- Email: [email protected]
Pilocytic astrocytoma: A retrospective study of 32 cases
Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
Contents lists available at SciVerse ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Pilocytic astrocytoma: A retrospective study of 32 cases Souki Cyrine a , Ziadi Sonia a,∗ , Trimeche Mounir a , Sriha Badderedine a , Tlili Kalthoum c , Krifa Hedi b , Mokni Moncef a a b c
Department of Pathology, University Hospital Farhat Hached, 4000 Sousse, Tunisia Department of Neurosurgery, University Hospital Sahloul, 4054 Sousse, Tunisia Department of Radiology, University Hospital Sahloul, 4054 Sousse, Tunisia
a r t i c l e
i n f o
Article history: Received 11 September 2012 Received in revised form 11 November 2012 Accepted 17 November 2012 Available online 21 December 2012 Keywords: Pilocytic astrocytoma Pathology Prognostic factors
a b s t r a c t Pilocytic astrocytoma (PA) is a neoplasia which is considered as a grade I astrocytoma by the World Health Organization (WHO). Its most common location is the cerebellum and it develops during the first two decades of life. Prognosis is mostly excellent if gross-total resection can be achieved, with 10-year survival rates of up to 95%. In rare cases, however, the patient has a bad outcome. Our aims were to retrospectively describe the clinicopathological features of 32 PAs, and identify factors that may be associated with aggressive behavior. The study included 21 males and 11 females with a median age of 10.5 years. Tumors demonstrated predilection for infratentorial location (74.9%), especially the cerebellum (59.3%), followed by cerebral ventricles (15.6%), supratentorial location (12.5%) and optic pathway (3.12%). Gross total resection was achieved in 14 tumors only. On histopathology, moderate cellularity (68.7%), microcystic changes (71.9%), Rosenthal fibers (62.5%) and eosinophilic granular bodies (53.2%) were present in the majority of cases. Atypia was present in 62.5% of cases, while endothelial proliferation and necrosis was noted in 3 and 2 cases, respectively. Median follow-up for all patients was 24 months. Four patients died in the postoperative period, one of whom was 62-year-old men and two others had brainstem location or invasion. Recurrence was observed in a 56-year-old patient whom first tumor was locally invasive. The patient died after the second surgery and anaplastic features was found in the recurrent tumor without previous radiotherapy. PA is a benign tumor, but some clinicopathological factors, such as partial resection, brainstem location and adult age have a worse prognosis. © 2012 Published by Elsevier B.V.
1. Introduction
2. Materials and methods
Pilocytic astrocytoma (PA) is a rare slowly growing glioma, classified as Grade I by the World Health Organization (WHO) occurring typically in children and young adults [1,2]. Most arise in the cerebellum, particularly in pediatric population; but can also occur in the cerebrum, diencephalon, brainstem, and optic pathway [1,3–5]. Generally, they are regarded as benign tumors in which gross total resection is often curative with a long-term survival greater than 95% [4,6]. Despite the very good prognosis, the outcome is not always favorable, mainly because of surgical morbidity and tumor recurrence. The behavior of this tumor after complete or incomplete resection is unpredictable. As well, pilocytic astrocytomas with anaplastic features are uncommon.
We carried out a retrospective study of 32 cases of pilocytic astrocytoma diagnosed in the Department of Pathology of the Farhat Hached University Hospital, Sousse, and registered in the Cancer Registry of the Center of Tunisia during 27-year period time (January 1984–December 2011). Medical records for each patient were reviewed to determine the sex, age at initial presentation, time to consultation, presenting symptoms and signs and treatment modalities (surgery or surgery and radiation therapy). The extent of resection of resection was estimated by analyzing the surgical reports and the results of postoperative neuroimaging. All results of neuroimaging (computed tomographic and/or magnetic resonance imaging scans) performed during the period of followup were obtained. They were judged on the presence or absence of residual or recurring tumor. Pathology materials from all biopsies and surgical procedures were reviewed to confirm diagnosis using established WHO criteria. Gross characteristics noted included size, consistency, color of surgical specimens, hemorrhagic and/or necrotic foci. Slides
∗ Corresponding author. Tel.: +216 98 633 982; fax: +216 73 226 702. E-mail address: [email protected] (Z. Sonia). 0303-8467/$ – see front matter © 2012 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.clineuro.2012.11.009
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225 Table 2 Correlation between tumor location and events.
Table 1 Clinical features of 32 patients with pilocytic astrocytoma. Patient characteristics Age (years) 0–10 11–20 21–30 >30 Median (range) Sex Male Female Time to consultation Median (range) Type of treatment Surgery Surgery + radiation Extent of resection Gross total resection Subtotal resection Biopsy Time to follow-up Median (range) Outcome Died Alive with tumor No evidence of disease Lost to follow-up
Number of cases (%)
1221
Events, n (%)
16 9 4 3 10.5 (1–62)
2 (12.5) 1 (11.1) 0 (0) 2 (66.7)
21 (65) 11 (34)
5 (23.8) 0 (0)
Tumor location
Number of cases (%)
Infratentorial Cerebellum Brainstem Spinal cord Supratentorial Hemisphere Corpus callosum Optic pathway Cerebral ventricles
24 (75) 19 (59.3) 4 (12.5) 1 (3.12) 4 (12.5) 3 (9.37) 1 (3.12) 1 (3.12) 5 (15.6)
Events, n (%) 3 (15.7) 2 (50)
1 (100) 1 (20)
7 months (15–36) 21 (95.4) 1 (4.5)
4 (19) 1 (100)
13 (59) 7 (31.8) 2 (9)
1 (7.7) 3 (43) 1 (50)
24 months (1–84) 5 3 11 13
were examined for cellular density and types, microcyst formation, Rosenthal fibers, eosinophilic granular bodies, vascularization, vascular hyalinization, chronic inflammation, microcalcification, hemorrhage, endothelial proliferation and necrosis. The evaluation also included an estimation of nuclear atypia and mitotic index. Glial fibrillary acidic protein (GFAP) immunostaining was performed at the time of initial diagnostic workup in a subset of cases. The clinical status of the patient at the end of the followup period was noted. Survival was estimated using Kaplan Meier method. Overall survival (OS) was defined as the time interval from diagnosis to death from any cause or, for patients remaining alive, the time interval from diagnosis to the last follow-up. Disease-free survival (DFS) was defined as the time interval from treatment to the time of disease progression or recurrence, to the last follow-up, or to death occurrence from any cause.
