Cerebral radiation necrosis: Incidence, outcomes, and risk factors with emphasis on radiation parameters and chemotherapy

Int. J. Radiation Oncology Biol. Phys., Vol. 65, No. 2, pp. 499 –508, 2006 Copyright © 2006 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/06/$–see front matter

doi:10.1016/j.ijrobp.2005.12.002

CLINICAL INVESTIGATION

Brain

CEREBRAL RADIATION NECROSIS: INCIDENCE, OUTCOMES, AND RISK FACTORS WITH EMPHASIS ON RADIATION PARAMETERS AND CHEMOTHERAPY JEREMY D. RUBEN, F.C.RAD.ONC. (S.A.),* MICHAEL DALLY, F.R.A.N.Z.C.R.,* MICHAEL BAILEY, PH.D.,† ROBIN SMITH, M.SC.,* CATRIONA A. MCLEAN, F.R.C.P.A.,‡ § AND PASQUAL FEDELE *William Buckland Radiotherapy Centre, Melbourne, Australia; †Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia; ‡Department of Anatomical Pathology, The Alfred Hospital, Melbourne, Australia; § Monash University, Melbourne, Australia Purpose: To investigate radiation necrosis in patients treated for glioma in terms of incidence, outcomes, predictive and prognostic factors. Methods and Materials: Records were reviewed for 426 patients followed up until death or for at least 3 years. Logistic regression analysis was performed to identify predictive and prognostic factors. Multivariate survival analysis was conducted using Cox proportional hazards regression. Separate analyses were performed for the subset of 352 patients who received a biologically effective dose (BED) >85.5 Gy2 (>45 Gy/25 fractions) who were at highest risk for radionecrosis. Results: Twenty-one patients developed radionecrosis (4.9%). Actuarial incidence plateaued at 13.3% after 3 years. In the high-risk subset, radiation parameters confirmed as risk factors included total dose (p < 0.001), BED (p < 0.005), neuret (p < 0.001), fraction size (p ⴝ 0.028), and the product of total dose and fraction size (p ⴝ 0.001). No patient receiving a BED <96 Gy2 developed radionecrosis. Subsequent chemotherapy significantly increased the risk of cerebral necrosis (p ⴝ 0.001) even when adjusted for BED (odds ratio [OR], 5.8; 95% confidence interval [CI], 1.6 –20.3) or length of follow-up (OR, 5.4; 95% CI, 1.5–19.3). Concurrent use of valproate appeared to delay the onset of necrosis (p ⴝ 0.013). The development of radionecrosis did not affect survival (p ⴝ 0.09). Conclusions: Cerebral necrosis is unlikely at doses below 50 Gy in 25 fractions. The risk increases significantly with increasing radiation dose, fraction size, and the subsequent administration of chemotherapy. © 2006 Elsevier Inc. Radiation necrosis, Glioma, Chemotherapy, Radiation injury, Brain.

Malignant gliomas are the most common primary brain tumors (1, 2), and glioblastoma multiforme (GBM) comprises over half of these tumors (3). Treatment for fit patients comprises maximal safe resection followed by radiotherapy (RT) ⫾ concurrent and adjuvant chemotherapy. Studies by the Brain Tumor Study Group and the Scandinavian Glioblastoma Study Group in the late 1970s and 1980s demonstrated a 6-month gain in survival with the addition of radiotherapy to management protocols (4 – 8); however, even with treatment, outcomes are poor. The median survival for patients with GBM is 10 –12 months (9), whereas 2-year survival rates for GBM and anaplastic astrocytoma are only 9% and 44%, respectively (2).

Radiation doses in the region of 46 –50 Gy are as efficacious as higher doses in the treatment of low-grade glioma (10, 11), but doses of 60 Gy provide better outcomes for high-grade gliomas and are commonly prescribed for these tumors (12–14). The subsequent development of radionecrosis in some glioma patients is therefore not surprising. Cerebral radionecrosis is a potentially devastating complication of radiotherapy to the brain. It was first described by Fischer and Holfelder in 1930 following radiotherapy administered for a basal cell carcinoma of the scalp (15). Since then, numerous case reports (16 –18), case series (19 –22), and reviews (23, 24) have documented its occurrence after radiation therapy for a variety of intra-axial and

Reprint requests to: Michael Dally, F.R.A.N.Z.C.R., William Buckland Radiotherapy Centre, The Alfred Hospital, Commercial Rd., Prahran, Victoria, 3004, Australia. Tel: (⫹61) 3 9276-2337; Fax: (⫹61) 3 9276-2916; E-mail: [email protected] Jeremy Ruben was supported by The Peter Grant Hay Fund.

Pasqual Fedele was supported by a Monash University Scholarship. Acknowledgment—We wish to thank Dr. Alistair Hunter for advice on radiobiologic calculations. Received Sept 28, 2005, and in revised form Nov 30, 2005. Accepted for publication Dec 1, 2005.

INTRODUCTION

499

500

I. J. Radiation Oncology

● Biology ● Physics

extra-axial diseases over a range of doses and fractionation schedules. While thresholds of 54 Gy (21) and 57.6 Gy (25) have been reported for fraction sizes of 1.8 –2.0 Gy delivered once daily, it is apparent that radionecrosis may occur below these limits (10, 23). The incidence of radionecrosis after conventional radiotherapy for primary brain tumors in modern radiation oncology practice is largely unknown. Moreover, data are lacking on the actuarial risk of radiation necrosis in the general glioma population. This is mainly due to a lack of knowledge of the populations from which patients were accrued to various series, as well as a lack of data regarding the radiation doses and schedules employed (23). Additional difficulties relate to differentiating necrosis from tumor recurrence radiologically and to the low rates of reoperation and autopsy in these patients (26). Few authors have reported on this, and almost all these studies took place before the routine availability of modern anatomic and functional neuroimaging (10, 21, 26 –28). Incidences ranging from 3% up to 24% for patients treated on an aggressive experimental protocol have been reported (26, 29). Known risk factors for the development of radiation necrosis include total radiation dose; fraction size; treatment duration; and, at least in patients who receive radiosurgery, the irradiated volume (21, 23, 30, 31). These risks are difficult to quantify owing to deficiencies in their reporting in the literature, which is usually retrospective and often anecdotal. Furthermore, many studies date from an era when magnetic resonance imaging (MRI) and even computed tomographic scans were not widely available (19, 21, 23, 32), whereas others rely predominantly on abnormal radiology findings for diagnosis (33). The risks of radiosurgical or interstitial radiation boosts are, however, well appreciated (34 –37). Clinical experience also suggests that chemotherapy increases the risk of subsequent necrosis, but this has not been definitively demonstrated (23, 38, 39). The deleterious clinical consequences of radiation necrosis are well appreciated. It is a condition that is associated with considerable morbidity and mortality despite therapy (23, 40). However, findings of some studies provocatively suggest that outcomes in patients with glioblastoma who survive after the development of radionecrosis may actually be more favorable than in those who do not develop this complication (36, 41, 42). The purpose of this study was to investigate cerebral radiation necrosis in patients with glioma treated at our center in terms of incidence, outcomes, predictive and prognostic factors.

