Original paper Radiation therapy in paediatric = gliomas: our institutional experience

Introduction =
Central nervous system (CNS) tumours are the = second=20 commonest cancers in Indian children after = lymphoma/leukaemias=20 [1]. More than 50% of all paediatric CNS tumours are = gliomas=20 [2,3]. In contrast to adults, low-grade gliomas (LGG) = are=20 commoner and carry better prognosis in children. = Location=20 wise, infra-tentorial location is more frequent as = compared to=20 those in adults [4]. Though surgery remains the = mainstay=20 treatment in paediatric gliomas, radiation therapy = (RT) also=20 plays an important role. It is mainly indicated as = adjuvant=20 treatment for all high-grade gliomas (HGG) and = incompletely=20 resected or progressive LGG; and as definitive = treatment, with=20 or without chemotherapy in inoperable gliomas.
The = planning and delivery of RT for children is = technically=20 challenging, especially in those requiring daily = anaesthesia.=20 Immobilization with head casts and other accessories = is=20 essential for precise treatment. Though RT plays a = vital role,=20 radiation-related late side-effects remain a concern. = With the=20 advances in imaging and radiation delivery methods, it = is=20 possible to confine the radiation dose to the target = area and=20 minimize the radiation to normal brain and therefore = reduce=20 late adverse effects [5]. The aim of our study is to = analyze=20 the clinical outcome of paediatric astrocytic glioma = patients=20 treated with RT.

Material and = methods=20
For this retrospective analysis, we retrieved the = case=20 records of all the paediatric astrocytic glioma = patients who=20 received RT in our department between 2004 and 2007. = Within=20 each case record, we analyzed the information = regarding the=20 patient’s demography, clinical details, = diagnosis, treatment=20 given, RT details, and survival. Inclusion criteria = consisted=20 of: (1) age less than 18 years, (2) histopathological=20 diagnosis of glioma of any grade or (3) convincing=20 radiological diagnosis of glioma in the absence of=20 histopathological diagnosis. Patients who defaulted = the=20 treatment in between, as well as patients with=20 medulloblastoma, ependymoma and other non-glioma = histology,=20 were excluded from the present analysis.
The = initial=20 clinical workup of the patients consisted of detailed=20 clinical/neurological examination, complete blood = count, liver=20 and kidney function tests, plain chest X-ray, and = computed=20 tomography (CT) or magnetic resonance imaging (MRI) of = the=20 brain (and spine, if necessary). After the surgery or = biopsy,=20 patients were referred for RT. Patients who were not = eligible=20 for biopsy due to lesions located in the=20 brainstem/inaccessible areas were directly sent to our = RT=20 clinic. Expected benefit and sequelae of RT were = explained in=20 detail to the parents of each child. All children who = were=20 thought to require anaesthesia during RT (children up = to 5=20 years of age in general, and uncooperative children) = were sent=20 to the pre-anaesthesia clinic (PAC) for assessment. = After the=20 PAC check-up, eligible patients were then called for = RT=20 planning. All the patients who required general = anaesthesia=20 were essentially admitted to the RT indoor unit. = Others were=20 treated on an out patient department basis. Patients = were=20 simulated on an X-ray/CT-simulator and treated on a = linear=20 accelerator/cobalt. Meticulous shielding of the field = was done=20 wherever necessary.
The various indications of RT = in our=20 patients were as follows: (1) high-grade tumours, (2)=20 inoperable tumours and (3) low-grade tumours with = post-surgery=20 residual or progressive disease. The usual dose of RT = was=20 50-56 Gy for low-grade gliomas and for children below = the age=20 of 5 years and 56-60 Gy for high-grade gliomas and = patients=20 above the age of 5 years.