Cerebellum (59.3%) was the most common infratentorial location. Seventeen PA tumors (65.3%) were cystic with solid mural nodules. The remaining ones were either predominantly cystic (19.2%) or solid (15.4%). Size of tumors varied from 1.5 to 6 cm. Hydrocephalus was detected in 14 cases (53.8%) and 2 patients showed signs of herniation. Information regarding treatment modalities was available in 22 patients. Gross total resection was achieved in 13 tumors (59%); 7 (31.8%) were subtotally resected and 2 patients underwent biopsy only. This later procedure was indicated in the oldest and the youngest patients of our series for neurological status and general condition deterioration. The first one received radiation therapy and died 2 months later. The medical record of the second 1-year-old patient was lost.
3.2. Histopathological features Histological features are illustrated in Fig. 1. The majority of PAs showed a moderate cellularity (68.7%) and microcystic changes (72%). Rosenthal fibers and eosinophilic granular bodies were present in 62.5% and 53% of tumors, respectively. Vascularization was prominent in 75% of cases and vascular hyalinization observed in 12 cases (37.5%). Endothelial proliferation was present focally in 2 cases and extensively in one case. Three tumors (9.3%) had conspicuous pleomorphic cells and low or moderate degree of atypia was noted in 17 cases (53.1%). Mitotic figures were absent in all tumors. Rare necrotic foci were present in 2 tumors (6.2%). Calcification, inflammation and foci of chronic hemorrhage were rarely seen.
3. Results 3.3. Follow-up 3.1. Clinical features In this study, data pertaining to survival and tumor characteristics were collected from 32 patients (Table 1). The male/female ratio was nearly 2 (21 males/11 females) with a median age at diagnosis of 10.5 years (range 1–62 years). The main presenting symptoms at the time of diagnosis were headache (78%), visual disturbances (43%), vomiting (37%) and gait disturbance (25%). Time to consultation varied from 15 to 36 months (median 7 months). At physical examination, 46% of patients presented with signs of increased intracranial hypertension, 40% with cerebellar syndrome and 28% had papillary edema. Skull X-ray was carried out in 6 patients. Although not specific, this exam allowed in all cases to detect classic signs of mass effect relative to an intracranial expanding process. More specific preoperative neuroimaging studies were performed in 26 patients and comprised 20 computed tomography (CT) and 12 magnetic resonance imaging (MRI) scans. The locations for the PA tumors are listed in Table 2. In our patient population, infratentorial location (75%) was the most common site, followed by cerebral ventricles (15.6%), supratentorial location (12.5%) and optic pathway (3.1%).
The mean follow-up period, available in 19 patients, was 29.7 months (median 24 months). Four patients died in the postoperative period; two of them had brainstem tumor location or invasion. The third patient was the oldest one who had not been operated due to his neurological status and general condition deterioration. The last patient died from unrelated septicemia of digestive origin. Tumor progression was observed in one of the remaining patients. This was the second oldest patient of 56 year old, who presented with a right cerebellar tumor and underwent a subtotal resection due to V4 invasion (Fig. 2A). This event supervened 29 months after first surgery. MRI scan showed a recurrent right cerebellar tumor extending to the V4 and accompanied by a second pineal localization (Fig. 2C and D). Pathological exam of surgical material (tumor recurrence) showed more densely cellular compartments with extensive ischemic necrotic foci, endothelial proliferation and variable degree of atypia (Fig. 3). The patient died in the postoperative period. The Kaplan Meier estimates of DFS and OS, displayed in Fig. 4, show the median DFS was 60.9 months (95% CI, 43.7–78.1 months) and median OS was 61.2 months (95% CI, 44.1–78.2 months).
1222
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
Fig. 1. Representative images of histologic features in PAs assessed in the current study. (A) Microcystic changes (hematoxylin-eosin (HE), ×100); (B) Rosenthal fibers (black arrow) (HE ×400); (C) endothelial proliferation (white arrow) (HE ×200); (D) moderate degree of atypia (HE ×400).
4. Discussion Pilocytic astroytomas account for up to 25% of all brain tumors encountered in pediatric neurosurgical practice but overall, only 2.3% of all brain tumors are classified as PA [1,7]. The annual incidence for all ages has been reported to be 2.9–4.8 cases per million people, and no gender predilection has been shown [10,11]. This is in contrast to the male predominance noted in our series which was similar to that reported by Malik and Komotar [5]. PAs commonly occur in the first two decades of life with few cases being reported in the >30-year age group, while occurrence beyond 50 years is exceptional [5]. In the current study, 3 (9%) patients were more than 30-year-old, with 2 (6%) being in the >50-year age group. The cerebellum and the region around the third ventricle are the most common sites of origin, although the entire neuraxis can be affected with preference for the optic nerve, optic chiasma, hypothalamus, cerebellum, brain stem, thalamus, basal ganglia and cerebral hemispheres [1,3–5]. Some studies found a correlation between age and tumor location [7–9]. Burkhard et al. [8] noticed that the majority of PAs in children were located in the cerebellum (67%), whereas those in adults most frequently involved supratentorial structures (55%). Our results showed a slight difference in frequency between children (10% supratentorial, 50% cerebellar) and adults (28% supratentorial, 42.8% cerebellar). However, ventricular location was predominant in children (20% vs. 0% in adults) and brainstem tumors were equally distributed in the two age groups.