Volume 65, Number 2, 2006

of their treatment were obtained from the original treating institution. The coexistence of hypertension and diabetes was noted. All patients were followed until death, or for at least 3 years, the time required for 74% of radiation necrosis to manifest (43). Mean duration of follow-up was 16.7 months. Thirteen patients were excluded from the study because of insufficient length of follow-up in 7 patients, incomplete treatment details in 3 patients, and unknown outcomes in 3 patients. Data were therefore complete for 426 patients and were used for analysis. A separate analysis was performed on a reduced data set of 352 patients who received a biologically effective dose (BED) greater than 85.5 Gy2 (equivalent to 45 Gy in 25 fractions using an ␣/␤ ratio of 2 for brain tissue [44]). The results reported are for these 352 patients who were at highest risk for radionecrosis, except where otherwise stated.

Diagnosis of radiation necrosis Radiation necrosis was diagnosed in 21 patients. A pathologic diagnosis was made on examination of surgical resection specimens in 17 patients and a biopsy specimen in 1 patient. Central pathologic review by a neuropathologist (C.M.) was obtained for all these cases. In the remaining 3 patients, radionecrosis was diagnosed by combinations of magnetic resonance spectroscopy (2 patients), positron emission tomography (1 patient), and thallium single-photon emission computed tomography studies (2 patients). Three patients, who had histories and investigations in keeping with cerebral radionecrosis but in whom the diagnosis was never definitively made, were not considered as having radionecrosis for the purposes of this study. Details of patients who developed radionecrosis are found in Table 1.

Statistical analysis Statistical analysis was performed using SAS version 8.2. (SAS Institute Inc., Cary, NC). Univariate analysis using logistic regression was performed to identify factors predictive for the development of brain radionecrosis and for patient outcomes once necrosis had manifested. Although there was insufficient power for definitive multivariate analysis, there was sufficient power for pairs of predictive variables to be compared using multiple logistic regression. Results from the logistic regression are presented as odds ratios (OR) with a 95% confidence interval (CI). Where risk variables are continuous, the OR can be seen to represent the risk associated with a 1-unit increase in the risk variable. Variables investigated included patient-related factors, treatment-related factors, and tumor-related factors, which are listed in Table 2. A multivariate survival analysis was conducted using Cox proportional hazards regression with results presented as a hazards ratio with a 95% CI. The relationship between latency until necrosis and all other variables was determined using linear regression. A two-sided p value of 0.05 was considered to be statistically significant.

Patient details METHODS AND MATERIALS Study population The records of patients who presented to our department between July 1, 1993, and November 31, 2002, were reviewed. A total of 439 patients with glioma who received all or part of their radiation therapy at our center were identified. Ten patients had received previous radiotherapy at another center; complete details

The mean age of patients at the time of commencement of radiotherapy was 57 years. The ratio of men to women was 1.7:1. The pathologic diagnosis was GBM in 317 patients, anaplastic astrocytoma in 88 patients, astrocytoma World Health Organization (WHO) Grade 2 in 3 patients, oligodendroglioma WHO Grade 2 in 8 patients, and WHO Grade 3 in 10 patients. Hypertension was present in 85 patients (20%), and 30 (7%) were diabetic. Ten patients were both hypertensive and diabetic; 357 (84%) patients

Table 1. Patients who developed radionecrosis Histology

RT dose/fraction number

Subsequent stereotactic RT dose/fraction

Total RT time* (days)

Total dose (Gy)

Total BED (Gy2)

Total neuret

Dose ⫻ fraction size (Gy2)

Chemotherapy

Latency (months)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

32 38 62 63 40 55 56 60 44 45 41 55 37 40 44 62 45 62 54 39 53

GBM AA GBM AA GBM GBM GBM GBM GBM AA GBM GBM AA AA GBM GBM AA GBM GBM GBM GBM

60 Gy/30 fr 54 Gy/28 fr 45 Gy/15 fr 45 Gy/20 fr 60 Gy/30 fr 45 Gy/20 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 50 Gy/25 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 60 Gy/30 fr 50 Gy/25 fr

— 18 Gy in 3 fr (salvage) 18 Gy in 3 fr (salvage) — — 15 Gy in 3 fr (initial boost) — — 18 Gy in 3 fr (salvage) 18 Gy in 3 fr (initial boost) 18 Gy in 3 fr (initial boost) — — — 24 Gy in 4 fr (salvage) — 18 Gy in 3 fr (salvage) — 18 Gy in 3 fr (initial boost) — 18 Gy in 3 fr (initial boost)

45 41 26 33 44 35 44 48 43 43 49 40 47 43 47 48 48 46 50 44 39

60 72 63 45 60 60 60 60 78 68 78 60 60 60 84 60 78 60 78 60 68

120 178 185 96 120 148 120 120 192 172 192 120 120 120 216 120 192 120 192 120 172

1,069 2,041 2,172 976 1,071 1,822 1,071 1,065 2,100 2,011 2,069 1,077 1,066 1,072 2,272 1,065 2,092 1,068 2,067 1,071 2,018

120 212 243 101 120 176 120 120 228 208 228 120 120 120 264 120 228 120 228 120 208

BCNU PCBZ/TMZ PCV BCNU BCNU BCNU BCNU BCNU — BCNU BCNU — BCNU TMZ TMZ BCNU BCNU BCNU — BCNU TMZ

32.10 26.41 26.22 21.02 20.83 19.55 17.71 9.10 8.90 7.88 7.79 6.31 5.81 5.49 5.03 4.89 4.57 4.20 3.81 3.55 2.07

Abbreviations: AA ⫽ anaplastic astrocytoma; BCNU ⫽ Carmustine; CCNU ⫽ Lomustine; GBM ⫽ glioblastoma multiforme; PCBZ ⫽ procarbazine; PCV ⫽ procarbazine, CCNU, vincristine; TMZ ⫽ temozolomide; RT ⫽ radiotherapy; BED ⫽ biologically equivalent dose. * The sum of the number of days in each treatment course if more than one was delivered.

● J. D. RUBEN et al.

Age at RT

Cerebral radionecrosis: incidence, outcomes, and risk factors

Patient no.

