Childs Nerv Syst DOI 10.1007/s00381-014-2374-y

ORIGINAL PAPER

Suprasellar pilocytic astrocytoma: one national centre’s experience Tafadzwa Mandiwanza & Chandrasekaran Kaliaperumal & Ayman Khalil & Muhammad Sattar & Darach Crimmins & John Caird

Received: 15 November 2013 / Accepted: 28 January 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Introduction Pilocytic astrocytomas in the supratentorial compartment make up 20 % of all brain tumours in children with only 5 % of these arising in the suprasellar region. Optic pathway gliomas or suprasellar gliomas are often seen in neurofibromatosis type 1 (NF1) patients. Given their location, suprasellar pilocytic astrocytomas are challenging to manage surgically with high morbidity rates from surgical resection. We assess our cohort of patients with suprasellar pilocytic astrocytoma and document our experience. Method A retrospective review of patients diagnosed with suprasellar glioma between 2000–October 2012. We included patients diagnosed with optic pathway glioma based on radiological features (with or without biopsy) and those who had a biopsy confirming pilocytic astrocytoma. Results Fifty-three patients included (sporadic tumours 24 and NF1 related 29). Fifteen sporadic and four NF1 patients were biopsied. Twelve sporadic and 13 NF1 patients were initially treated with chemotherapy while only 1 patient had radiotherapy initially. Progression was noted in 58 % of the sporadic group and 24 % of the NF1 group. The only significant factor for progression was NF1 status (p=0.026). Conclusion Management should be guided by individual patient circumstance. In our cohort, chemotherapy did not significantly improve progression free survival; however, NF1 status significantly correlated with the decreased progression. T. Mandiwanza (*) : C. Kaliaperumal : M. Sattar : D. Crimmins : J. Caird Department of Paediatric Neurosurgery, Children’s University Hospital, Temple Street, Dublin 1, Ireland e-mail: [email protected] A. Khalil Department of Neurosurgery, Beaumont Hospital, Beaumont Road, Dublin 9, Ireland

Keywords Pilocytic astrocytoma . Suprasellar . Optic pathway glioma . Neurofibromatosis type 1

Introduction Tumours of the suprasellar region pose a challenge to surgical management given the eloquence of the area and resulting high morbidity. In children, suprasellar gliomas (involving optic pathway, hypothalamus or both) consist of 5 % of CNS tumours with 6–20 % of neurofibromatosis type 1 (NF1) patients having these lesions [7, 21]. These lesions can present with visual deficit, endocrinopathies, seizures or hydrocephalus; however, in a majority of cases, they are an incidental finding [3, 11, 16, 17]. The behaviour of suprasellar pilocytic astrocytomas can be unpredictable in some cases although most run an indolent course and rarely undergo malignant transformation [1]. We assess our current cohort of patients with suprasellar glioma and our management of these cases.

Methods and materials This is a single-centre retrospective review of patients below the age of 16 years diagnosed with a suprasellar lesion from 2000 to October 2012. Patients were selected from the neurooncology multidisciplinary team database and from the patient database of our institution using the keywords “optic glioma”, “suprasellar glioma”, “suprasellar lesion” and “neurofibromatosis type 1”. We included those who had radiological diagnosis of suprasellar/optic pathway glioma (OPG). Of those who had had a biopsy, we only included those with Grade 1 astrocytoma. Demographic information was gathered as well as information regarding initial presentation, NF1 status, initial management (biopsy, surgical resection, chemotherapy,

Childs Nerv Syst

radiotherapy), progression, time to progression and any subsequent progression and treatment. The age at diagnosis was recorded in months as were the time to progression and follow-up periods. Descriptive statistics and Kaplan-Meier analysis were applied to the data collected.

Results Fifty-six patients met the initial radiological criteria; however, three patients were excluded as histology revealed their lesions to be a germinoma, a grade 2 astrocytoma and Langerhans cell histiocytosis, respectively. The data of 53 patients were analysed for this study. The median follow-up period was 62 months with a mean follow-up of 72.9 46.2 months. Overall, in both NF1 and sporadic tumour groups, the average age at diagnosis was 52.2±47.6 months.. Neurofibromatosis status The sporadic group made up 45.3 % (24/53) while the NF1 group was 54.7 % (29/53) with a female preponderance. The mean age at diagnosis for the NF1 group was 58.3 months (range 4–202 months). Radiological features Lesions were divided into three groups depending on their MRI findings and based on the Dodge classification for staging of gliomas of the optic nerves [4]: stage A - optic nerves, stage B - chiasmal lesions with or without optic nerve involvement and stage C - hypothalamic or other adjacent structures. Stage A=16, stage B=15 and stage C=22. Table 1 illustrates the radiological findings, their distribution between the sporadic and NF1 groups and the surgical management. Stage C was the commonest radiological type in our cohort and was more common in the sporadic group.