There are no clinical features that are unique to these tumors; signs and symptoms are usually of several months duration and directly related to the size, location and presence of associated hydrocephalus [10]. These symptoms are commonly headache, vomiting, mental change, gait ataxia, tremors, and dysmetria [9]. A spontaneous hemorrhage occurs in 8% of cases [9]. On neuroimaging, the stereotypical PA is described as a cystic cerebellar mass, with an enhancing mural nodule. Murray et al. [10] described 4 predominant imaging patterns: a mass with a nonenhancing cyst and an intensely enhancing mural nodule (21% of cases); a mass with an enhancing cyst wall and an intensely enhancing mural nodule (46%); a necrotic mass with a central nonenhancing zone (16%); and a predominantly solid mass with minimal to no cyst-like component (17%). Pilocytic astroytomas has a noteworthy benign biologic behavior that translates into an extremely high survival rate (94% at 10 years) that is by far the best of any glial tumor [11]. Several studies have shown that the extent of resection and tumor location are the main prognostic factors. PAs in eloquent and deep locations such as optic pathway, brainstem, and diencephalon present greater challenges in terms of accessibility and total resection. Surgery for tumors in these regions is typically subtotal in order to avoid complications [1]. In this regard, brainstem tumors or cerebellar PAs involving the brainstem, which Fernandez et al. referred to as transitional form, carry a worse prognosis than pure cerebellar astrocytomas, mainly because of the extent of resection [3]. As well, PAs involving the optic pathways were found to be associated with worse event-free survival (EFS) compared with those arising in other locations [1,3]. Another critical distinction in terms of tumor
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
1223
Fig. 2. (A) At the time of initial diagnosis CT scan showing an enhancing mass within cerebellum; (B) MR image showed no contrast enhancement at the surgical site; (C) MRI scan showed a recurrent right cerebellar tumor extending to the V4 29 months later and accompanied by a second pineal localization (D).
location reported is between PA of the anterior visual pathways and tumors that involve the hypothalamic/chiasmatic pathway. The later have a higher rate of local recurrence than tumors affecting the optic nerve and the orbit [12] for which complete surgical resection is usually successful but not feasible for disease that has progressed into or beyond the chiasma [13]. Two of four patients presenting with brainstem PAs died at the postoperative period. Data on patient with optic nerve tumor was unavailable in the medical record. Many studies noted that the recurrence rate in PA in any location is highly dependent on the success of surgical resection. A total resection led to 5- and 10-year survival rates of 100% and recurrence rates of 2–5.4%, whereas 42–45% of partially removed PAs have been demonstrated to recur [3]. However, the clinical course after subtotal removal appears to be less predictable since long-term stable disease and even partial involution have been also described [7,11]. This spontaneous regression may also occur apart from any prior surgery, especially in children with neurofibromatosis 1. Numerous theories have been proposed for this finding and include postsurgical apoptosis, host immune reaction, thrombosis or infarction of tumor vessels, alteration of growth kinetics, removal of the offending carcinogen, genetic programming, and
hormonal factors [11]. Our study also confirms the critical importance of the extent of surgery as a determinant of survival and recurrence since evolution in 6 of 7 patients who underwent subtotal resection was characterized by death (3 patients) or neurological sequelea (3 patients). The seventh patient, for whom tumor was located at the cerebellar hemisphere, was lost to follow. PA is histologically characterized by a biphasic pattern with alternating densely and loosely packed areas. Loose areas contain protoplasmic cells, microcysts, and granular bodies. Compact areas are composed of bipolar cells with presence of Rosenthal fibers. Other features that are variably present include sclerotic vessels, vascular proliferation, calcification and chronic inflammation, while nuclear pleomorphism, mitosis and necrosis are rare [5,14]. Some PAs with unusual histological features may show early local recurrence or dissemination, necessitating adjuvant therapy [12]. Elements that separate aggressive PAs from other tumors are not clearly identified. Indeed, the malignancy criteria usually used in grading diffuse gliomas are irrelevant to the patient’s prognosis [3]. Necrosis may be present in up to 10.4% of all PAs but these are generally infarct-like, associated with low proliferative activity and lack of perinecrotic palissading similar to that observed in the current series. Generally, it is said to be of no prognostic significance,
1224
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
Fig. 3. Photomicrographs of tumor specimens from patient with recurrent cerebellar tumor showing high-grade areas. (A) Densely cellular compartments (HE ×100); (B) extensive ischemic necrotic foci (HE ×100); (C) endothelial proliferation (HE ×100); (D) variable degree of atypia (HE ×200).
even though it was associated with decreased event-free survival in some studies [1,4,5]. The same is true of only occasional mitoses and microvascular proliferation. Nevertheless, in their study, Rodriguez et al. found that mitotic activity exceeding 4/10 HPF was of prognostic significance, distinguishing PA with anaplastic features from conventional, WHO grade I PAs [4]. Published descriptions of anaplastic features in PA are rare and have been reported most often in recurrent pilocytic astrocytomas after radiation therapy
[4,6]. The 2007 WHO criteria for making such a diagnosis remain poorly defined, but include pseudopalisading necrosis, high cellularity, cytologic anaplasia, endothelial proliferation, and multiple mitoses per HPF. Some of these criteria were found in the recurrent PA of our 56-year-old patient whose tumor invaded the fourth ventricle and was accompanied by a second pineal location. These anaplastic features developed without previous radiotherapy. The term “atypical pilocytic astrocytoma” has been advocated by some authors, but is not currently sanctioned by the WHO [2]. Whether
Fig. 4. Kaplan Meier estimates of overall survival (A) and disease-free survival (B).
