Clinical manifestations

Presentation and course

The clinical manifestations of pilocytic astrocytomas in adults are similar to those of other brain tumors, except for their more indolent growth and onset of symptoms (142). Symptoms and signs develop due to the localized effects of tumor growth on specific regions of the brain, elevated intracranial pressure, or a combination of these factors. Pilocytic astrocytomas in adults are most commonly located in the cerebrum; however, they can occur anywhere along the midline axis of the brain, including the cerebellum, brainstem, third ventricle, hypothalamus, sellar region, and optic pathways (37; 56; 215; 169; 173; 49; 52; 91; 185). Pilocytic astrocytomas can develop in such diverse locations; as a result, the clinical evolution of signs and symptoms is variable. The most common symptoms are seizures, headaches, nausea, visual disturbances, and weakness (37; 56; 149; 52). The typical duration of symptoms before diagnosis is 2 to 4 years (37; 56). On occasion, patients have a more rapid clinical presentation and diagnosis (eg, days, weeks, or months) or have such a benign indolent course that the diagnosis is made after decades or as an incidental finding at autopsy. A rapid presentation and diagnosis is most common with the onset of seizure activity, especially if the seizures are generalized. In other cases, the pilocytic tumor can be the cause of longstanding intractable epilepsy that requires surgical resection for improved control (146). On neurologic examination, the most frequent signs are papilledema, hemianopsia and other visual disturbances, pyramidal tract signs, and weakness (37; 149; 52). The progression of neurologic signs is typically slow, although rare patients experience a fulminant course due to intratumoral hemorrhage or acute hydrocephalus (116).

In adults, pilocytic astrocytomas most commonly occur in the cerebrum. Although they can develop within any lobe, a predilection exists for the temporal and parietal lobes. At the time of diagnosis, these patients are usually young adults, with a reported mean age of 20 to 25 years (Table 1) (37; 56; 115; 54; 91). However, even patients in the sixth and seventh decades have been reported with pilocytic astrocytomas (02; 52). The typical young age of these patients is 8 to 12 years younger than the corresponding median age of 34 years for patients with grade 1 diffuse fibrillary astrocytoma of the cerebrum (56). The most common symptoms at presentation are either progressive headache or seizure activity, occurring in 40% to 60% of patients (37; 56; 149; 02; 52). The headaches are usually generalized and may be associated with nausea or emesis. The type of seizure will vary depending on the location of the tumor. Tumors that involve an anteromedial temporal or frontal lobe will usually cause partial complex seizures, with or without secondary generalization. Involvement of the thalamus or diencephalon may cause generalized tonic-clonic seizures. Tumors affecting the primary motor, sensory, and visual cortices will usually cause simple partial seizures that may secondarily generalize. Visual disturbances involving constriction of visual fields, reduced visual acuity, or diplopia are noted in less than 20% of patients. Symptomatic weakness of the extremities is present in 10% to 20% of patients, usually in the form of a hemiparesis. On neurologic examination, the most common sign is papilledema, found in 40% to 80% of patients (37; 56; 149). In one third of patients, hemianopsia and other visual abnormalities can be detected. Hemiparetic weakness and upper motor neuron signs are present in 20% of patients. Less common findings include cranial nerve palsies, dysphasia, and behavioral changes, each of which is noted in less than 10% of patients.

Table 1. Summary of Demographic, Treatment, and Survival Data from Published Reports of Adult Patients with Pilocytic Astrocytoma

n: 7 (57)

n: 23 (37)

n: 8 (115)

n: 11 (24)

n: 5 (54)

n: 8 (91)

