimaki, Tanaka, Nakamura, Ueki, Tanaka and Matsuda12 showed that the MST of 30 patients with newly diagnosed GBM treated with EBRT and IORT was 118 weeks. Sakai et al. Reference Sakai, Yamada, Andoh, Hirata, Nishimura, Miwa, Shimizu and Yanagawa11 and Fujiwara et al. Reference Fujiwara, Honma, Ogawa, Irie, Kuyama, Nagao, Takashima, Hosokawa, Ohkawa and Tanabe14 described significantly longer survival in a group of MG patients who underwent EBRT with IORT, compared with a control group of MG patients who underwent EBRT alone. Although these results represent a clinical advantage of IORT in the initial management of MG, the level of evidence for a clinical effect of IORT as a therapeutic modality is not high enough, because a multivariate statistical analysis was not done in those reports.
SRS is also a local irradiation boost, and is superior to IORT from the following points of view: SRS is less invasive than IORT in terms of the lack of need for a craniotomy. The target for irradiation can be precisely located using a computerised stereotactic planning system on visualised three-dimensional CT/MR imaging in SRS. Also, the anti-tumour effect of SRS can be easily and precisely evaluated on CT/MR imaging. IORT was used for the treatment of GBM since 1982 in our department, but since 2000 SRS has been incorporated into the initial management approach for this tumour. This is the first report of clinical results of GBM treatment using SRS compared with those using IORT in a single institute.
In this study, the survival of patients in the SRS group was significantly prolonged compared with that of those in the IORT and non-local boost groups, and SRS was revealed to be a significant prognostic factor for survival both in univariate and multivariate analyses, as was gross total resection of tumour. The local tumour progression/recurrence rate during the 9 months after initial surgery of patients in the IORT or SRS groups was significantly lower than that in the non-local boost group. This may be the reason why a local high-dose irradiation boost led to the prolongation of survival in GBM patients in this series. In addition, the local tumour progression/recurrence rate during the first 9 months after initial surgery in the SRS group was superior to that in the IORT group. This superiority may be due to the meticulousness and precision of therapeutic planning for the irradiation boost in SRS. This is a retrospective review and, therefore, may be subject to potential selection bias. Although the IORT and SRS were treated in prospective trials, the sequential nature of the studies may make direct comparisons difficult. Actually, there were significant differences in the extent of surgery and adjuvant therapy among the IORT, SRS and non-local boost groups. However, SRS was revealed to be a significant, independent positive prognostic factor for survival in a multivariate analysis. Therefore, the superiority of SRS for survival is unlikely to be dependent on the differences of these factors in this study.
Planning of the irradiation boost dose and field should be performed according to meticulous therapeutic strategies. In most reports regarding SRS for GBM, the marginal irradiation dose ranged from 15 to 20 Gy.Reference Buatti, Friedman, Bova and Mendenhall24–Reference Prisco, Weltman, de Hanriot and Brandt34 Generally, 20 Gy of SRS can be estimated to convert to 40–50 Gy of EBRT. Therefore, 20 Gy or less is assumed to be an appropriate dose, considering the effect on tumour cells and the periphery, which have already received about 50 Gy of EBRT before SRS.Reference Shinoda, Yano, Ando, Ohe, Sakai, Saio and Shimokawa35 The targeted field of SRS has been designed on CT/MR imaging. The fact that most recurrences of MG were within 2 cm outside of the contrast-enhancing lesion indicates the significance of local tumour control of the primary site for the management of MG. According to a recent methionine positron emission tomography (PET) study, GBM cells mostly invade over the Gd-enhancing area on T1-weighted MR imaging and extend more widely with increasing tumour size.Reference Miwa, Shinoda, Yano, Okumura, Iwama, Nakashima and Sakai36 There may be a significant correlation between tumour cell invasion and the Gd contrast-enhancing area, as follows: y = 0.535x – 8.865, x = maximum tumour diameter (mm), x > 16.6, y = 98% tumour cell invasion distance beyond the Gd contrast-enhancing area (mm). Target planning of SRS may be performed considering the facts of such PET studies hereafter.
CONCLUSION
The local irradiation boost, such as IORT and SRS, led to clearly better results on the survival of GBM patients. Especially, SRS is less invasive, allows for meticulous target planning of the irradiation boost and permits precise evaluation of the anti-tumour effect on CT/MR imaging. In addition, SRS was superior to IORT in terms of survival prolongation as well as suppression of local tumour recurrence/progression at the primary site in this series, and was a significant, positive prognostic factor for survival as well as gross total resection of the tumour in both univariate and multivariate analyses in this study. Gross total resection followed by EBRT and SRS may be a rational treatment to achieve the best clinical result in GBM.