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
such an intermediate category is justified remains uncertain, but unfortunately there is currently insufficient data to clearly define it. In the same way, association between the proliferation indices and outcomes was proved by some studies and not by others [1,3,7,15]. The significance of p53 status has been found, in some series, to correlate with tumor proliferative index. In contrast, other studies failed to detect either p53 immunoreactivity or TP53 mutations in PAs [1]. Several other factors were studied and found to be associated with a worse patient prognosis although controversies still exist. These include oligodendroglioma-like histology [1], multicentric disease, CSF dissemination [3,10], calcification and vascular hyalinization [1]. The two later factors are rather felt to reflect tumor chronicity or degenerative changes [1]. On neuroimaging, solid cerebellar PAs are presumed to be more aggressive than cystic PAs [16], essentially because cystic PAs are the most favorable type for achieving total resection [3,17]. The majority of the available data on PA stem from pediatric series. Only very few studies to date have focused on PA in adults. Several authors have suggested that PA in this age group take a relatively benign course [7]. In contrast, Stüer et al. [7] who reported a retrospective study of 44 adult patients (>16 years) noticed that 30% presented tumor recurrence or progression and 18% died after a median follow-up of 55 months. Increased proliferative activity and anaplastic histological features was present in a surprisingly large number of patients. Our results are consistent with these findings since death occurred in 42% of our adult patients in comparison with the rate of 16% observed in 16-year-old age group. In summary, this study shows that the survival rates for patients with PA are excellent and the prognosis is related mainly to the extent of surgical excision that depends widely of the location of tumor. Prognostic value of other specific histopathological features is still controversial and necessitates multinational and multiinstitutional studies that can accrue sufficient number of patients in a reasonably short period. The identification of such features may improve our ability to determine which PAs are more likely to exhibit clinical progression and require more vigilant observation or even an adjuvant treatment. Competing interests The authors declare that they have no competing interests.
1225
References [1] Tibbetts KM, Emnett RJ, Gao F, Perry A, Gutmann DH, Leonard JR. Histopathologic predictors of pilocytic astrocytoma event-free survival. Acta Neuropathologica 2009;117(6):657–65. [2] Scheithauer BW, Hawkins C, Tihan T, Vandenberg SR, Burger PC. Pilocytic astrocytoma. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, editors. WHO classification of tumours of the central nervous system. 4th ed. Lyon, France: IARC; 2007. p. 14–21. [3] Fernandez C, Figarella-Branger D, Girard N, Bouvier-Labit C, Gouvernet J, Paz Paredes A, et al. Pilocytic astrocytomas in children: prognostic factors—a retrospective study of 80 cases. Neurosurgery 2003;53(3):544–53. [4] Rodriguez FJ, Scheithauer BW, Burger PC, Jenkins S, Giannini C. Anaplasia in pilocytic astrocytoma predicts aggressive behavior. American Journal of Surgical Pathology 2010;34(2):147–60. [5] Malik A, Deb P, Sharma MC, Sarkar C. Neuropathological spectrum of pilocytic astrocytoma—an Indian series of 120 cases. Pathology Oncology Research 2006;12(3):164–71. [6] Shibahara I, Kawaguchi T, Kanamori M, Yonezawa S, Takazawa H, Asano K, et al. Pilocytic astrocytoma with histological malignant features without previous radiation therapy—case report. Neurologia Medico-Chirurgica 2011;51(2):144–7. [7] Stüer C, Vilz B, Majores M, Becker A, Schramm J, Simon M. Frequent recurrence and progression in pilocytic astrocytoma in adults. Cancer 2007;110(12):2799–808. [8] Burkhard C, Di Patre PL, Schüler D, Schüler G, Yas¸argil MG, Yonekawa Y, et al. A population-based study of the incidence and survival rates in patients with pilocytic astrocytoma. Journal of Neurosurgery 2003;98(6):1170–4. [9] Kim MS, Kim SW, Chang CH, Kim OL. Cerebellar pilocytic astrocytomas with spontaneous intratumoral hemorrhage in adult. Journal of Korean Neurosurgical Society 2011;49(6):363–6. [10] Murray RD, Penar PL, Filippi CG, Tarasiewicz I. Radiographically distinct variant of pilocytic astrocytoma: a case series. Journal of Computer Assisted Tomography 2011;35(4):495–7. [11] Koeller KK, Rushing EJ. From the archives of the AFIP: pilocytic astrocytoma: radiologic—pathologic correlation. Radiographics 2004;24(6):1693–708. [12] Tihan T, Ersen A, Qaddoumi I, Sughayer MA, Tolunay S, Al-Hussaini M, et al. Pathologic characteristics of pediatric intracranial pilocytic astrocytomas and their impact on outcome in 3 countries: a multi-institutional study. American Journal of Surgical Pathology 2012;36(1):43–55. [13] Hoyama E, Cruz AA, Colli BO, Matos JR, Chahud F. Isolated low grade pilocytic astrocytoma of the optic nerve in the elderly: case report. Arquivos Brasileiros de Oftalmologia 2008;71(1):97–100. [14] Taillibert S, Pedretti M, Sanson M. Current classification of gliomas. Presse Medicale 2004;33(18):1274–7. [15] Bowers DC, Gargan L, Kapur P, Reisch JS, Mulne AF, Shapiro KN, et al. Study of the MIB-1 labeling index as a predictor of tumor progression in pilocytic astrocytomas in children and adolescents. Journal of Clinical Oncology 2003;21(15):2968–73. [16] Strong JA, Hatten Jr HP, Brown MT, Debatin JF, Friedman HS, Oakes WJ, et al. Pilocytic astrocytoma: correlation between the initial imaging features and clinical aggressiveness. American Journal of Roentgenology 1993;161(2):369–72. [17] Klein DM, McCullough DC. Surgical staging of cerebellar astrocytomas in childhood. Cancer 1985;56(Suppl. 7):1810–1.