During the course of RT, = patients were assessed at least once weekly. Complete = blood=20 count was done for each patient every week. Follow-up = was done=20 every month for the first 6 months and then every 3 = months=20 till 2 years. Neuroimaging (CT or MRI) was performed = every 6=20 months or earlier in case of symptomatic = deterioration.=20
Statistical analysis was performed using the = statistical=20 software SPSS, version 11.5. The EFS was defined as = the period=20 from the date of completion of RT to the date of = occurrence of=20 the event, i.e. death or recurrence. EFS was = determined using=20 the Kaplan-Meier survival method [6] with respect to = age=20 (below and above 5 years), sex, location of the tumour = (brainstem vs. non-brainstem), histopathological=20 characteristics (low-grade vs. high-grade), extent of = surgical=20 resection, dose and duration of RT, and use of = chemotherapy.=20 Log-rank test was used to determine the p-value and p = <=20 0.05 was considered significant. =

Results=20
A total of 70 children with glioma received RT = between=20 2004 and 2007. Their median age was 10 years (range = 2-18=20 years). About 80% of patients were above 5 years of = age. There=20 were 42 males and 28 females in the study group. = Duration of=20 presenting symptoms ranged from 10 days to 18 months = (median 4=20 months). Various symptoms at clinical presentation are = shown=20 in Table 1. Headache was the commonest symptom.=20
Histopathologically, pilocytic astrocytoma was the = commonest subtype. Numbers and percentages of = histological=20 subtypes of tumours are provided in Table 2. The = brainstem=20 lesions which could not be biopsied were put in a = separate=20 histopathological variety. Distribution of patients = according=20 to histopathological grade of the tumour was as = follows: grade=20 I – 22, grade II – 7, grade III – 7, = grade IV – 8, unknown=20 grade – 5, and brainstem – 21 patients. = Patients who had=20 brainstem lesions and unknown grade lesions were = treated like=20 HGG. Figure 1 shows the anatomical location of tumours = in our=20 patients. Brainstem was the commonest site (21 = patients),=20 followed by cerebral cortex (20 patients). Most = brainstem=20 tumours were diffuse (18 out of 21). The remaining 3 = brainstem=20 lesions were exophytic tumours but all three of them = were=20 inoperable due to poor general condition of the = patient or=20 refusal by the parents.
Of the 70 patients, 45 = underwent=20 some form of surgery before RT and the remaining 25 = patients=20 were scheduled for RT directly since they were not = eligible=20 for surgery. Table 3 shows the details of surgical = treatment.=20 Twenty-five patients could not undergo any surgical = procedure=20 because of an inaccessible lesion site or poor general = condition. Of the 45 patients, 11 underwent gross = total=20 excision and 32 underwent debulking surgery in the = form of=20 near-total excision or partial removal of the tumour.=20
Details of RT are outlined in Table 4. Almost all = patients=20 received local field RT, except one who received whole = brain=20 RT in view of the multi-centricity of the tumour. Most = patients received a total dose of 56 Gy in 28 = fractions over=20 5.5 weeks. One patient, who had spinal cord glioma = with=20 paraplegia, received palliative radiotherapy with 20 = Gy/5=20 fractions/1 week in view of the poor general = condition. Due to=20 deteriorating clinical condition, one patient could = not=20 complete the prescribed dose of 56 Gy and could = receive only a=20 total dose of 38 Gy. One patient with multicentric = tumour=20 received whole brain RT for a total dose of 40 Gy/20=20 fractions/ 4 weeks. All other patients received a = minimum=20 total dose of 45 Gy. Nineteen patients received = concurrent=20 and/or adjuvant chemotherapy (mostly temozolomide). A = total of=20 64 chemotherapy cycles were given in these 19 patients = (average 3 cycles per patient).