501

502

I. J. Radiation Oncology

● Biology ● Physics

Volume 65, Number 2, 2006

Table 2. Univariate analysis of factors predictive for radionecrosis ⱖ85.5 Gy2 subset (n ⫽ 352) Odds ratio (95% confidence interval) Host factors Length of survival Diabetes Hypertension Age at radiotherapy Gender Concurrent medication Carbamazepine Phenytoin Valproate Dexamethasone Tumor factors Pathologic type Intracerebral location Pathologic grade (WHO) Therapy-related Radiation dose Total BED (Gy2) Total dose (Gy) Total neuret Fraction size Product of total dose ⫻ fraction size (Gy2) Overall treatment time (radiotherapy) Use of chemotherapy Extent of initial surgery

p value

Entire cohort (n ⫽ 426) Odds ratio (95% confidence interval)

p value

1.02 (1.01–1.03) 0.70 (0.09–5.48) 0.47 (0.11–2.05) 0.97 (0.94–1.00) 0.77 (0.31–1.88)

⬍0.001 0.74 0.31 0.053 0.56

1.02 (1.01–1.03) 0.65 (0.08–5) 0.40 (0.09–1.76) 0.96 (0.94–0.99) 0.76 (0.31–1.85)

⬍0.001 0.68 0.23 0.01* 0.54

1.7 (0.63–4.56) 1.79 (0.68–4.72) 0.64 (0.14–2.82) 2.93 (0.67–12.8)

0.29 0.24 0.55 0.15

2.05 (0.77–5.49) 1.88 (0.72–4.95) 0.73 (0.17–3.23) 2.52 (0.58–11.0)

0.15 0.2 0.68 0.22

0.98 (0.37–2.59) 0.69 (0.98–1.02) 1.07 (0.44–2.57)

0.42 0.08 0.89

0.83 (0.31–2.20) 2.18 (0.90–5.28) 0.93 (0.39–2.22)

0.38 0.08 0.86

1.01 (1.00–1.02) 1.06 (1.02–1.09) 1.00 (1.00–1.00) 1.82 (1.07–3.09) 1.01 (1–1.02) 1.06 (1–1.11) 5.79 (1.67–20) 0.97 (0.58–1.63)

0.005 ⬍0.001 ⬍0.001 0.028 0.001 0.041† 0.002 0.91

1.01 (1.01–1.02) 1.06 (1.03–1.10) 1.00 (1.00–1.0) 1.67 (0.94–2.98) 1.01 (1.01–1.02) 1.07 (1.02–1.12) 7.85 (2.28–27.1) 0.87 (0.52–1.45)

⬍0.001 ⬍0.001 ⬍0.001 0.08 ⬍0.001 0.005† ⬍0.001 0.59

Abbreviations: BED ⫽ biologically effective dose; WHO ⫽ World Health Organization. * Insignificant when adjusted for BED or survival. † Insignificant when adjusted for BED.

were of an Eastern Cooperative Oncology Group (ECOG) performance status of ⱕ2 on presentation to our department.

Treatment details All patients received initial external-beam radiotherapy (EBRT). This was typically delivered using conformal three-dimensional planning with 2–3 beams. Margins of 2–3 cm were employed for malignant gliomas based on the initial contrast-enhancing volume on T1-weighted MRI, taking care to include edema evident on T2 imaging. The median dose delivered was 50 Gy (range, 16 – 60), and the median fraction size used was 2 Gy (range, 1.6 –3.75). A subsequent course of radiation was administered to 56 patients to a median dose of 18 Gy (range, 10 –50 Gy). This took the form of stereotactic radiotherapy in 51 patients (12%), conformal EBRT in 2 (0.5%), and interstitial brachytherapy in 3 (0.7%). Formulae used in the BED calculations for low-dose-rate brachytherapy are detailed in the Appendix. A total of 194 patients received chemotherapy as part of their management. Chemotherapy regimens used included carmustine (BCNU), procarbazine or temozolomide monotherapy, and combination chemotherapy with procarbazine/lomustine (CCNU)/vincristine (PCV).

Follow-up All patients received regular follow-up after treatment by the treating radiation oncologist and a neuro-oncologist. Follow-up

included routine 3– 6 monthly clinic visits and periodic MRI evaluation.

RESULTS Incidence Cerebral radionecrosis was diagnosed in 21 patients, an incidence of 4.9%. When the analysis was restricted to patients who received a BED greater than 85.5 Gy2, which renders the population comparable to the studies of Marks et al. (21) and Soffietti et al. (28), the prevalence of radionecrosis increased to 6%. As expected, the risk of radionecrosis increased with increasing survival after radiotherapy (p ⬍ 0.001). The actuarial incidence for the entire cohort was 2.9%, 5.1%, 9.3%, and 13.3% at 6, 12, 24, and 36 months, respectively (Fig. 1). Interval to development of necrosis The shortest latent period from completion of radiotherapy to the diagnosis of necrosis was 2.1 months, and the mean latent interval was 11.6 months (range, 2–32 months). Sixty-six percent of cases had manifested by 1 year and 85% by 2 years. On linear regression analysis neither total dose in Gy, BED in Gy2, nor neuret were predictive for latency to necrosis (neuret is defined in Appendix 2). The

Cerebral radionecrosis: incidence, outcomes, and risk factors

● J. D. RUBEN et al.

503

Fig. 1. Kaplan-Meier estimate of time to occurrence of radiation necrosis. RT ⫽ radiotherapy.

Factors predictive for the development of radionecrosis Effect of radiation dose, fractionation, and time. On logistic regression analysis, total BED in Gy2, total dose in Gy, and neuret were all highly significant risk factors for radionecrosis (p ⫽ 0.005, p ⬍ 0.001, and p ⬍ 0.001, respectively). No patient who received a BED less than 96 Gy2 manifested radionecrosis. Only 1 of 154 patients who received a BED of between 85.5–116 Gy2 (45 Gy/25 fractions– 61.2 Gy/34 fractions) developed radiation necrosis. This patient received 45 Gy in 20 fractions over 32 days. A total of 155 patients received 60 Gy in 30 fractions and radionecrosis developed in 10 (6.5%). Nine of the 10 also received chemotherapy. In contrast, radionecrosis developed in 5 of 26 patients (19%) who received subsequent stereotactic radiotherapy (median BED 72 Gy2, range 48 –96 Gy2) following the identical regimen. The risks of radiation necrosis across various dose ranges are illustrated in Fig. 2. The product of total dose ⫻ fraction size in Gy2, identified by Lee et al. (20) as a predictive variable for radionecrosis, was confirmed as such by our study (p ⬍ 0.001). Average fraction size also reached statistical significance (p ⫽ 0.028). This was calculated using the proportional contribution to the total BED of each course of radiotherapy if more than one was delivered using different fraction sizes. Overall treatment time (the sum of the number of days in each radiotherapy course if more than one was delivered) was significant on logistic regression analysis (p ⫽ 0.041), but not after adjusting for BED in Gy2 (p ⫽ 0.067). Overall duration of treatment (time period from the