Contents lists available at SciVerse ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Pilocytic astrocytoma: A retrospective study of 32 cases Souki Cyrine a , Ziadi Sonia a,∗ , Trimeche Mounir a , Sriha Badderedine a , Tlili Kalthoum c , Krifa Hedi b , Mokni Moncef a a b c
Department of Pathology, University Hospital Farhat Hached, 4000 Sousse, Tunisia Department of Neurosurgery, University Hospital Sahloul, 4054 Sousse, Tunisia Department of Radiology, University Hospital Sahloul, 4054 Sousse, Tunisia
a r t i c l e
i n f o
Article history: Received 11 September 2012 Received in revised form 11 November 2012 Accepted 17 November 2012 Available online 21 December 2012 Keywords: Pilocytic astrocytoma Pathology Prognostic factors
a b s t r a c t Pilocytic astrocytoma (PA) is a neoplasia which is considered as a grade I astrocytoma by the World Health Organization (WHO). Its most common location is the cerebellum and it develops during the first two decades of life. Prognosis is mostly excellent if gross-total resection can be achieved, with 10-year survival rates of up to 95%. In rare cases, however, the patient has a bad outcome. Our aims were to retrospectively describe the clinicopathological features of 32 PAs, and identify factors that may be associated with aggressive behavior. The study included 21 males and 11 females with a median age of 10.5 years. Tumors demonstrated predilection for infratentorial location (74.9%), especially the cerebellum (59.3%), followed by cerebral ventricles (15.6%), supratentorial location (12.5%) and optic pathway (3.12%). Gross total resection was achieved in 14 tumors only. On histopathology, moderate cellularity (68.7%), microcystic changes (71.9%), Rosenthal fibers (62.5%) and eosinophilic granular bodies (53.2%) were present in the majority of cases. Atypia was present in 62.5% of cases, while endothelial proliferation and necrosis was noted in 3 and 2 cases, respectively. Median follow-up for all patients was 24 months. Four patients died in the postoperative period, one of whom was 62-year-old men and two others had brainstem location or invasion. Recurrence was observed in a 56-year-old patient whom first tumor was locally invasive. The patient died after the second surgery and anaplastic features was found in the recurrent tumor without previous radiotherapy. PA is a benign tumor, but some clinicopathological factors, such as partial resection, brainstem location and adult age have a worse prognosis. © 2012 Published by Elsevier B.V.
1. Introduction
2. Materials and methods
Pilocytic astrocytoma (PA) is a rare slowly growing glioma, classified as Grade I by the World Health Organization (WHO) occurring typically in children and young adults [1,2]. Most arise in the cerebellum, particularly in pediatric population; but can also occur in the cerebrum, diencephalon, brainstem, and optic pathway [1,3–5]. Generally, they are regarded as benign tumors in which gross total resection is often curative with a long-term survival greater than 95% [4,6]. Despite the very good prognosis, the outcome is not always favorable, mainly because of surgical morbidity and tumor recurrence. The behavior of this tumor after complete or incomplete resection is unpredictable. As well, pilocytic astrocytomas with anaplastic features are uncommon.
We carried out a retrospective study of 32 cases of pilocytic astrocytoma diagnosed in the Department of Pathology of the Farhat Hached University Hospital, Sousse, and registered in the Cancer Registry of the Center of Tunisia during 27-year period time (January 1984–December 2011). Medical records for each patient were reviewed to determine the sex, age at initial presentation, time to consultation, presenting symptoms and signs and treatment modalities (surgery or surgery and radiation therapy). The extent of resection of resection was estimated by analyzing the surgical reports and the results of postoperative neuroimaging. All results of neuroimaging (computed tomographic and/or magnetic resonance imaging scans) performed during the period of followup were obtained. They were judged on the presence or absence of residual or recurring tumor. Pathology materials from all biopsies and surgical procedures were reviewed to confirm diagnosis using established WHO criteria. Gross characteristics noted included size, consistency, color of surgical specimens, hemorrhagic and/or necrotic foci. Slides
∗ Corresponding author. Tel.: +216 98 633 982; fax: +216 73 226 702. E-mail address: [email protected] (Z. Sonia). 0303-8467/$ – see front matter © 2012 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.clineuro.2012.11.009
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225 Table 2 Correlation between tumor location and events.
Table 1 Clinical features of 32 patients with pilocytic astrocytoma. Patient characteristics Age (years) 0–10 11–20 21–30 >30 Median (range) Sex Male Female Time to consultation Median (range) Type of treatment Surgery Surgery + radiation Extent of resection Gross total resection Subtotal resection Biopsy Time to follow-up Median (range) Outcome Died Alive with tumor No evidence of disease Lost to follow-up
Number of cases (%)
1221
Events, n (%)
16 9 4 3 10.5 (1–62)
2 (12.5) 1 (11.1) 0 (0) 2 (66.7)
21 (65) 11 (34)
5 (23.8) 0 (0)
Tumor location
Number of cases (%)
Infratentorial Cerebellum Brainstem Spinal cord Supratentorial Hemisphere Corpus callosum Optic pathway Cerebral ventricles
24 (75) 19 (59.3) 4 (12.5) 1 (3.12) 4 (12.5) 3 (9.37) 1 (3.12) 1 (3.12) 5 (15.6)
Events, n (%) 3 (15.7) 2 (50)
1 (100) 1 (20)
7 months (15–36) 21 (95.4) 1 (4.5)
4 (19) 1 (100)
13 (59) 7 (31.8) 2 (9)
1 (7.7) 3 (43) 1 (50)
24 months (1–84) 5 3 11 13
were examined for cellular density and types, microcyst formation, Rosenthal fibers, eosinophilic granular bodies, vascularization, vascular hyalinization, chronic inflammation, microcalcification, hemorrhage, endothelial proliferation and necrosis. The evaluation also included an estimation of nuclear atypia and mitotic index. Glial fibrillary acidic protein (GFAP) immunostaining was performed at the time of initial diagnostic workup in a subset of cases. The clinical status of the patient at the end of the followup period was noted. Survival was estimated using Kaplan Meier method. Overall survival (OS) was defined as the time interval from diagnosis to death from any cause or, for patients remaining alive, the time interval from diagnosis to the last follow-up. Disease-free survival (DFS) was defined as the time interval from treatment to the time of disease progression or recurrence, to the last follow-up, or to death occurrence from any cause.