The follow-up = period=20 ranged from 2 to 50 months, with a median of 18 = months. The=20 overall EFS for all patients is depicted in Fig. 2. = The 3-year=20 EFS was 44% and the median EFS period was 18 months. = Age (<=20 5 years vs. > 5 years) and sex of the patients did = not=20 correlate with EFS. As shown in Fig. 3, the 3-year EFS = for LGG=20 and HGG was 46% and 41% respectively (p =3D 0.9). As = shown in=20 Fig. 4, patients with brainstem lesions had = statistically=20 significantly lower 3-year EFS than those with = non-brainstem=20 gliomas (28% vs. 56%, p < 0.01). Extent of surgery = was=20 associated with significantly improved EFS (Fig. 5). = The=20 3-year EFS in patients who underwent partial resection = and no=20 surgery was 58% and 25% respectively (p < 0.042). = The dose=20 of radiation (56 Gy vs. 60 Gy), dose per fraction (1.8 = Gy vs.=20 2.0 Gy) and type of external beam radiation therapy = machine=20 (Cobalt-60 vs. linear accelerator) had no correlation = with the=20 outcome. Though a small proportion of patients = received=20 chemotherapy (19/70), the 3-year EFS was slightly (but = not=20 significantly) better with chemotherapy (Fig. 6). =
Due to=20 the short follow-up period, late toxicity was not = analyzed in=20 this study. Interruption during the course of RT was=20 considered as indirect incidence of acute toxicity. = Eleven=20 patients had interruptions of more than 1-week = duration during=20 RT; 6 of them belonged to the RT-alone group (6/51) = and 5=20 belonged to the chemoradiotherapy group (5/19).=20

Discussion
Facilities for RT of = paediatric CNS tumours are limited to a few centres in = a=20 developing country such as India for various reasons.=20 Management of such patients requires involvement of = multiple=20 specialties including neurosurgery and neuroradiology. = Most=20 children below the age of 5 years require anaesthesia = every=20 day during the delivery of RT and it is therefore a = cumbersome=20 treatment. Facilities of RT, in general, are limited. = Our=20 centre is an apex medical institute and its cancer = centre is=20 one of the 25 regional cancer centres in India. Our = experience=20 being reported here represents the outcome of = paediatric=20 glioma patients treated at an apex medical centre. = Most series=20 [7-10] on paediatric gliomas are not pure glioma = series and=20 they include patients with a wide spectrum of tumours=20 including medulloblastomas, ependymomas, etc. The = series by=20 Varan et al. [10], though titled “astrocytic = tumours in=20 children”, excluded patients with optic glioma, = brainstem=20 glioma, and diencephalic tumours. Our series comprised = patients with paediatric gliomas only and therefore = our=20 results may not be strictly comparable with these = studies=20 [7-10]. Apparently, EFS for the entire group of = patients in=20 our series seems to be lower than the literature since = our=20 series had a sizeable number of brainstem lesions = which are=20 known to have worse prognosis.
As reported in = other series=20 [7-8,11], the extent of surgery before RT is = associated with=20 improved survival. We also noticed significantly = better=20 survival (58% vs. 25%) in patients who had undergone = surgery=20 than those who had no surgery (Fig. 5).
RT = provides better=20 results in LGG as compared with HGG. In our series, = the 3-year=20 EFS rate was marginally better in favour of LGG (Fig. = 3) but=20 it was not statistically significant. This may be due = to two=20 factors. Firstly, LGG generally have good prognosis = and=20 routine adjuvant RT is not required. However, in the = case of=20 patients who show symptomatic and radiological = progression=20 during the follow-up, RT alone or after re-resection = plays a=20 vital role; the outcome may not be as good as in those = who=20 undergo complete resection and do not show further=20 progression. Several patients with LGG in our series = were=20 considered for RT when they had either significant=20 post-surgery residual disease or progression on MRI = with or=20 without deteriorating clinical symptoms. The other = reason=20 could be that, in our setup, patients find it = difficult to=20 comply with the MRI surveillance policy very strictly = due to=20 MRI costs and the long distances to travel. Therefore, = several=20 of these patients, when they show progression, report = for the=20 follow-up with frank clinical progression with large = tumours=20 not amenable to complete re-resection. RT for such = patients=20 may be less effective. Therefore, early use of RT in = patients=20 with post-surgery residual disease may be worth = re-visiting.=20 There is, however, a study [12] in adult patients with = LGG=20 showing better progression-free survival with use of = RT=20 immediately after sur-gery/biopsy as compared to RT = used due=20 to progression. In this study, Arm A patients received = RT=20 within 8 weeks after resection or biopsy and Arm B = patients=20 received RT when they showed clinical and/or = radiological=20 progression. Progression-free survival at 5 years was=20 significantly better in Arm A as compared to Arm B = (44% vs.=20 37%, p =3D 0.02). We feel that for incompletely = resected=20 tumours, RT should be initiated early since the modern = 3D=20 conformal RT techniques such as stereotactic RT, = intensity=20 modulated RT (IMRT), etc. have shown reduced morbidity = in=20 several trials [5,9,13].