Pilocytic astrocytomas of the cerebellum occur less frequently than their cerebral counterparts. They are located in 1 or both of the cerebellar hemispheres in 55% to 65% of patients, in the vermis in 15% to 20% of patients, and in a combination of these areas in 15% to 30% of patients (69). In general, a slight predilection exists for hemispheric involvement in adults. Invasion of the brainstem is noted in almost one third of cases. The age at diagnosis is between 20 and 40 years for most patients, although reports are available on cases that developed in the sixth, seventh, and eighth decades (69). In a study of 132 patients that compared patients having pilocytic astrocytomas with patients having diffuse cerebellar astrocytomas (25 pilocytic tumor patients over 20 years old), Hayostek and colleagues found the median ages to be significantly different (p less than 0.001): 12 versus 52 years, respectively (69). The most common symptoms of headache, nausea, and emesis, noted in greater than 95% of patients, are usually related to increased intracranial pressure (49; 69). Elevated pressure can be due to the presence of the tumor, hydrocephalus, or both. Primary cerebellar symptoms and signs (ie, ipsilateral limb dysmetria, dysarthria, and ataxia) also occur in greater than 90% of patients. In 30% to 40% of patients, visual disturbances are noted, usually manifesting as diplopia. Hemiparesis and meningismus are uncommon, occurring in 7% to 12% of patients. On neurologic examination, the most common findings are papilledema, gait ataxia, dysmetria, nystagmus, and abducens nerve palsy (49; 69); however, the majority of these patients (over 80%) remain fully active at home or work despite these deficits (69).

Optic pathway pilocytic astrocytomas include tumors that arise in the optic nerve, chiasm, or tract. They are uncommon in adults and tend to involve the chiasm and tract more than the optic nerve (21). These tumors are frequently associated with neurofibromatosis type 1. Pilocytic optic pathway tumors that occur in patients with neurofibromatosis generally have a less favorable outcome, as do tumors that develop in older individuals (21; 05). The symptoms and signs of optic pathway pilocytic tumors are slowly progressive, with occasional periods of spontaneous stabilization and indolence. Rarely, these tumors can have spontaneous involution after biopsy (10). Lesions of the intraorbital optic nerve present with painless ipsilateral proptosis and variable loss of vision. In addition, optic atrophy and strabismus may be noted. Intracranial tumors of the optic nerve or chiasm produce bilateral visual loss in greater than 90% of patients. Headaches, optic atrophy, strabismus, and papilledema may also occur. Tumors of the optic tract usually cause bilateral visual loss and headache. Infiltrative spread along the tracts is common, often involving the dorsal midbrain and brainstem or hypothalamus. A rare presentation of optic pathway pilocytic astrocytomas is subarachnoid hemorrhage, with acute headache, vomiting, and loss of consciousness (58).

Approximately 20% of brainstem gliomas are pilocytic astrocytomas, usually occurring in patients under 18 years of age (173). The tumors typically arise in the pons, with subsequent growth rostrally through the brainstem. Less often the mass is focal and may have an exophytic component. The most common presenting signs and symptoms are cranial nerve dysfunction (eg, double vision, strabismus, dysphonia, dysphagia, facial weakness) and cerebellar abnormalities (eg, ataxia, dysarthria, dysmetria), each noted in greater than 75% of patients (04). Symptoms of increased intracranial pressure, typically caused by hydrocephalus, are present in 65% to 70% of cases. In more than one half of the patients, pyramidal tract dysfunction is evident. Personality changes, sensory deficits, and disturbances of consciousness are each present in one quarter of patients. Growth of the tumor and clinical progression are generally slow. Rarely, pilocytic astrocytomas of the brainstem can arise in elderly patients (27). In the case reported by Burkhardt and coworkers, the patient was an 85-year-old male with Parkinsonian symptoms (27). At autopsy, the tumor was noted to have infiltrated into the substantia nigra in the anterior midbrain. None of the classic neuropathologic features of Parkinson disease were noted (eg, Lewy bodies).

Hypothalamic pilocytic astrocytomas typically develop in the periventricular region around the third ventricle. They are rare in adults and follow an indolent course similar to that in children. The most common presenting symptoms are referable to endocrine dysfunction, including diabetes insipidus, hypothyroidism, infertility, and hypogonadism. Less frequently, these tumors may cause hydrocephalus and associated pressure-related symptoms.