Cerebellum (59.3%) was the most common infratentorial location. Seventeen PA tumors (65.3%) were cystic with solid mural nodules. The remaining ones were either predominantly cystic (19.2%) or solid (15.4%). Size of tumors varied from 1.5 to 6 cm. Hydrocephalus was detected in 14 cases (53.8%) and 2 patients showed signs of herniation. Information regarding treatment modalities was available in 22 patients. Gross total resection was achieved in 13 tumors (59%); 7 (31.8%) were subtotally resected and 2 patients underwent biopsy only. This later procedure was indicated in the oldest and the youngest patients of our series for neurological status and general condition deterioration. The first one received radiation therapy and died 2 months later. The medical record of the second 1-year-old patient was lost.
3.2. Histopathological features Histological features are illustrated in Fig. 1. The majority of PAs showed a moderate cellularity (68.7%) and microcystic changes (72%). Rosenthal fibers and eosinophilic granular bodies were present in 62.5% and 53% of tumors, respectively. Vascularization was prominent in 75% of cases and vascular hyalinization observed in 12 cases (37.5%). Endothelial proliferation was present focally in 2 cases and extensively in one case. Three tumors (9.3%) had conspicuous pleomorphic cells and low or moderate degree of atypia was noted in 17 cases (53.1%). Mitotic figures were absent in all tumors. Rare necrotic foci were present in 2 tumors (6.2%). Calcification, inflammation and foci of chronic hemorrhage were rarely seen.
3. Results 3.3. Follow-up 3.1. Clinical features In this study, data pertaining to survival and tumor characteristics were collected from 32 patients (Table 1). The male/female ratio was nearly 2 (21 males/11 females) with a median age at diagnosis of 10.5 years (range 1–62 years). The main presenting symptoms at the time of diagnosis were headache (78%), visual disturbances (43%), vomiting (37%) and gait disturbance (25%). Time to consultation varied from 15 to 36 months (median 7 months). At physical examination, 46% of patients presented with signs of increased intracranial hypertension, 40% with cerebellar syndrome and 28% had papillary edema. Skull X-ray was carried out in 6 patients. Although not specific, this exam allowed in all cases to detect classic signs of mass effect relative to an intracranial expanding process. More specific preoperative neuroimaging studies were performed in 26 patients and comprised 20 computed tomography (CT) and 12 magnetic resonance imaging (MRI) scans. The locations for the PA tumors are listed in Table 2. In our patient population, infratentorial location (75%) was the most common site, followed by cerebral ventricles (15.6%), supratentorial location (12.5%) and optic pathway (3.1%).
The mean follow-up period, available in 19 patients, was 29.7 months (median 24 months). Four patients died in the postoperative period; two of them had brainstem tumor location or invasion. The third patient was the oldest one who had not been operated due to his neurological status and general condition deterioration. The last patient died from unrelated septicemia of digestive origin. Tumor progression was observed in one of the remaining patients. This was the second oldest patient of 56 year old, who presented with a right cerebellar tumor and underwent a subtotal resection due to V4 invasion (Fig. 2A). This event supervened 29 months after first surgery. MRI scan showed a recurrent right cerebellar tumor extending to the V4 and accompanied by a second pineal localization (Fig. 2C and D). Pathological exam of surgical material (tumor recurrence) showed more densely cellular compartments with extensive ischemic necrotic foci, endothelial proliferation and variable degree of atypia (Fig. 3). The patient died in the postoperative period. The Kaplan Meier estimates of DFS and OS, displayed in Fig. 4, show the median DFS was 60.9 months (95% CI, 43.7–78.1 months) and median OS was 61.2 months (95% CI, 44.1–78.2 months).
1222
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
Fig. 1. Representative images of histologic features in PAs assessed in the current study. (A) Microcystic changes (hematoxylin-eosin (HE), ×100); (B) Rosenthal fibers (black arrow) (HE ×400); (C) endothelial proliferation (white arrow) (HE ×200); (D) moderate degree of atypia (HE ×400).
4. Discussion Pilocytic astroytomas account for up to 25% of all brain tumors encountered in pediatric neurosurgical practice but overall, only 2.3% of all brain tumors are classified as PA [1,7]. The annual incidence for all ages has been reported to be 2.9–4.8 cases per million people, and no gender predilection has been shown [10,11]. This is in contrast to the male predominance noted in our series which was similar to that reported by Malik and Komotar [5]. PAs commonly occur in the first two decades of life with few cases being reported in the >30-year age group, while occurrence beyond 50 years is exceptional [5]. In the current study, 3 (9%) patients were more than 30-year-old, with 2 (6%) being in the >50-year age group. The cerebellum and the region around the third ventricle are the most common sites of origin, although the entire neuraxis can be affected with preference for the optic nerve, optic chiasma, hypothalamus, cerebellum, brain stem, thalamus, basal ganglia and cerebral hemispheres [1,3–5]. Some studies found a correlation between age and tumor location [7–9]. Burkhard et al. [8] noticed that the majority of PAs in children were located in the cerebellum (67%), whereas those in adults most frequently involved supratentorial structures (55%). Our results showed a slight difference in frequency between children (10% supratentorial, 50% cerebellar) and adults (28% supratentorial, 42.8% cerebellar). However, ventricular location was predominant in children (20% vs. 0% in adults) and brainstem tumors were equally distributed in the two age groups.