beginning of the first course of radiation to the end of the last course if more than one was administered) was not related to necrosis risk (p ⫽ 0.18). Effect of chemotherapy and concurrent medications. The use of adjuvant chemotherapy was strongly related to subsequent development of necrosis (p ⫽ 0.006 for the reduced data set, and p ⫽ 0.001 for the whole group). This significance was maintained after adjusting for BED (OR, 4.9; 95% CI, 1.4 –17.1, p ⫽ 0.01) or length of follow-up (OR, 4.4; 95% CI, 1.2–15.5, p ⫽ 0.02). Apart from valproate, neither the use of steroids nor anticonvulsants concurrent with radiotherapy had any influence on subsequent radionecrosis (Table 2). The numbers of patients using phenobarbital were, however, too small to detect a meaningful impact even if one were to exist. Extent of initial surgery. Neither complete macroscopic resection (n ⫽ 181) nor tumor debulking (n ⫽ 103) predisposed toward radionecrosis compared with stereotactic biopsy alone (n ⫽ 142) despite some suggestion from the n=23

% developing necrosis

latent interval was significantly lengthened for patients using valproate concurrently with radiotherapy (p ⫽ 0.013), but not for those using other anticonvulsants. Dexamethasone use had no significant effect.

25%

n=23

20%

17%

n=11

15% 10% 5%

22%

n=291

9%

4%

n=4 0%

0%

BED in Gy2 Equivalent in 2Gy fractions

>85-120Gy2 >120-155Gy2 >155-190Gy2 >190-225Gy2

44-60Gy

62-76Gy

78-94Gy

96-112Gy

>225Gy2

>112Gy

Fig. 2. Incidence of radiation necrosis across a range of radiation doses. BED ⫽ biologically equivalent dose.

504

I. J. Radiation Oncology

● Biology ● Physics

literature that surgical trauma might predispose toward later necrosis (22, 23) (p ⫽ 0.9). Constitutional factors. Neither hypertension nor diabetes were associated with the development of cerebral radionecrosis (p ⫽ 0.3 and p ⫽ 0.7, respectively). Similarly, patient age at treatment did not reach significance on logistic regression analysis (p ⫽ 0.526) even when adjusted for radiation dose delivered (p ⫽ 0.27) or survival (p ⫽ 0.8). Necrosis occurred equally readily in men as in women. Tumor-related factors. With respect to radionecrosis risk, no differences were detected between the different histologic types of glioma or between different anatomic tumor locations within the brain. Survival Median survival for patients with GBM, anaplastic astrocytoma, anaplastic oligodendroglioma, WHO Grade 2 oligodendroglioma, and WHO Grade 2 glioma was 10.3 months, 24.6 months, 59.2 months, 56 months, and 188 months, respectively. On multivariate analysis, radionecrosis had no effect on overall survival, hazards ratio ⫽ 0.72 (95% CI, 0.42–1.21). In patients with radionecrosis, the presence of persistent tumor was a risk factor for death from disease (p ⫽ 0.048) on univariate analysis but did not predict for length of survival (p ⫽ 0.31). DISCUSSION Most previous studies reporting on the incidence of radiation necrosis took place before the routine availability of MRI imaging, and before the utility of positron emission tomography scanning and thallium scintigraphy was appreciated for the diagnosis of this entity. A 5% incidence of radiation necrosis was reported by Marks et al. in 139 patients who received 4,500 rad or more for the treatment of primary brain tumors (21). Mikhael reported an incidence of 3.3% in patients irradiated for glioma and found that the radionecrosis coincided with the volumes of brain receiving 4,500 rad or more on dose reconstruction maps (26). Of 107 patients in a series published by Soffietti et al., cerebral radionecrosis was noted in 3 of the 30 who underwent autopsy (2.8% overall) (28). However, that series included patients who received concurrent misonidazole, as well as patients who received hyperfractionated radiotherapy— treatment options no longer pursued for glioma. The North Central Cancer Treatment Group/Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group intergroup trial of high- vs. low-dose radiotherapy for low-grade glioma is noteworthy because it was conducted prospectively and took place in the era of modern neuroimaging (10). That study reported on Grades 3–5 radionecrosis and found a crude incidence of 6% and 1% in the 64.8-Gy and 50.4-Gy arms, respectively. Possibly the highest reported incidence of “therapy-induced necrosis” of the brain is 24% in patients on an experimental protocol who received hyperfractionated radiotherapy with concurrent carboplatin, followed by adjuvant PCV chemotherapy (29). In contrast, an incidence of

Volume 65, Number 2, 2006

just 0.2% (45) and 0% (46) has been estimated after radiotherapy for pituitary adenomas. The development of radiation necrosis was documented in 4.9% of patients in our series. This rose to 6% if only the 352 patients who received a BED ⱖ85.5 Gy2 were considered; a prevalence comparable to previously reported figures (10, 21, 26 –28). This is despite the subsequent delivery of hypofractionated-stereotactic or interstitial irradiation in 13% of our patients after radical radiotherapy, the risks of which are well appreciated. In one series, reoperation was required in half of the patients who received a radiosurgical boost, and isolated radiation necrosis accounted for half of these cases (34). In studies reporting brachytherapy results, 40 –50% of patients who received interstitial radiotherapy by permanent implantation of 125I seeds manifested radionecrosis (35–37). In keeping with these findings, radionecrosis developed in 18% of patients in our series who received a subsequent course of stereotactic radiation. The comparable incidence rates between our study and older ones, despite the risks associated with additional radiation courses in 13% of our patients, might be ascribed to modern conformal radiotherapy planning and the avoidance of a contralateral beam unless tumor extends to the corpus callosum. It may also reflect a volume effect, as earlier studies used mostly whole-brain radiation fields to treat malignant gliomas, whereas modern practice employs margins around the enhancing tumor and surrounding edema (21, 28). We believe that our result of 5– 6% accurately portrays the prevalence of cerebral radionecrosis in patients treated for glioma before the recent adoption of concurrent and adjuvant temozolomide chemotherapy as standard of care in GBM (47). All patients received regular clinical and MRI follow-up, and only 3 of the 34 patients in whom the possibility of radionecrosis was considered significant did not have this diagnosis definitively proven or disproven. Several case reports and small series have reported that chemotherapy alone may produce changes mimicking radiation damage, including cerebral necrosis (48 –51). The reports of Burger et al. (38), among others, highlighted a possible additive necrosis risk with the addition of chemotherapy but lacked either the required patient numbers or data about the original study population and radionecrosis incidence to provide confirmation (23, 52, 53). It has therefore long been supposed that chemotherapy may have an additive effect on the development of cerebral necrosis in the setting of radiotherapy. This clinical suspicion is confirmed by our study (Table 2). The administration of chemotherapy produced a greater than fourfold increase in risk for cerebral necrosis even after adjusting for BED or survival. With the recent adoption of concurrent and adjuvant temozolomide as a standard of care in the treatment of GBM, as well as the potential for longer survival in sotreated patients, a modest increase in the incidence of radionecrosis may be expected in the future. On the other hand, 23 GBM patients in our series received stereotactic boosts as part of their initial management, but the use of a