Table 1 Radiological classification and surgical management in both NF1 and sporadic groups Radiological subtype

NF1 status

Stage A (16)

Sporadic NF1 Sporadic NF1 Sporadic

2 14 4 11 17

NF1

5

Stage B (15) Stage C (22)

Primary surgical management

Partial resection Biopsy only Multiple cyst aspiration Biopsy and resection Biopsy only Partial resection

1 1 1 2 12 1

Surgical management Fifteen of the sporadic group had a biopsy (10 - stereotactic, 3 - open, 2 - endoscopic biopsy). Histology was pilocytic astrocytoma in the 15 patients. In the NF1 group, four patients had an open biopsy and partial surgical resection. Histology was pilocytic astrocytoma in all these patients. Two of the surgical resections were not performed in our institution. Of the two resections done in our centre, both were sporadic tumours and one died 10 days post-operatively. The other patient had surgical resection a progressive lesion and troublesome hydrocephalus and required a radiotherapy following resection. Of the 53 patients, 33 % required CSF diversion with a shunt (14 in the sporadic group and 4 in the NF1 group). Radiotherapy Progressive disease was treated with radiotherapy in eight patients (six sporadic and two NF1). Of these six sporadic tumour patients, five had had previous chemotherapy. The mean age at which radiotherapy was given in the sporadic group was 7 years (92 months). Of the two NF1 patients treated with radiotherapy, one had previous chemotherapy while the other (aged 16) had previous partial resection of his lesion. Radiotherapy was administered in these patients at ages 3 and 16, respectively. Chemotherapy Twelve patients in the sporadic group had neoadjuvant chemotherapy (seven - vincristine and carboplatin, three - vincristine, etopamide and carboplatin, two - vincristine, carboplatin, cyclophosphamide and cisplatin). Four of these were treated without prior biopsy. Twelve NF1 patients had neoadjuvant chemotherapy (11 vincristine and carboplatin, 1 - vincristine, carboplatin, cyclophosphamide and cisplatin). Nine patients were treated with chemotherapy for progressive disease (five - sporadic group and four - NF1 group). For three patients who had more the one progression, chemotherapy was the second- and third-line treatments of choice (Table 2). Observation Eleven patients in the sporadic group had no initial treatment although seven of these had a biopsy. In the NF1 group, 15 patients were observed clinically and none of them had prior biopsy. Overall, seven of the patients clinically observed went on to have disease progression.

Childs Nerv Syst Table 2 Summary of the patients who had more than one progression

NF1 status Initial management Time to first progression (months) Second-line treatment Time to second progression (months) Third-line treatment Time stable (months)

Patient 1

Patient 2

Patient 3

No

Yes Observation

Yes SIOP-LGG2

17

4

Biopsy Vinblastine 13

SIOP-LGG2

Vinblastine

24

15

Avastin + Irinotecan 4

Vinblastine

Bevacizumab and Irinotecan 8

SIOP-LGG2 62

10

SIOP-LGG2 International Society of Paediatric Oncology-Low Grade Glioma 2 protocol 2: vincristine, carboplatin and etoposide."

Progression

Fig. 1 Progression in NF1 group versus sporadic group

Fourteen (58 %) patients in the sporadic group had tumour progression with a mean of 26.7 progression-free months. The mean age at progression in this group was 82.2 months in this group with the majority of these having been initially diagnosed under the age of 48 months. Radiological tumour progression, deterioration in visual acuity or both occurred in 24 % (7/29) of the NF1 group. The mean age at diagnosis of these patients was 35.6 months and only one patient was diagnosed under the age of 12 months. The average time from diagnosis to progression was 48.6 months (range 4–146 months). There was a significant difference between the sporadic and NF1 group in the progression-free survival (PFS) at 5 years, with 34 % in the non-NF1 group and 73 % in the NF1 group, p=0.002 (Fig. 1). The 5-year PFS in the NF1 patients not treated with chemotherapy was 85 % (126.6 months; confidence interval (CI) 103.1-150.1) versus 72 % (118.3 months; CI 87.1-148.8) in those with NF1 treated with chemotherapy (Fig. 2). This difference was not statistically significant (p=0.56). In the sporadic tumour group, the 5-year PFS was 50 % (66.5 months; CI 23.5-109.4) in those who did not have chemotherapy and 52 % (39.5 months; CI 27.6-51.4) in those who were treated with chemotherapy (Fig. 3). Statistically, this difference was not significant (p=0.14). The age at diagnosis was analysed as a prognostic factor. Progression in both groups occurred in 71.4 % of those diagnosed below the age of 6 years in comparison to 28.6 % of those diagnosed above the age of 6 years. In the NF1 group, we found a 5-year PFS of 85 % for those diagnosed below the age of 6 years and 67 % in those diagnosed over the age of