In rare cases, pilocytic astrocytomas may develop in the pituitary gland, sellar region, cerebellopontine angle, the meninges, and fornix (169; 16; 185; 19; 177). Tumors that arise in the pituitary gland or sellar region are clinically characterized by endocrine dysfunction and may include diabetes insipidus, hypothyroidism, infertility, hypogonadism, and hyperprolactinemia (169; 185). Pilocytic tumors arising from the posterior pituitary gland (ie, pituicytomas) can sometimes invade the skull base and erode into the sinuses (185). Pilocytic astrocytomas of the cerebellopontine angle can mimic acoustic schwannomas with symptoms of unilateral hearing loss, vertigo, and ataxia (16). In an extremely rare report, Skopelitou and colleagues describe a 22-year-old woman with a dermoid cyst of the ovary that contained a pilocytic astrocytoma (182). Similarly, Bohner and colleagues reported a 25-year-old woman with a pilocytic astrocytoma that developed primarily within the meninges and had a poor clinical course (19). Genetic analysis revealed a missense mutation within p53 and reduced expression of the PTEN tumor suppressor protein. On very rare occasions, pilocytic astrocytomas can arise from the fornix and mimic a colloid cyst (177).

Another rare case report from Munshey and colleagues describes a 47-year-old male who presented with a sudden onset of headache, nausea, and emesis and who was found to have a hemorrhagic right parietal-occipital mass (136). He also had a longer history of progressive lower back pain. An MRI of the spine revealed a mass in the sacral region. The cerebral and spinal masses were resected, and both were consistent with pilocytic astrocytoma. Hemorrhage at presentation, with the presence of a spinal drop metastasis, is very rare and unusual for a pilocytic astrocytoma. A similar report by Galgano and coworkers describes a 30-year-old woman who presented with spontaneous hemorrhage into a cerebellar pilocytic astrocytoma (55). In their literature review, some degree of hemorrhage is present in 5% to 8% of patients at presentation with pilocytic astrocytomas. A report by Prasad and colleagues reviewed the literature for cases of hemorrhagic presentation in adults with pilocytic astrocytomas (157). A total of 26 cases were noted, with a mean age of 37 years (range 21-75 years), and a male:female ratio of 2.1:1. The most common symptoms were sudden headache and emesis, visual changes, limb weakness, and seizures. A supratentorial location for the tumor was most common. All tumors had intratumoral hemorrhage, but intraventricular, subarachnoid, and subdural forms of hemorrhage were also noted.

Prognosis and complications

The overall prognosis for survival and intact neurologic function is good for adult patients with pilocytic astrocytomas compared to that for patients with other brain neoplasms; it is similar to, although not as favorable as, that for children with pilocytic tumors (56; 149; 02; 215; 49; 52; 25; 86). In several studies, the overall 5-, 10-, and 20-year survival rates have been estimated at 85%, 80%, and 75% respectively for supratentorial tumors, with slightly higher percentages for infratentorial tumors (Table 1) (198; 115; 49; 52; 69; 91; 64). In an analysis of adult patients with cerebral pilocytic tumors, the median survival for 6 of the 7 patients was 9 years (56). In a similar study by Katsetos and colleagues, which included 8 adult patients with cerebral lobar pilocytic tumors, the mean survival was 8 or more years (91). The majority of patients maintain adequate neurologic function that allows continued activity at home or work (56; 149; 02; 198; 49; 52). A study by Brown and associates prospectively followed 20 adult patients with supratentorial pilocytic astrocytomas (25). This study confirmed the favorable prognosis in adults with this tumor and noted an estimated 10-year survival rate of 95%. The majority of patients maintained stable neurologic function after surgical resection and did not require irradiation or chemotherapy. Although it is more common in children and patients with NF-1, spontaneous involution of pilocytic astrocytomas can occur in adults (10). This prompted the authors to recommend observation and expectant treatment in this group of patients. A study from the Mayo Clinic reviewed the SEER data on a series of 865 adult patients with pilocytic astrocytomas, from 1973 through 2008 (86). In this cohort, survival rates declined significantly with age, from 96.5% 5-year survival in patients 5 to 19 years of age, to 52.9% 5-year survival in adult patients 60 years of age or older. Overall, pilocytic astrocytomas were associated with a higher mortality in adult patients when compared to children and teenagers. A long-term follow-up report of data from the NCCTG-867251 study by Brown and colleagues suggests that pilocytic astrocytomas in adults will usually have a benign course (23). They followed 20 adult patients with pilocytic astrocytomas after biopsy and radiotherapy or observation after partial or complete resection for a median of 20.8 years. The 20-year time to progression and overall survival rates were 95% and 90%, respectively. Only 2 patients in the cohort had died during follow-up, and they had nontumor-related deaths. The authors recommended only using radiotherapy for salvage after a partial or complete resection.