Cerebral radionecrosis: incidence, outcomes, and risk factors

radiosurgical boost is likely to decrease substantially following the negative results of Radiation Therapy Oncology Group 93-05, although it remains useful as salvage therapy (54). Therefore, the incidence of radiation necrosis in glioma patients treated in the current era awaits further study. The total dose of radiation appears to be the most important risk factor for subsequent necrosis in our study and others. In the North Central Cancer Treatment Group/Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group trial, patients who received 60.8 Gy had a significantly higher actuarial incidence of necrosis than those who received 50.4 Gy (10). Only 1 patient developed radionecrosis at a dose of 50.4 Gy in 28 fractions in that study. In the report by Marks et al., only 1 patient developed necrosis in an area radiated to less than 50 Gy in 27 fractions (21). Of the 80 radionecrosis cases selected from the literature by Sheline et al., only 20 had received doses ⱕ50 Gy (23). Of these, 17 (81%) received fraction sizes larger than 2.5 Gy. In the present study, we did not observe radionecrosis at a BED lower than 96 Gy2. Only 1 affected patient received a total dose lower than 50 Gy, but she received fractions of 2.25 Gy (Patient 4, Table 1). Radiation dose in Gy, BED in Gy2, and neuret were all highly significant for necrosis on logistic regression analysis. BED is based on linear-quadratic formalism and places emphasis on the ␣/␤ ratio of tissue, whereas the neuret utilizes a power function and also places some emphasis on time factors. Despite the limitations of a power function to model radiation effect and the brain’s limited capacity for repopulation (which would diminish any impact of a time factor), both the neuret and the BED in Gy2 were strong predictors for radionecrosis from our data. There are limited data to support the possible role of a time factor in the radiobiology of radionecrosis. Lee et al. found that overall treatment time (OTT) is a risk factor for subsequent temporal lobe radionecrosis after radiotherapy for nasopharyngeal cancer (31). In that study, OTT was shown to be significant on a multivariate analysis including 24 patients with radionecrosis. Interestingly, a previous similar study by the same authors showed no influence of OTT on necrosis risk (20). Likewise, no influence of OTT on necrosis risk was demonstrated in the study by Marks et al. (21) or in the present study after adjusting for radiation dose. However, in both the present study and that of Marks et al., patients received fairly uniform fractionation schedules and so treatment duration was reasonably invariable relative to dose. The narrow range of this parameter coupled with relatively small numbers of events hamper the detection of a time effect should one actually exist. Classic radiobiologic teaching holds that time factors are less important in the reaction of late-responding tissues to radiotherapy, given their limited capacity for repopulation. Although the vast majority of laboratory (44, 55) and clinical data (56 –59) on the topic support this view, some animal studies (60, 61), an analysis of the British Institute of Radiology trials (62), and the European Organization for Research and Treatment of Cancer 22851 trial (63) cast

● J. D. RUBEN et al.

505

some doubt on its absoluteness. It appears that accelerated radiotherapy in particular carries a high risk of cerebral radionecrosis (31, 63, 64). The influence of fraction size has been assessed by Lee et al. using the product of total dose and fraction size (Dd) in Gy2. Dd was found to be the most significant predictive factor for necrosis in that study (20). Our study confirms the significance of Dd and also demonstrates that average fraction size is a significant independent risk factor. Necrosis usually manifests after a latency period of many months, but has been reported as early as 3 months (40), and as late as 47 years after radiotherapy (65). The shortest latency period observed in our study was 2.1 months in a patient who received a stereotactic boost after EBRT (cumulative BED 172 Gy2). An analysis of BED vs. interval to necrosis did not reveal any relationship between increasing dose delivered and a shorter latent interval. Sheline et al. were similarly unable to show such a relationship in their analysis of 80 cases using the neuret as a measure of biologic dose (23). The mean interval from the end of radiotherapy until onset of necrosis of 11.6 months in our study is comparable to those described by others in similar patient populations (21, 28). It is interesting that according to our data and others (21, 23, 28), the latency period for radionecrosis appears approximately 5 times shorter in patients with glioma than in other patients—notably patients treated for nasopharyngeal carcinoma (31, 66). This may support the theory that the brain parenchyma immediately surrounding a glioma is especially sensitive to chemotherapy and radiation, as has been suggested by Burger et al. (38). An alternative possibility is that much of the necrosis in this setting is tumor necrosis, whereas the same process in patients with nasopharyngeal tumors represents only the necrosis of late-responding normal brain tissue. Our observation that valproate lengthens the interval to necrosis is by no means definitive but is intriguing in the context of an animal study by Kulinskii and Klimova (67). In that study, valproate was found to be radioprotective by increasing tissue concentrations of ␥-aminobutyric acid. Other than valproate and chemotherapy agents, no other drugs were found to influence the development of radionecrosis. Although it has been suggested that steroid treatment during radiotherapy may offer protection from radiation damage (28, 38), patients with Cushing’s disease appear to be especially susceptible to radiation necrosis according to two reports (68, 69). The use of dexamethasone during radiotherapy had no influence on the development of cerebral necrosis in our patients. Although both diabetes and hypertension might be expected to augment the vasculopathic component of the radionecrotic process, neither of these variables was related to the development of radiation necrosis in our study. This is in keeping with the animal data of Hopewell and Wright, which showed that although hypertension accelerated vascular damage at lower radiation doses of 2,000 –3,000 rad, at doses ⱖ4,000 rad both normotensive and hypertensive rats exhibited the same degree of necrosis (70).