6 years. Age analysed in the sporadic group yielded a 5-year PFS of 34 % in those diagnosed below the age of 6 years and 33 % in those diagnosed over the age of 6 years. On multivariate analysis, having NF1 was associated with a five-time reduction in progression (risk ratio 0.26, p=0.026, 95 % CI). This was adjusted for chemotherapy and age at diagnosis, which were not found to be major predictors (p= 0.19 and 0.60, respectively).

Discussion Suprasellar gliomas/optic-chiasmatic-hypothalamic glioma (OCHG) usually affect the optic nerve; nonetheless, they can affect chiasmal structures and the hypothalamus accounting for their clinical presentations as endocrinopathies or visual deficits. Over 75 % however are asymptomatic and diagnosed

NF1 Group 29

Chemotherapy 11

Observation 15

Biopsy and/ resection 3

chemotherapy 2

Fig. 2 Initial treatment of NF1 patients

Radiotherapy 1

Childs Nerv Syst Fig. 3 Initial treatment of sporadic group

Sporadic group 24

Biopsy Only 15

Chemotherapy 8

Biopsy and resection 1

Chemotherapy 4

Observation 4

Observation 7

Chemotherapy

on the screening of other stigmata of NF1 [21]. This condition is closely associated with NF1 and indeed over half of our patients have phenotypic or genetic features of NF1. Of our cohort, only 41 % were asymptomatic with the rest presenting with visual deficits, pressure symptoms or endocrine-related symptoms. Hydrocephalus is reported to occur in 10–20 % of these patients; however, our cohort has a higher incidence at 33 % [19, 25]. This may possibly be due to a high proportion of chiasmatic Dodge stage C lesions, which are more likely to obstruct the CSF circulation. The clinical management of these lesions can be challenging. Some of these lesions can be indolent and have been labelled self-limiting by some authors, especially in NF1 patients, while other lesions can exhibit aggressive behaviour despite the histology of pilocytic astrocytoma [1]. Multidisciplinary management, with the involvement of endocrinologists, ophthalmologists, neurosurgeons and oncologists, is advised given the wide spectrum of presenting symptoms and complications [20].

Chemotherapy is the modality of choice in children under the age of 5 as it delays the use of radiotherapy in NF1 patients and is utilised for progressive tumours. There are reported 10-year survival rates of 80–90 % with chemotherapy although some patients do not respond to it and some have tumour relapse [9, 12, 22]. Visual deterioration is rarely reversed, and in some cases, it continues to deteriorate after chemotherapy [2, 12]. A combination of carboplatin and vincristine has been shown to delay radiotherapy in young children with progressive low-grade glioma [24]. In our group, 67 % of NF1 patents initially treated with chemotherapy went on to have disease progression. The fact that these patients required chemotherapy initially suggests that they had more aggressive tumours from the outset. The difference we are seeing may be due the tumour biology as opposed to a negative effect of chemotherapy.

Surgical treatment

Radiotherapy

Low-grade gliomas, in general, have a slower course and are known to have good prognosis when completely resected [3, 8]. Considerable debate exists as to whether surgical treatment is optimal in these patients. Surgery is usually performed to ameliorate mass effect for those with severe neurological symptoms or visual deterioration and for progressive lesions; however, it is usually associated with high morbidity rates [1, 2, 20, 23, 26]. Surgery for chiasmatic lesions has been associated with a risk of infarct with no positive prognostic implications from partial resection in this area [6]. In our department, we do not generally favour surgical management of these patients as we have seen very poor outcomes in those who have had procedures elsewhere prior to attending our service. Of the two patients we have managed surgically, there was one death and the surviving patient had severe morbidity.