In a small percentage of pilocytic astrocytomas, a more clinically aggressive course is seen (24; 183; 123; 184). In this group of patients, the tumor has a more rapid onset of symptoms and signs and is less responsive to standard therapy. In a study of 11 patients with "clinically aggressive" pilocytic astrocytomas, Brown and colleagues noted progression through surgical resection or radiation therapy after a median of only 9 months (range, 0.5 to 71 months) (24). After the addition of chemotherapy, the median time to progression was only 7.5 months. In a subgroup of 5 patients who died, the median survival from diagnosis was 29 months, significantly less than the expected survival of more typical pilocytic astrocytoma patients, and similar to that of patients with anaplastic astrocytoma (Table 1). In a report by Stuer and colleagues, of 44 adult patients with newly diagnosed or recurrent pilocytic astrocytomas, tumor recurrence or progression was noted in 14 (30%) (184). In addition, 8 of 44 patients (18%) died as a direct result of the tumor. The 5- and 10-year overall survival rates were only 87% and 77%, respectively. Similarly, the 5- and 10-year progression-free survival rates were only 72% and 67%, respectively. Factors that conferred a poor prognosis included subtotal resection, anaplastic pathologic features, and a Ki-67 labeling index greater than 5%. Kurwale and colleagues evaluated a series of 118 patients with pilocytic astrocytomas and divided them into nonrecurrent and early symptomatic recurrence cohorts (103). There was a trend for VEGF expression and endothelial proliferation to be more pronounced in the nonrecurrent patient group, but the difference did not reach statistical significance. Based on these results, the authors concluded that VEGF expression and degree of endothelial proliferation were not predictive for early recurrence. Some authors have termed these more aggressive pilocytic tumors, "anaplastic" pilocytic astrocytomas, when the tumor has typical features of a high-grade astrocytoma, along with the presence of Rosenthal fibers (50). In a small case series and literature review, Fiechter and coworkers noted an improved prognosis in comparison to other high-grade gliomas, with a median survival closer to 8 to 10 years.

When recurrences do occur, the tumor usually retains the original pilocytic histology, even when progression develops after many years (148). Most authors feel degeneration of these tumors to anaplastic astrocytoma or glioblastoma is uncommon, and may represent misdiagnosis of the initial tissue specimen at least in some cases (37; 02; 215; 173; 52; 91; 190). At the time of tumor recurrence, an indolent growth pattern often persists, allowing extended survival in a large proportion of patients. However, some authors have reported series of adult patients in which the recurrence rate was more pronounced, with a higher incidence of progression to more malignant pathology (47). In a report of 20 adult patients with pilocytic astrocytoma, the 12- and 24-month freedom from recurrence rates were 94% and 76%, respectively. Of the patients that recurred and required further surgery, the progression rate to anaplastic pathology was 75% (3 of 4 patients).