506

I. J. Radiation Oncology

● Biology ● Physics

It is apparent from this study that viable tumor often coexists within areas of radiation necrosis; 10 of the 18 pathologically examined cases (55%). This is intuitive, because gliomas usually recur within 2 cm of the primary site, and radionecrosis, when it develops, is most likely to do so within this same region because it receives the highest dose. Because of this, although stereotactic biopsy has been reported as a useful diagnostic modality in this setting (42), we would caution that unless guided by a positron emission tomography scan or other functional imaging modality, diagnostic biopsy alone may be misleading (71), nor would biopsy relieve mass effect. Management of radiation necrosis is predominantly surgical in the face of raised intracranial pressure or if symptoms require rapid control or progress on conservative management (72). However, there is ample evidence that surgery is not always necessary, and resolution may be obtained after corticosteroid therapy alone in some cases (40, 73–76). Four of the 21 patients in our series were managed conservatively with eventual resolution of the necrotic process. No survival benefit was noted for maximal resection of necrotic debris vs. conservative management. On multivariate survival analysis of our data, radionecrosis had no effect on survival, although a report by Floyd et al. suggested a survival advantage for those patients on a hypofractionated intensity-modulated radiotherapy protocol who subsequently developed radionecrosis (41). Interest-

Volume 65, Number 2, 2006

ingly, in our series, 2 patients who showed no evidence of persistent tumor at the time of resection of their necrotic foci are still alive with no evidence of disease 6 years later. A third patient with necrosis diagnosed by magnetic resonance spectroscopy and whose MRI abnormalities subsequently resolved with conservative management is similarly alive with no evidence of disease 5 years later (Patients 1, 10, and 4, Table 1). CONCLUSIONS The most significant risk factor for radiation necrosis is the total dose of radiotherapy delivered, although fraction size is also important (20, 23). Our study has shown that chemotherapy after radiation significantly increases the risk of cerebral necrosis by approximately fivefold. The recent adoption of concurrent and adjuvant temozolomide chemotherapy may therefore be expected to increase the incidence of radionecrosis in the future; this should be borne in mind when assessing patients for possible disease progression. While the present study is the largest so far published that allows for calculation of incidence rates, the small number of patients with radionecrosis (twenty one) and the absence of data for irradiated volumes hinder the construction of a meaningful multivariate model of predictive and prognostic factors for radionecrosis. This problem is likely to be overcome only through the evaluation of data from cooperative group trials enrolling large numbers of patients.

REFERENCES 1. Laperriere N, Zuraw L, Cairncross G. Radiotherapy for newly diagnosed malignant glioma in adults: A systematic review. Radiother Oncol 2002;64:259 –273. 2. CBTRUS. Statistical report: Primary brain tumors in the United States, 1997–2001. Chicago, IL: Central Brain Tumor Registry of the United States; 2004 –2005. 3. Surawicz TS, McCarthy BJ, Kupelian V, et al. Descriptive epidemiology of primary brain and CNS tumors: Results from the Central Brain Tumor Registry of the United States, 1990 – 1994. Neuro-oncol 1999;1:14 –25. 4. Walker MD, Alexander E Jr., Hunt WE, et al. Evaluation of BCNU and/or radiotherapy in the treatment of anaplastic gliomas: A cooperative clinical trial. J Neurosurg 1978;49: 333–343. 5. Walker MD, Green SB, Byar DP, et al. Randomized comparisons of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. N Engl J Med 1980;303: 1323–1329. 6. Kristiansen K, Hagen S, Kollevold T, et al. Combined modality therapy of operated astrocytomas grade III and IV. Confirmation of the value of postoperative irradiation and lack of potentiation of bleomycin on survival time: A prospective multicenter trial of the Scandinavian Glioblastoma Study Group. Cancer 1981;47:649 – 652. 7. Sandberg-Wollheim M, Malmstrom P, Stromblad LG, et al. A randomized study of chemotherapy with procarbazine, vincristine, and lomustine with and without radiation therapy for astrocytoma grades 3 and/or 4. Cancer 1991;68:22–29. 8. Shapiro WR, Young DF. Treatment of malignant glioma. A controlled study of chemotherapy and irradiation. Arch Neurol 1976;33:494 –500.

9. Grossman SA, Batara JF. Current management of glioblastoma multiforme. Semin Oncol 2004;31:635– 644. 10. Shaw E, Arusell R, Scheithauer B, et al. Prospective randomized trial of low- versus high-dose radiation therapy in adults with supratentorial low-grade glioma: Initial report of a North Central Cancer Treatment Group/Radiation Therapy Oncology Group/Eastern Cooperative Oncology Group study. J Clin Oncol 2002;20:2267–2276. 11. Karim AB, Maat B, Hatlevoll R, et al. A randomized trial on dose-response in radiation therapy of low-grade cerebral glioma: European Organization for Research and Treatment of Cancer (EORTC) Study 22844. Int J Radiat Oncol Biol Phys 1996;36:549 –556. 12. Bleehen NM, Stenning SP. A Medical Research Council trial of two radiotherapy doses in the treatment of grades 3 and 4 astrocytoma. The Medical Research Council Brain Tumour Working Party. Br J Cancer 1991;64:769 –774. 13. Walker MD, Strike TA, Sheline GE. An analysis of doseeffect relationship in the radiotherapy of malignant gliomas. Int J Radiat Oncol Biol Phys 1979;5:1725–1731. 14. Coffey RJ, Lunsford LD, Taylor FH. Survival after stereotactic biopsy of malignant gliomas. Neurosurgery 1988;22:465– 473. 15. Fischer A, Holfelder H. Lokales amyloid in Gehirn [Local amyloid in the brain]. Dtsch Z Chir 1930;227:475– 483. 16. Bampoe J, Nag S, Leung P, et al. Brain necrosis after permanent low-activity iodine-125 implants: Case report and review of toxicity from focal radiation. Brain Tumor Pathol 2000;17: 139 –145. 17. Hoshi M, Hayashi T, Kagami H, et al. Late bilateral temporal lobe necrosis after conventional radiotherapy. Neurol Med Chir (Tokyo) 2003;43:213–216.