RT remains an essential = component=20 of treatment in HGG even after gross total excision. = The=20 3-year EFS rate of 41% in HGG in our study is more or = less the=20 same as reported by Bauman et al. [7] and Varan et al. = [10] in=20 their series.
The various RT parameters did not = alter the=20 outcome significantly. The duration of overall = treatment (more=20 than 8 weeks) seems to have a correlation with worse = outcome=20 in our series but it could not reach statistical = significance=20 due to the small number of patients (11 out of 70) = completing=20 RT in 8 weeks or more time. Various studies [14,15] = have tried=20 altered fractionation schedules in the form of=20 hyperfractionation [14] and accelerated fractionation = [15],=20 especially in malignant gliomas and brainstem gliomas, = but=20 failed to show any significant benefit over = conventional=20 fractionation.
Several trials [16-19] have = reported their=20 results with the use of chemotherapy in paediatric = gliomas.=20 Various chemotherapeutic drugs are used in different=20 combinations for both LGG [17] and HGG [16,18,19].=20 Temozolomide, which has shown promising results in = adult=20 glioblastoma multiforme [20], has shown varying = results in=20 paediatric gliomas. In our study we have used = chemotherapy on=20 a selective basis. Though there was no rigid policy = for=20 selection of patients for chemotherapy, it was avoided = in=20 patients with poor general condition, low = socioeconomic status=20 and unwillingness for chemotherapy. Temozolomide for=20 glioblastoma multi-forme and combination of = carboplatin and=20 etoposide for other HGG and brainstem tumours were the = preferred regimens. Our results have shown a trend = towards=20 improved EFS, though not statistically significant, = with the=20 use of chemotherapy (Fig. 6).
Brainstem glioma is = an=20 important group of CNS tumours in children. = Conventional RT=20 remains the standard of care for such patients, but = the=20 outcome continues to be poor, with 5-year survival = ranging=20 from 20% to 30% [3,4,21,22]. The 3-year EFS of 28% in = our=20 series is consistent with that reported in the = lite-rature.=20 Chemotherapy with newer agents such as temo-zolomide = [23] has=20 been tried in some recent studies in order to improve = the=20 survival, and the results are encouraging.=20

Conclusions
Our results of RT = in HGG=20 and brainstem gliomas in children are consistent with = those in=20 the literature. However, the results of LGG in our = series are=20 probably not very decent, possibly because of our = policy of=20 delaying the RT till progression after surgery. We = therefore=20 suggest a revisit of the policy of early use of RT for = incompletely resected LGG even though they are = asymptomatic.=20

Disclosure
Authors report no = conflict=20 of interest.

References =
1. Indian=20 Council of Medical Research. Population-based cancer = registry.=20 Indian Council of Medical Research, New Delhi, India=20 2001-2003.
2. Packer R.J. Childhood brain = tumors:=20 accomplishments and ongoing challenges. J Child Neurol = 2008;=20 23: 1122-1127.
3. Freeman C.R., Farmer = J.P., Taylor R.E.=20 Central nervous system tumors in children. In: = Halperin E.C.,=20 Parez C.A., Brady L.W. (eds.). Perez and Brady’s = principles=20 and practice of radiation oncology. 5th edition. = Lippincott=20 Williams  Wilkins, Philadelphia 2008, pp. = 1822.
4.=20 Halperin E.C., Constine L.S., Tarbell N.J., et al. = Pediatric=20 radiation oncology. 4th edition. Lippincott Williams = =20 Wilkins, Philadelphia 2005.
5. Kirsch D.G., = Tarbell N.J.=20 Conformal radiation therapy for childhood CNS tumors.=20 Oncologist 2004; 9: 442-450.
6. Kaplan = E.L., Meier P.=20 Nonparametric estimation from incom-plete = observations. J Am=20 Stat Assoc 1958; 53: 457-481.
7. Bauman G., = Fisher B.,=20 Cairney E., et al. Radiotherapy for pediatric central = nervous=20 system tumors: a regional cancer centre experience. J=20 Neurooncology 2004; 68: 285-294.