The complications of pilocytic astrocytomas vary depending on the location of the tumor. One of the most common complications is a persistent seizure disorder secondary to involvement of the cerebrum or diencephalon. The seizures are usually simple partial, simple partial with secondarily generalized tonic-clonic activity, or partial complex. Tumors that arise in the midline structures near the ventricular system may cause hydrocephalus at the time of presentation or subsequent recurrence. Hydrocephalus is most frequent with neoplasms of the cerebellum, hypothalamus, rostral brainstem, and optic chiasm. Hypothalamic involvement, either by a primary tumor or a secondarily infiltrating lesion, may cause endocrine disturbances (ie, diabetes insipidus, hypothyroidism, or infertility). On rare occasions, pilocytic tumors can spread into the cerebrospinal fluid, causing leptomeningeal metastasis and multifocal disease (173; 132; 123; 28; 48; 125; 61). In the study by Mamelak and colleagues, which included several adult patients and exclusively used gadolinium-enhanced MRI for follow-up, the estimated incidence of meningeal spread was 12% (123). The most frequent symptoms were diffuse or localized back pain and cognitive changes. Less common symptoms or signs included cranial neuropathies, radicular pain from root irritation, bowel and bladder dysfunction, and lower extremity weakness from lumbar root or cauda equina involvement. Tumors of the hypothalamus were much more likely to cause leptomeningeal metastases than tumors located elsewhere. Patients with meningeal metastasis had a poorer prognosis and reduced survival compared to patients with more typical pilocytic astrocytomas. However, some reports suggest that even with meningeal spread, patients can have prolonged survival (125). In a series of pilocytic astrocytoma patients with leptomeningeal disease, the median survival was 65 months, with a 5-year overall survival rate of 55.5%. The report by Buschmann and coworkers also supports an increased incidence of meningeal spread from pilocytic tumors of the chiasmo-hypothalamic region and recommends frequent screening with enhanced MRI scans of the spinal axis (28). The case reported by Grahnke and colleagues involved an initial pilocytic astrocytoma arising in the cerebellar vermis in a 66-year-old male (61). After undergoing surgical resection and chemo-radiotherapy with temozolomide, the patient developed acute right-sided hearing loss 7 years later, and was found to have tumor recurrence in the right cerebellopontine angle -- the mass looked like a typical acoustic schwannoma, but was proven at surgical resection to be a pilocytic astrocytoma similar to the original cerebellar tumor. Spontaneous hemorrhage is an uncommon complication of pilocytic astrocytomas. A report by White and coworkers reviewed 138 histologically proven cases (mostly adult, mean age 23 years), and noted hemorrhage in 11 (8%) (203). Although rare, pilocytic astrocytomas can also infiltrate along white matter tracts and spread across the midline (48).

Clinical vignette

K.S. was a 24-year-old woman with an unremarkable past medical history who first noted headaches 5 months prior to diagnosis. The headaches slowly worsened over the next few months and sometimes were associated with nausea and emesis. The patient would often awaken in the morning, or be aroused from sleep, with a headache. A tremor developed in her right hand that affected her handwriting. The patient denied any visual changes, focal weakness, speech abnormalities, or alterations of memory. After the headaches became continuous, she was evaluated at a local emergency room where a contrast-enhanced CT scan showed a large, cystic, partially enhancing, left temporoparietal mass. Edema and mild mass affect were noted around the lesion. The patient was placed on dexamethasone and brought to the operating room, where an extensive resection of the solid portion of the mass, as well as the cyst, was performed. On pathologic examination, the tumor was initially labeled an anaplastic astrocytoma (WHO grade 3); however, on further review it was more consistent with a juvenile pilocytic astrocytoma (WHO grade 1). After the operation, the headache and tremor were significantly improved. A postoperative enhanced MRI scan did not disclose any residual tumor. Because of the clean MRI scan and pathologic diagnosis of pilocytic astrocytoma, her physicians elected to follow her expectantly with frequent neurologic examinations and neuroimaging. Other treatment options, such as radiation therapy and chemotherapy, were deferred and would only be necessary if recurrent disease was not amenable to surgical resection. The patient remains free of recurrent tumor and is working full time.