Cerebral radionecrosis: incidence, outcomes, and risk factors

18. Coghlan KM, Magennis P. Cerebral radionecrosis following the treatment of parotid tumors: A case report and review of the literature. Int J Oral Maxillofac Surg 1999;28:50–52. 19. Glass JP, Hwang TL, Leavens ME, et al. Cerebral radiation necrosis following treatment of extracranial malignancies. Cancer 1984;54:1966 –1972. 20. Lee AW, Foo W, Chappell R, et al. Effect of time, dose, and fractionation on temporal lobe necrosis following radiotherapy for nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 1998;40:35– 42. 21. Marks JE, Baglan RJ, Prassad SC, et al. Cerebral radionecrosis: Incidence and risk in relation to dose, time, fractionation and volume. Int J Radiat Oncol Biol Phys 1981;7:243–252. 22. Morris JG, Grattan-Smith P, Panegyres PK, et al. Delayed cerebral radiation necrosis. Q J Med 1994;87:119 –129. 23. Sheline GE, Wara WM, Smith V. Therapeutic irradiation and brain injury. Int J Radiat Oncol Biol Phys 1980;6:1215–1228. 24. Kramer S. The hazards of therapeutic irradiation of the central nervous system. Clin Neurosurg 1968;15:301–318. 25. Liebel S, Sheline G. Tolerance of the brain and spinal cord to conventional irradiation. In: Sheline GE, editor. Radiation injury to the nervous system. New York: Raven Press; 1991. p. 239 –256. 26. Mikhael MA. Dosimetric considerations in the diagnosis of radiation necrosis of the brain. In: Kagan A, editor. Radiation damage to the nervous system. A delayed therapeutic hazard. New York: Raven Press; 1980. p. 59 –91. 27. Marsa GW, Goffinet DR, Rubinstein LJ, et al. Megavoltage irradiation in the treatment of gliomas of the brain and spinal cord. Cancer 1975;36:1681–1689. 28. Soffietti R, Sciolla R, Giordana MT, et al. Delayed adverse effects after irradiation of gliomas: Clinicopathological analysis. J Neurooncol 1985;3:187–192. 29. Kumar AJ, Leeds NE, Fuller GN, et al. Malignant gliomas: MR imaging spectrum of radiation therapy- and chemotherapyinduced necrosis of the brain after treatment. Radiology 2000; 217:377–384. 30. Shaw E, Scott C, Souhami L, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: Final report of RTOG protocol 90-05. Int J Radiat Oncol Biol Phys 2000;47:291–298. 31. Lee AW, Kwong DL, Leung SF, et al. Factors affecting risk of symptomatic temporal lobe necrosis: Significance of fractional dose and treatment time. Int J Radiat Oncol Biol Phys 2002;53:75– 85. 32. Kramer S. The hazards of therapeutic irradiation of the central nervous system. Clin Neurosurg 1968;15:301–318. 33. al-Mefty O, Kersh JE, Routh A, et al. The long-term side effects of radiation therapy for benign brain tumors in adults. J Neurosurg 1990;73:502–512. 34. Shrieve DC, Alexander E 3rd, Black PM, et al. Treatment of patients with primary glioblastoma multiforme with standard postoperative radiotherapy and radiosurgical boost: Prognostic factors and long-term outcome. J Neurosurg 1999;90: 72–77. 35. Wowra B, Schmitt HP, Sturm V. Incidence of late radiation necrosis with transient mass effect after interstitial low dose rate radiotherapy for cerebral gliomas. Acta Neurochir (Wien) 1989;99:104 –108. 36. Gutin PH, Prados MD, Phillips TL, et al. External irradiation followed by an interstitial high-activity iodine-125 implant “boost” in the initial treatment of malignant gliomas: NCOG study 6G-82-2. Int J Radiat Oncol Biol Phys 1991;21:601– 606. 37. Leibel SA, Gutin PH, Wara WM, et al. Survival and quality of life after interstitial implantation of removable high-activity iodine-125 sources for the treatment of patients with recurrent

38.

39.

40. 41. 42. 43. 44.

45. 46.

47. 48. 49.

50. 51.

52. 53.

54.

55.

● J. D. RUBEN et al.

507

malignant gliomas. Int J Radiat Oncol Biol Phys 1989;17: 1129 –1139. Burger PC, Mahley MS Jr., Dudka L, et al. The morphologic effects of radiation administered therapeutically for intracranial gliomas: A postmortem study of 25 cases. Cancer 1979; 44:1256 –1272. Rubinstein JL, Herman MM, Long TF, et al. Leukoencephalopathy following combined therapy of central nervous system leukemia and lymphoma. Acta Neuropathol Suppl (Berl) 1975;(Suppl. 6):251–255. Edwards M, Wilson C. Treatment of radiation necrosis. In: Kagan A, editor. Radiation damage to the nervous system. New York: Raven Press; 1980. p. 129 –153. Floyd NS, Woo SY, Teh BS, et al. Hypofractionated intensitymodulated radiotherapy for primary glioblastoma multiforme. Int J Radiat Oncol Biol Phys 2004;58:721–726. Forsyth PA, Kelly PJ, Cascino TL, et al. Radiation necrosis or glioma recurrence: Is computer-assisted stereotactic biopsy useful? J Neurosurg 1995;82:436 – 444. Kramer S, Southard M, Mansfield C. Radiation effect and tolerance of the central nervous system. Front Rad Ther Oncol 1972;6:332–345. van der Kogel A. The nervous system: Radiobiology and experimental pathology. In: Scherer E, Streffe C, Trott KR, editors. Radiopathology of organs and tissues. Berlin: Springer-Verlag; 1991. p. 191–212. Becker G, Kocher M, Kortmann RD, et al. Radiation therapy in the multimodal treatment approach of pituitary adenoma. Strahlenther Onkol 2002;178:173–186. Zaugg M, Adaman O, Pescia R, et al. External irradiation of macroinvasive pituitary adenomas with telecobalt: A retrospective study with long-term follow-up in patients irradiated with doses mostly of between 40 – 45 Gy. Int J Radiat Oncol Biol Phys 1995;32:671– 680. Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 2005;352:987–996. Mahaley MS Jr., Whaley RA, Blue M, et al. Central neurotoxicity following intracarotid BCNU chemotherapy for malignant gliomas. J Neurooncol 1986;3:297–314. Di Chiro G, Oldfield E, Wright DC, et al. Cerebral necrosis after radiotherapy and/or intraarterial chemotherapy for brain tumors: PET and neuropathologic studies. Am J Roentgenol 1988;150:189 –197. Allen JC, Rosen G, Mehta BM, et al. Leukoencephalopathy following high-dose iv methotrexate chemotherapy with leucovorin rescue. Cancer Treat Rep 1980;64:1261–1273. Allen JC, Thaler HT, Deck MD, et al. Leukoencephalopathy following high-dose intravenous methotrexate chemotherapy: Quantitative assessment of white matter attenuation using computed tomography. Neuroradiology 1978;16:44 – 47. Safdari H, Fuentes JM, Dubois JB, et al. Radiation necrosis of the brain: Time of onset and incidence related to total dose and fractionation of radiation. Neuroradiology 1985;27:44 – 47. Rubinstein LJ, Herman MM, Long TF, et al. Disseminated necrotizing leukoencephalopathy: A complication of treated central nervous system leukemia and lymphoma. Cancer 1975;35:291–305. Souhami L, Seiferheld W, Brachman D, et al. Randomized comparison of stereotactic radiosurgery followed by conventional radiotherapy with carmustine to conventional radiotherapy with carmustine for patients with glioblastoma multiforme: Report of Radiation Therapy Oncology Group 93-05 protocol. Int J Radiat Oncol Biol Phys 2004;60:853– 860. Herrmann T, Voigtmann L, Knorr A, et al. The time– dose relationship for radiation-induced lung damage in pigs. Radiother Oncol 1986;5:127–135.