There are no clinical features that are unique to these tumors; signs and symptoms are usually of several months duration and directly related to the size, location and presence of associated hydrocephalus [10]. These symptoms are commonly headache, vomiting, mental change, gait ataxia, tremors, and dysmetria [9]. A spontaneous hemorrhage occurs in 8% of cases [9]. On neuroimaging, the stereotypical PA is described as a cystic cerebellar mass, with an enhancing mural nodule. Murray et al. [10] described 4 predominant imaging patterns: a mass with a nonenhancing cyst and an intensely enhancing mural nodule (21% of cases); a mass with an enhancing cyst wall and an intensely enhancing mural nodule (46%); a necrotic mass with a central nonenhancing zone (16%); and a predominantly solid mass with minimal to no cyst-like component (17%). Pilocytic astroytomas has a noteworthy benign biologic behavior that translates into an extremely high survival rate (94% at 10 years) that is by far the best of any glial tumor [11]. Several studies have shown that the extent of resection and tumor location are the main prognostic factors. PAs in eloquent and deep locations such as optic pathway, brainstem, and diencephalon present greater challenges in terms of accessibility and total resection. Surgery for tumors in these regions is typically subtotal in order to avoid complications [1]. In this regard, brainstem tumors or cerebellar PAs involving the brainstem, which Fernandez et al. referred to as transitional form, carry a worse prognosis than pure cerebellar astrocytomas, mainly because of the extent of resection [3]. As well, PAs involving the optic pathways were found to be associated with worse event-free survival (EFS) compared with those arising in other locations [1,3]. Another critical distinction in terms of tumor
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
1223
Fig. 2. (A) At the time of initial diagnosis CT scan showing an enhancing mass within cerebellum; (B) MR image showed no contrast enhancement at the surgical site; (C) MRI scan showed a recurrent right cerebellar tumor extending to the V4 29 months later and accompanied by a second pineal localization (D).
location reported is between PA of the anterior visual pathways and tumors that involve the hypothalamic/chiasmatic pathway. The later have a higher rate of local recurrence than tumors affecting the optic nerve and the orbit [12] for which complete surgical resection is usually successful but not feasible for disease that has progressed into or beyond the chiasma [13]. Two of four patients presenting with brainstem PAs died at the postoperative period. Data on patient with optic nerve tumor was unavailable in the medical record. Many studies noted that the recurrence rate in PA in any location is highly dependent on the success of surgical resection. A total resection led to 5- and 10-year survival rates of 100% and recurrence rates of 2–5.4%, whereas 42–45% of partially removed PAs have been demonstrated to recur [3]. However, the clinical course after subtotal removal appears to be less predictable since long-term stable disease and even partial involution have been also described [7,11]. This spontaneous regression may also occur apart from any prior surgery, especially in children with neurofibromatosis 1. Numerous theories have been proposed for this finding and include postsurgical apoptosis, host immune reaction, thrombosis or infarction of tumor vessels, alteration of growth kinetics, removal of the offending carcinogen, genetic programming, and
hormonal factors [11]. Our study also confirms the critical importance of the extent of surgery as a determinant of survival and recurrence since evolution in 6 of 7 patients who underwent subtotal resection was characterized by death (3 patients) or neurological sequelea (3 patients). The seventh patient, for whom tumor was located at the cerebellar hemisphere, was lost to follow. PA is histologically characterized by a biphasic pattern with alternating densely and loosely packed areas. Loose areas contain protoplasmic cells, microcysts, and granular bodies. Compact areas are composed of bipolar cells with presence of Rosenthal fibers. Other features that are variably present include sclerotic vessels, vascular proliferation, calcification and chronic inflammation, while nuclear pleomorphism, mitosis and necrosis are rare [5,14]. Some PAs with unusual histological features may show early local recurrence or dissemination, necessitating adjuvant therapy [12]. Elements that separate aggressive PAs from other tumors are not clearly identified. Indeed, the malignancy criteria usually used in grading diffuse gliomas are irrelevant to the patient’s prognosis [3]. Necrosis may be present in up to 10.4% of all PAs but these are generally infarct-like, associated with low proliferative activity and lack of perinecrotic palissading similar to that observed in the current series. Generally, it is said to be of no prognostic significance,
1224
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
Fig. 3. Photomicrographs of tumor specimens from patient with recurrent cerebellar tumor showing high-grade areas. (A) Densely cellular compartments (HE ×100); (B) extensive ischemic necrotic foci (HE ×100); (C) endothelial proliferation (HE ×100); (D) variable degree of atypia (HE ×200).
even though it was associated with decreased event-free survival in some studies [1,4,5]. The same is true of only occasional mitoses and microvascular proliferation. Nevertheless, in their study, Rodriguez et al. found that mitotic activity exceeding 4/10 HPF was of prognostic significance, distinguishing PA with anaplastic features from conventional, WHO grade I PAs [4]. Published descriptions of anaplastic features in PA are rare and have been reported most often in recurrent pilocytic astrocytomas after radiation therapy
[4,6]. The 2007 WHO criteria for making such a diagnosis remain poorly defined, but include pseudopalisading necrosis, high cellularity, cytologic anaplasia, endothelial proliferation, and multiple mitoses per HPF. Some of these criteria were found in the recurrent PA of our 56-year-old patient whose tumor invaded the fourth ventricle and was accompanied by a second pineal location. These anaplastic features developed without previous radiotherapy. The term “atypical pilocytic astrocytoma” has been advocated by some authors, but is not currently sanctioned by the WHO [2]. Whether
Fig. 4. Kaplan Meier estimates of overall survival (A) and disease-free survival (B).