Biological basis

Etiology and pathogenesis

The cells of origin of pilocytic astrocytomas, as well as the initial steps in neoplastic transformation of these cells, remain unknown. Some authors feel these tumors are derived from persistent glial cells of the embryonic subependymal plate, which have pluripotent developmental potential (149; 02).

On gross and microscopic pathologic examination, pilocytic astrocytomas of adults are often similar in appearance to the more common childhood variety of pilocytic neoplasm (37; 56; 149; 173; 49; 52; 122). In the World Health Organization classification scheme, pilocytic astrocytomas of adults and children are listed as astrocytoma, grade 1 (214). In adults, these tumors are most commonly located in the cerebrum (with a temporoparietal predilection), hypothalamus, optic pathways, brainstem, cerebellum, and diencephalon. Most often, pilocytic astrocytomas appear as well-circumscribed, yellowish-to-grayish pink and translucent masses that often contain multiple small or singular large cysts (37; 215). Tumors with large cysts are usually located in the cerebrum or cerebellum and typically demonstrate a single, solid, mural tumor nodule along the cyst wall. The cyst may contain a variable amount of clear-to-yellowish fluid. Small hemorrhages or regions of calcification may be noted within solid portions of the tumor. The margins of pilocytic astrocytomas typically appear distinct or encapsulated on histological inspection, but can be more diffuse or infiltrative in up to 50% to 60% of cases (173; 38). Tumor cells can often be seen extending into regional meninges.

On microscopic examination of adult pilocytic astrocytomas, 2 distinct forms are recognized: (1) the more frequently diagnosed classic "juvenile" form and (2) an "adult" variant (37; 173; 122). Various mixtures of these 2 forms can also be seen. The juvenile form of pilocytic astrocytoma demonstrates a biphasic pattern in which mild to moderately cellular areas of elongated, fusiform unipolar and bipolar "piloid" cells are interspersed with hypocellular, microcystic regions (215; 173). Within the pilocytic zones, the piloid cells form streams, whorls, parallel arrays, and other patterns (often with interdigitating processes), and often longitudinally sheath nearby blood vessels. Piloid cells and their processes stain with glial fibrillary acidic protein, demonstrating their astrocytic lineage. The amount of cellular and nuclear pleomorphism is mild to moderate, whereas mitoses are generally uncommon (ie, less than 1 to 2 mitoses per high-power field). Regions of necrosis and florid endothelial proliferation are rare. Occasional tumors will display robust vascularization (5% to 12%), suggestive of arteriovenous malformation (115; 52).

These behave in a benign biological fashion similar to tumors that do not contain excessive vasculature. Rosenthal fibers are densely eosinophilic rod-shaped structures that are commonly found within the pilocytic zones.

They are contained within the processes of piloid cells and appear to be aggregates of glial filaments that contain significant concentrations of alpha-beta-crystallin protein (43; 191). Rosenthal fibers are immunoreactive with glial fibrillary acidic protein; intense staining is usually visible toward the outer surface of the fiber, whereas milder staining occurs near the center.

The microcystic zones of juvenile pilocytic astrocytomas are remote from blood vessels and are characterized by a more hypocellular appearance.

The cells are not piloid; rather, they are multipolar and stellate in appearance, resembling protoplasmic astrocytes (37; 215; 173; 52). Numerous microcysts are interspersed throughout the loose, reticulated regions of cellularity. Many of the cysts contain an eosinophilic, protein-rich fluid.

In some tumors, the microcysts coalesce into larger cysts, some of which become macroscopic. Granular bodies are eosinophilic hyaline granules present within cell processes of stellate astrocytes that are only detected in microcystic zones (215; 173).

They apparently form as small autophagic vacuoles become incorporated into larger membrane-bound bodies during progressive pinocytosis of protein material (173). The exact composition of granular bodies is unknown, but they do contain large amounts of -1-antitrypsin and -1-antichymotrypsin (91; 92).