508

I. J. Radiation Oncology

● Biology ● Physics

56. Overgaard J, Hansen HS, Specht L, et al. Five compared with six fractions per week of conventional radiotherapy of squamous-cell carcinoma of head and neck: DAHANCA 6 and 7 randomised controlled trial. Lancet 2003;362:933–940. 57. Fu KK, Pajak TF, Trotti A, et al. A Radiation Therapy Oncology Group (RTOG) Phase III randomized study to compare hyperfractionation and two variants of accelerated fractionation to standard fractionation radiotherapy for head-andneck squamous cell carcinomas: First report of RTOG 9003. Int J Radiat Oncol Biol Phys 2000;48:7–16. 58. Bourhis J, Calais G, Lapeyre M, et al. Concomitant radiochemotherapy or accelerated radiotherapy: Analysis of two randomized trials of the French Head and Neck Cancer Group (GORTEC). Semin Oncol 2004;31:822– 826. 59. Taylor JM, Mendenhall WM, Lavey RS. Dose, time, and fraction size issues for late effects in head-and-neck cancers. Int J Radiat Oncol Biol Phys 1992;22:3–11. 60. Hornsey S, Myers R, Coultas PG, et al. Turnover of proliferative cells in the spinal cord after X irradiation and its relation to time-dependent repair of radiation damage. Br J Radiol 1981;54:1081–1085. 61. Field SB, Hornsey S. Slow repair after x-rays and fast neutrons. Br J Radiol 1977;50:600 – 601. 62. Chappell R, Nondahl DM, Rezvani M, et al. Further analysis of radiobiological parameters from the First and Second British Institute of Radiology randomized studies of larynx/ pharynx radiotherapy. Int J Radiat Oncol Biol Phys 1995;33: 509 –518. 63. Horiot JC, Bontemps P, van den Bogaert W, et al. Accelerated fractionation (AF) compared to conventional fractionation (CF) improves loco-regional control in the radiotherapy of advanced head and neck cancers: Results of the EORTC 22851 randomized trial. Radiother Oncol 1997;44: 111–121. 64. Jen YM, Hsu WL, Chen CY, et al. Different risks of symptomatic brain necrosis in NPC patients treated with different altered fractionated radiotherapy techniques. Int J Radiat Oncol Biol Phys 2001;51:344 –348. 65. Babu R, Huang PP, Epstein F, et al. Late radiation necrosis of the brain: Case report. J Neurooncol 1993;17:37– 42.

Volume 65, Number 2, 2006

66. Woo E, Lam K, Yu YL, et al. Temporal lobe and hypothalamic-pituitary dysfunctions after radiotherapy for nasopharyngeal carcinoma: A distinct clinical syndrome. J Neurol Neurosurg Psychiatry 1988;51:1302–1307. 67. Kulinskii VI, Klimova AD. [The radioprotective effect of GABA-tropic substances, gamma-hydroxybutyrate and piracetam]. Radiobiologiia 1993;33:133–136. 68. Grattan-Smith PJ, Morris JG, Langlands AO. Delayed radiation necrosis of the central nervous system in patients irradiated for pituitary tumours. J Neurol Neurosurg Psychiatry 1992;55:949 –955. 69. Aristizabal SA, Boone ML, Laguna JF. Endocrine factors influencing radiation injury to central nervous tissue. Int J Radiat Oncol Biol Phys 1979;5:349 –353. 70. Hopewell JW, Wright EA. The nature of latent cerebral irradiation damage and its modification by hypertension. Br J Radiol 1970;43:161–167. 71. Ehrenfeld CE, Maschke M, Dorfler A, et al. Is stereotactic biopsy a reliable method to differentiate tumor from radiation necrosis? Clin Neuropathol 2002;21:9 –12. 72. Gutin P. Treatment of radiation necrosis of the brain. In: Sheline GE, editor. Radiation injury to the nervous system. New York: Raven Press; 1991. p. 271–281. 73. Woo E, Lam K, Yu YL, et al. Cerebral radionecrosis: Is surgery necessary? J Neurol Neurosurg Psychiatry 1987;50: 1407–1414. 74. Lee AW, Ng SH, Ho JH, et al. Clinical diagnosis of late temporal lobe necrosis following radiation therapy for nasopharyngeal carcinoma. Cancer 1988;61:1535–1542. 75. Shaw PJ, Bates D. Conservative treatment of delayed cerebral radiation necrosis. J Neurol Neurosurg Psychiatry 1984;47: 1338 –1341. 76. Marks MP, Delapaz RL, Fabrikant JI, et al. Intracranial vascular malformations: Imaging of charged-particle radiosurgery. Part II. Complications. Radiology 1988;168:457– 462. 77. Joiner MC, Bentzen SM. Time– dose relationships: The linearquadratic approach. In: Steele GG, editor. Basic clinical radiobiology. 3rd ed. London: Arnold; 2002. p. 120 –133. 78. Steele GG. The dose rate effect: Brachytherapy and targeted radiotherapy. In: Steele GG, editor. Basic clinical radiobiology. London: Arnold; 2002. p. 192–204.

APPENDIX 1 CALCULATING THE BED FOR LOW-DOSE-RATE BRACHYTHERAPY For incomplete repair, the isoeffective dose at 2 Gy per fraction (ID2) ⫽ D(D.g ⫹ {␣/␤})/(2 ⫹ ␣/␤), where D is total dose and g is the incomplete repair factor (77).

g ⫽ 2[␮.t ⫺ 1 ⫹ exp(⫺␮.t)]/(␮.t)2, where ␮ (repair constant) is 0.693/t1/2 and t is exposure time (78). t1/2 is the repair half time, assumed to be 6 h for brain. BED ⫽ ID2(1 ⫹ {2/ ␣/␤}).

APPENDIX 2 The Neuret is a unit of biological dose devised by Sheline et al. (23). It represents a modification of the Nominal Standard Dose (NSD) formula which gives more weight to

fractionation and less to time factors than the standard Ellis NSD formula. Neuret ⫽ D ⫻ N⫺0.44 ⫻ T⫺0.06