The use of radiotherapy is controversial, with a high risk of long-term adverse effects to cognition and endocrine function in very young children and infants. In older children and adults, the British neuro-oncology society recommends radiotherapy as the initial treatment in cases with severe ophthalmologic symptoms, severe neurological symptoms or for progressive symptoms [20]. There is evidence that radiotherapy has vision sparing effect, with visual improvement in 24– 36 % of those treated aged over 5. In NF1 patients, radiotherapy is not routinely recommended, except in extreme cases, due to a higher risk of secondary tumours in that group of patients [12, 18, 20]. Tumour stabilization occurs in 42–92 % although the duration of stabilization is not quantified and may be transient [2, 9, 12, 15]. In our group, no patient had radiotherapy under the age of 7 years and those treated with this modality have

Childs Nerv Syst

remained progression-free since treatment. One child had improvement of his vision from 6/9 bilaterally to 6/6 bilaterally after radiotherapy. Progression The clinical course of suprasellar pilocytic astrocytomas can be unpredictable with a “wait-and-see” attitude being adopted for asymptomatic or stable disease. Treatment is reserved for those with progressive disease or symptomatic disease [12]. Ages at diagnosis and NF1 status have been found to be prognostic factors. Age less than 1 year has been found to be the most significant statistically proven prognostic factor. Patients diagnosed at this age or below have shorter progression-free survival rates than older patients [1, 5, 10, 13, 21]. Our findings suggest that diagnosis age is a factor in the NF1 group where a difference in a 5-year PFS was noted. In the sporadic tumour group, however, age of diagnosis did not yield a difference in a 5-year PFS. NF1 patients rarely have tumour progression after their first decade of life, and in some cases, tumours actually regress [2, 14, 15, 24]. Regression may be complete radiological involution or partial reduction in tumour size. Reports of regression have occurred in some NF1 patients with a mean age of diagnosis of 52.9 months who did not receive any treatment [15]. Our results correlate with this, with a statistically longer 5-year PFS in our NF1 patients of 73 % (p=0.002).

Conclusion These lesions are rare and almost solely a paediatric disease. Management should be guided by individual patient circumstances with radiotherapy being delayed in young children and infants as well. Small cohort numbers preclude us from making conclusive statements and long-term follow-up studies will be required to determine the impact of this condition on the child growing into an adult. From our cohort, NF1 status was the only significant factor influencing progression. Acknowledgment The first author is in receipt of an Educational grant from DePuy Synthes.

References 1. Aishail E, Rutka J, Becker L, Hoffman H (1997) Optic chiasmatichypothalamic glioma. Brain Pathology 7:799–806 2. Astrup J (2003) Natural history and clinical management of optic pathway glioma. British Journal of Neurosurgery 17(4):327–335