The adult variant of pilocytic astrocytoma does not demonstrate a biphasic pattern. Instead, these tumors display more densely packed bundles of broad bipolar fibrillated cells that form monotonous, interweaving, intersecting patterns (37; 173).

The tumor cells do not sheath blood vessels and have no predilection for microcystic degeneration. Rosenthal fibers may be numerous, whereas granular bodies are generally uncommon. The mean age of patients with the adult variant tends to be older than that of patients with the more classic juvenile form: 29.9 years versus 20.3 years (37; 173). Anaplastic degeneration of pilocytic astrocytomas is generally uncommon but tends to occur more frequently in the adult variant (173). This is demonstrated in a case report of an adult patient with a cerebellar pilocytic astrocytoma that was transitioning into a glioblastoma multiforme (176). The different regions of the tumor had a transition zone in between them, which had an identical molecular signature. Another paper reported a series of 3 adult patients with pilocytic astrocytomas that presented with anaplastic histological features at diagnosis (209). All of the tumors displayed regions of necrosis, marked nuclear pleomorphism, and a very high mitotic rate.

Some investigators contend that pilocytic astrocytomas are a separate pathologic entity from other low-grade astrocytomas and should not be graded using the typical neuropathologic classification criteria (eg, Kernohan, St. Anne-Mayo, and Daumas-Duport classification schemes) (52; 54; 69; 40). Because these tumors may contain features suggestive of anaplasia (eg, mitoses, pleomorphism, vascularity), they are often misclassified as higher-grade tumors. For example, in the study by Forsyth and colleagues, the majority of pilocytic astrocytomas would have been classified as grade 2 or higher using the St. Anne-Mayo or Kernohan grading systems (52). Furthermore, in these studies, the survival of patients with pilocytic astrocytomas was not found to correlate with tumor grade or presence of "anaplastic" histological features as predicted by the common grading systems (52; 69). In a study of interobserver concordance among neuropathologists, pilocytic astrocytomas accounted for a significant percentage of misclassification, especially supratentorial tumors with a uniphasic histological pattern or anaplastic features (40). In contrast to the above data, a study from the Mayo Clinic suggests that the presence of anaplastic features in a pilocytic astrocytoma may predict more aggressive behavior (165). The authors reviewed a subgroup of 34 pilocytic astrocytomas among a larger cohort of 2200 cases. Overall and progression-free survival was significantly shorter in patients with tumors that had more pronounced mitotic activity and the presence of necrosis, paralleling that of grade 2 and 3 diffuse astrocytomas.

Ganglioside content and glycolipid markers are other methods of differentiating pilocytic astrocytomas from higher-grade gliomas (197; 206). In a study of 31 gliomas (7 pilocytic) using microbore high-performance liquid chromatography, Wagener and colleagues noted that pilocytic tumors had high GD3 content, moderately reduced levels of GD1b/GT1b/GQ1b, and low levels of GD1a (197); however, it remains unclear how clinically useful this information will be because glioblastoma multiforme has a similar profile. Yates and coworkers noted high levels of GD3 and globoside as well as similarly reduced levels of GD1b and GD1a (206). Similarly, nestin expression can be used to differentiate pilocytic astrocytoma from more aggressive astrocytic tumors (03). Nestin, an intermediate filament protein, is consistently expressed at low levels compared to normal brain, whereas higher-grade fibrillary astrocytomas demonstrate progressively higher expression.