8. Jenkin = D. Long-term=20 survival of children with brain tumors. Oncology = (Williston=20 Park) 1996; 10: 715-719.
9. Merchant T.E., = Zhu Y.,=20 Thompson S.J., et al. Preliminary results from a phase = II=20 trial of conformal radiation therapy for pediatric = patients=20 with localized low grade astrocytoma and ependymoma. = Int J=20 Radiat Oncol Biol Phys 2002; 52: 325-332.
10. = Varan A.,=20 Aky=FCz C., Akalan N., et al. Astrocytic tumors in = children:=20 treatment results from a single institution. Childs = Nerv Syst=20 2007; 23: 315-319.
11. Pollack I.F. The role of = surgery in=20 pediatric gliomas. J Neurooncol 1999; 42: 271-288. =
12.=20 Karim A.B., Afra D., Cornu P., et al Randomized trial = on the=20 efficacy of radiotherapy for cerebral low-grade glioma = in the=20 adult: European Organization for Research and = Treatment of=20 Cancer Study 22845 with the Medical Research Council = study=20 BRO4: an interim analysis. Int J Radiat Oncol Biol = Phys 2002;=20 52: 316-324.
13. Marcus K.J., Goumnerova L., = Billett A.L.,=20 et al. Stereotactic radiotherapy for localized low = grade=20 gliomas in children: final results of a prospective = trial. Int=20 J Radiat Oncol Biol Phys 2005; 61: 374-379.
14. = Freeman=20 C.R. Hyperfractionated radiotherapy for diffuse = intrinsic=20 brain stem tumors in children. Pediatr Neurosurg 1996; = 24:=20 103-110.
15. Lewis J., Lucraft H., Gholkar A. = UKCCSG study=20 of accelerated radiotherapy for pediatric brain stem = gliomas.=20 United Kingdom Childhood Cancer Study Group. Int J = Radiat=20 Oncol Biol Phys 1997; 38: 925-929.
16. Pollack = I.F.,=20 Boyett J.M., Finlay J.L. Chemotherapy for high grade = gliomas=20 of childhood. Childs Nerv Syst 1999; 15: 529-544. =
17. Khaw=20 S.L., Coleman L.T., Downie P.A., et al. Temozolomide = in=20 pediatric low-grade glioma. Pediatr Blood Cancer 2007; = 49:=20 808-811.
18. Verschuur A.C., Grill J., = Lelouch-Tubiana A.,=20 et al. Temozolomide in pediatric high-grade glioma: a = key for=20 combination therapy? Br J Cancer 2004; 91: 425-429. =
19.=20 Lashford L.S., Thiesse P., Jouvet A., et al.; a United = Kingdom=20 Children’s Cancer Study Group and French Society = for Pediatric=20 Oncology Intergroup Study. Temozolomide in malignant = gliomas=20 of childhood: a United Kingdom Children’s Cancer = Study Group=20 and French Society for Pediatric Oncology Intergroup = Study. J=20 Clin Oncol 2002; 20: 4684-4691.
20. Stupp R., = Mason W.P.,=20 van den Bent M.J., et al.; European Organisation for = Research=20 and Treatment of Cancer Brain Tumor and Radiotherapy = Groups;=20 National Cancer Institute of Canada Clinical Trials = Group.=20 Radiotherapy plus concomitant and adjuvant = temozolomide for=20 glioblastoma. N Engl J Med 2005; 352: 987-996.
21. = Broniscer A., Gajjar A. Supratentorial high-grade = astrocytoma=20 and diffuse brainstem glioma: two challenges for the = pediatric=20 oncologist. Oncologist 2004; 9: 197-206.
22. = Freeman C.R.,=20 Farmer J.P. Pediatric brain stem gliomas: a review. = Int J=20 Radiat Oncol Biol Phys 1998; 40: 265-271.
23. = Broniscer=20 A., Iacono L., Chintagumpala M., et al. Role of = temozolomide=20 after radiotherapy for newly diagnosed diffuse = brainstem=20 glioma in children. Cancer 2005; 103: 133-139.