3. Dirven C, Mooij J, Molenaar W (1997) Cerebellar pilocytic astrocytoma: a treatment protocol based upon analysis of 73 cases and a review of the literature. Childs Nervous System 13(1):17–23 4. Dodge H, Lowe J, Craigh W, Dockerty M, Kearns T, Holman C, Hayles A (1958) Gliomas of the optic nerves. Archives of neurology and psychiatry 79(6):607–621 5. Grill J, Laithier V, Rodriguez D, Raquin MA, Pierre-Kahn A, Kalifa C (2000) When do children with optic pathway tumours need treatment? An oncological perspective in 106 patients treated in a single centre. European Journal of Pediatrics 159(9):692–696. doi:10.1007/ s004310000531 6. Hupp M, Falkenstein F, Bison B, Mirow C, Kraub J, Gnekow A, Solymosi L, Warmuth-Metz M (2012) Infarction following chiasmatic low-grade glioma resection. Childs Nervous System 28(3):391– 398 7. Janss A, Grundy R, Cnaan A, Savino P, Packer R, Zackai E, Goldwein J, Sutton L, Radcliffe J, Molloy P (1995) Optic pathway and hypothalamic/chiasmatic gliomas in children younger than age 5 years with a 6-year foIIow-up. Cancer 75(4):1051–1059 8. Keles G, Lamborn K, Berger M (2001) Low-grade hemispheric gliomas in adults: a critical review of extent of resection as a factor influencing outcome. Journal of Neurosurgery 95:735–745 9. Kretschmar C, Linggood R (1991) Chemotherapeutic treatment of extensive optic pathway tumors in infants. Journal of NeuroOncology 10:263–270 10. Massimino M, Spreafico F, Cefalo G, Riccardi R, Tesoro-Tess J, Gandola L, Riva D, Ruggiero A, Valentini L, Mazza E, Genitori L, Di Rocco C, Navarria P, Casanova M, Ferrari A, Luksch R, Terenzani M, Balestrini M, Colosimo C, Fossati-Bellani F (2002) High response rate to cisplatin/etoposide regimen in childhood low-grade glioma. Journal of Clinical Oncology 20(20):4209–4216 11. McKinney P (2004) Brain tumours: incidence, survival and aetiology. Journal of Neurology Neurosurgery Psychiatry 75(suppl 2):ii12–ii17 12. Moreno L, Bautista F, Ashley S, Duncan C, Zacharoulis S (2010) Does chemotherapy affect the visual outcome in children with optic pathway glioma? A systematic review of the evidence. European Journal of Cancer 46(12):2253–2259. doi:10.1016/j.ejca.2010.03.028 13. Opocher E, Kremer LCM, Da Dalt L, van de Wetering MD, Viscardi E, Caron HN, Perilongo G (2006) Prognostic factors for progression of childhood optic pathway glioma: a systematic review. European Journal of Cancer 42(12):1807–1816. doi:10.1016/j.ejca.2006.02. 022 14. Parsa C, Hoyt C, Lesser R, Weinstein J, Strother C, Muci-Mendoza R, Ramella M, Manor R, Fletcher W, Repka M, Garrity J, Ebner R, Monteiro M, McFadzean R, Rubtsova I, Hoyt W (2001) Spontaneuos regression of optic gliomas: thirteen cases documented by serial neuroimaging. Archives of Ophthalmology 119(4):516–529 15. Piccirilli M, Lenzi J, Delfinis C, Trasimeni G, Salvati M, Raco A (2006) Spontaneous regression of optic pathways gliomas in three patients with neurofibromatosis type I and critical review of the literature. Childs Nervous System 22(10):1332–1337. doi:10.1007/ s00381-006-0061-3 16. Rosychuk R, Witol A, Wilson B, Stobart K (2012) Central nervous system (CNS) tumour trend in children in a western Canadian province: a population-based 22-year retrospective study. Journal of Neurology 259:1131–1136 17. Sawamura Y, Kamada K, Kamoshima Y, Yamaguchi S, Tajima T, Tsubaki J, Fujimaki T (2008) Role of surgery for optic pathway/ hypothalamic astrocytomas in children. Neuro-oncology 10(5):725– 733 18. Sharif S, Ferner R, Birch JM, Gillespie JE, Gattamaneni HR, Baser ME, Evans DGR (2006) Second primary tumors in neurofibromatosis I patients treated for optic glioma: substantial risks after radiotherapy. Journal of Clinical Oncology 24(16):2570–2575. doi:10.1200/jco. 2005.03.8349

Childs Nerv Syst 19. Shuper A, Kornreich L, Michowitz S, Schwartz M, Yaniv I, Ij C (2000) Visual pathway tumors and hydrocephalus. Pediatric Hematology and Oncology 17(6):463–468 20. British Neuro-oncology Society/National cancer Action team (2011) Guidelines on the diagnosis and management of Optic Pathway Glioma (OPG): Rare Brain and CNS Tumour Guidelines. http:// www.bnos.org.uk/documents/rare_tumours_guidelines/Optic% 20pathway%20glioma%20guidelines.pdf. Accessed 19 Feb 2013 21. Sylvester CL, Drohan LA, Sergott RC (2006) Optic-nerve gliomas, chiasmal gliomas and neurofibromatosis type 1. Current Opinion in Ophthalmology 17(1):7–11. doi:10.1097/01.icu.0000193070.32369. 37 22. Tacke U, Karger D, Spreer J, Berlis A, Nikkhah G, Korinthenberg R (2011) Incidence of vasculopathy in children with hypothalamic/ chiasmatic gliomas treated with brachytherapy. Childs Nervous System 27(6):961–966. doi:10.1007/s00381-010-1370-0

23. Valdueza J, Lohmann F, Dammann O, Hagel C, Eckert B, Freckmann N (1994) Analysis of 20 primarily surgically treated chiasmatic/ hypothalamic pilocytic astrocytomas. Acta Neurochirurgica 126(1): 44–50 24. von Hornstein S, Kortmann R, Pietsch T, Emser A, Warmuth-Metz M, Soerensen N, Straeter R, Graf N, Thieme B, Gnekow A (2011) Impact of chemotherapy on disseminated low-grade glioma in children and adolescents: report from the HIT-LGG 1996 Trial. Pediatric Blood & Cancer 56(7):1046–1054 25. Wilne S, Collier J, Kennedy C, Koller K, Grundy R, Walker D (2007) Presentation of childhood CNS tumours: a systematic review and meta-analysis. Lancet Oncology 8(8):685–695. doi:10.1016/s14702045(07)70207-3 26. Zoeller G, Braithwaite C, Sandberg D (2010) Malignant transformation of an optic pathway glioma without prior radiation therapy. Case report. Journal of Neurosurgery pediatrics 5(5):507–510