S. Cyrine et al. / Clinical Neurology and Neurosurgery 115 (2013) 1220–1225
such an intermediate category is justified remains uncertain, but unfortunately there is currently insufficient data to clearly define it. In the same way, association between the proliferation indices and outcomes was proved by some studies and not by others [1,3,7,15]. The significance of p53 status has been found, in some series, to correlate with tumor proliferative index. In contrast, other studies failed to detect either p53 immunoreactivity or TP53 mutations in PAs [1]. Several other factors were studied and found to be associated with a worse patient prognosis although controversies still exist. These include oligodendroglioma-like histology [1], multicentric disease, CSF dissemination [3,10], calcification and vascular hyalinization [1]. The two later factors are rather felt to reflect tumor chronicity or degenerative changes [1]. On neuroimaging, solid cerebellar PAs are presumed to be more aggressive than cystic PAs [16], essentially because cystic PAs are the most favorable type for achieving total resection [3,17]. The majority of the available data on PA stem from pediatric series. Only very few studies to date have focused on PA in adults. Several authors have suggested that PA in this age group take a relatively benign course [7]. In contrast, Stüer et al. [7] who reported a retrospective study of 44 adult patients (>16 years) noticed that 30% presented tumor recurrence or progression and 18% died after a median follow-up of 55 months. Increased proliferative activity and anaplastic histological features was present in a surprisingly large number of patients. Our results are consistent with these findings since death occurred in 42% of our adult patients in comparison with the rate of 16% observed in 16-year-old age group. In summary, this study shows that the survival rates for patients with PA are excellent and the prognosis is related mainly to the extent of surgical excision that depends widely of the location of tumor. Prognostic value of other specific histopathological features is still controversial and necessitates multinational and multiinstitutional studies that can accrue sufficient number of patients in a reasonably short period. The identification of such features may improve our ability to determine which PAs are more likely to exhibit clinical progression and require more vigilant observation or even an adjuvant treatment. Competing interests The authors declare that they have no competing interests.
1225
References [1] Tibbetts KM, Emnett RJ, Gao F, Perry A, Gutmann DH, Leonard JR. Histopathologic predictors of pilocytic astrocytoma event-free survival. Acta Neuropathologica 2009;117(6):657–65. [2] Scheithauer BW, Hawkins C, Tihan T, Vandenberg SR, Burger PC. Pilocytic astrocytoma. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, editors. WHO classification of tumours of the central nervous system. 4th ed. Lyon, France: IARC; 2007. p. 14–21. [3] Fernandez C, Figarella-Branger D, Girard N, Bouvier-Labit C, Gouvernet J, Paz Paredes A, et al. Pilocytic astrocytomas in children: prognostic factors—a retrospective study of 80 cases. Neurosurgery 2003;53(3):544–53. [4] Rodriguez FJ, Scheithauer BW, Burger PC, Jenkins S, Giannini C. Anaplasia in pilocytic astrocytoma predicts aggressive behavior. American Journal of Surgical Pathology 2010;34(2):147–60. [5] Malik A, Deb P, Sharma MC, Sarkar C. Neuropathological spectrum of pilocytic astrocytoma—an Indian series of 120 cases. Pathology Oncology Research 2006;12(3):164–71. [6] Shibahara I, Kawaguchi T, Kanamori M, Yonezawa S, Takazawa H, Asano K, et al. Pilocytic astrocytoma with histological malignant features without previous radiation therapy—case report. Neurologia Medico-Chirurgica 2011;51(2):144–7. [7] Stüer C, Vilz B, Majores M, Becker A, Schramm J, Simon M. Frequent recurrence and progression in pilocytic astrocytoma in adults. Cancer 2007;110(12):2799–808. [8] Burkhard C, Di Patre PL, Schüler D, Schüler G, Yas¸argil MG, Yonekawa Y, et al. A population-based study of the incidence and survival rates in patients with pilocytic astrocytoma. Journal of Neurosurgery 2003;98(6):1170–4. [9] Kim MS, Kim SW, Chang CH, Kim OL. Cerebellar pilocytic astrocytomas with spontaneous intratumoral hemorrhage in adult. Journal of Korean Neurosurgical Society 2011;49(6):363–6. [10] Murray RD, Penar PL, Filippi CG, Tarasiewicz I. Radiographically distinct variant of pilocytic astrocytoma: a case series. Journal of Computer Assisted Tomography 2011;35(4):495–7. [11] Koeller KK, Rushing EJ. From the archives of the AFIP: pilocytic astrocytoma: radiologic—pathologic correlation. Radiographics 2004;24(6):1693–708. [12] Tihan T, Ersen A, Qaddoumi I, Sughayer MA, Tolunay S, Al-Hussaini M, et al. Pathologic characteristics of pediatric intracranial pilocytic astrocytomas and their impact on outcome in 3 countries: a multi-institutional study. American Journal of Surgical Pathology 2012;36(1):43–55. [13] Hoyama E, Cruz AA, Colli BO, Matos JR, Chahud F. Isolated low grade pilocytic astrocytoma of the optic nerve in the elderly: case report. Arquivos Brasileiros de Oftalmologia 2008;71(1):97–100. [14] Taillibert S, Pedretti M, Sanson M. Current classification of gliomas. Presse Medicale 2004;33(18):1274–7. [15] Bowers DC, Gargan L, Kapur P, Reisch JS, Mulne AF, Shapiro KN, et al. Study of the MIB-1 labeling index as a predictor of tumor progression in pilocytic astrocytomas in children and adolescents. Journal of Clinical Oncology 2003;21(15):2968–73. [16] Strong JA, Hatten Jr HP, Brown MT, Debatin JF, Friedman HS, Oakes WJ, et al. Pilocytic astrocytoma: correlation between the initial imaging features and clinical aggressiveness. American Journal of Roentgenology 1993;161(2):369–72. [17] Klein DM, McCullough DC. Surgical staging of cerebellar astrocytomas in childhood. Cancer 1985;56(Suppl. 7):1810–1.