Some studies have evaluated the role of cell adhesion molecules and extracellular matrix proteins in brain tumor invasiveness (53; 208; 06). In a study of the role of integrins in tumor cell migration, pilocytic astrocytomas were found to be much less mobile compared to high-grade astrocytomas (53). Ylagan and Quinn assessed the correlation between CD44 expression and degree of malignancy in astrocytic tumors (208). No correlation was noted between CD44 expression and tumor grade: pilocytic astrocytoma had 89% CD44 staining, whereas glioblastoma multiforme had 84% CD44 staining. In a study of migration ability using a myelin substrate, Amberger and colleagues demonstrated a correlation between capacity for migration and tumor grade (06). Pilocytic astrocytomas were significantly (p less than 0.001) less able to migrate on the myelin substrate compared to high-grade tumors. Some data suggest that pilocytic astrocytomas have a significantly more robust binding capacity to saccharide epitopes in the extracellular matrix when compared to more aggressive fibrillary astrocytomas (30). The same group of authors has also demonstrated that the levels of expression of galectins correlate with grade in astrocytomas (29). Galectins have specificity to beta-galactosides and are involved in growth regulation, cell adhesion, and cell migration. They are minimally expressed in pilocytic astrocytomas compared to grade 3 and grade 4 tumors. A paper has evaluated the expression of galectin-3 in a series of 90 glioma specimens of different grade (140). In contrast to the reports cited above, the immunohistochemical expression of galectin-3 was high in pilocytic astrocytomas and glioblastoma multiforme and low or undetectable in other glial tumors. The staining was diffuse in the pilocytic tumors and more focused around regions of necrosis and pseudopalisading cells in glioblastoma. The authors suggested that galectin-3 could be used to differentiate pilocytic astrocytomas from other diffuse astrocytomas. A microarray analysis of several cohorts of sporadic pilocytic astrocytomas noted high expression of matrilin-2, an extracellular matrix protein (178). Matrilin-2 was not overexpressed in glioblastoma multiforme. In addition, expression of matrilin-2 and its protein was considered a biomarker for more clinically aggressive pilocytic tumors.

DNA flow cytometric studies of pilocytic astrocytomas have demonstrated that the majority of tumors (50% to 78%) are diploid, with S-phase fractions of between 2% and 4.5% (52; 69; 190; 64). The only exceptions to these data are a small subset of histologically malignant cerebellar pilocytic astrocytomas that had a mean S-phase fraction of 7.6%, significantly higher than more typical pilocytic tumors (190). No correlations have been found between flow cytometric measures of ploidy or S-phase fraction and patient survival. Histological characteristics of tumors did not correlate with flow cytometric analyses.

Analysis of apoptosis rate in a series of 78 pilocytic astrocytomas demonstrated a low apoptotic index (64). In addition, there was no significant correlation between apoptotic index and indices of cell proliferation (ie, Ki-67, S-phase fraction). However, the apoptotic index did correlate with p53 positivity, although the absolute number of p53 positive tumors was small. These results are in contrast to a study by Nakamizo and colleagues, in which pilocytic tumors had elevated apoptotic ratios in comparison to higher grade astrocytic tumors (138). The authors speculate that the elevated apoptotic ratio may account for the less aggressive biological behavior of these tumors.

Other measures of the biological potential of pilocytic tumors include bromodeoxyuridine and Ki-67 labeling studies and positron emission tomography scans of glucose metabolism using fluorine-18 fluorodeoxyglucose and silver staining of nucleolar organizing regions. Several investigators have evaluated the proliferative activity of these tumors after bromodeoxyuridine labeling (80; 132). The labeling indices, another measure of the percentage of cells in S-phase, were consistently low (mean labeling index 1.0%) in 51 tumors studied by Ito and colleagues and Mishima and colleagues. Six of their patients were more than 18 years of age. The labeling index tended to decrease with advancing age (80). Similar results have been found using Ki-67 and MIB-1 immunohistochemical techniques, with a mean labeling index ranging from 0.6% to 2.0% for most authors (94; 121; 60; 64; 189; 22). In a study by Dirven and colleagues, the MIB-1 labeling index was 4.2% for pilocytic tumors, similar to the labeling index for low-grade fibrillary astrocytomas (44). However, Bowers and coworkers report that tumors with MIB-1 labeling indices of greater than 2.0 suggest a more aggressive phenotype and are associated with shorter progression-free survival than tumors with lower indices. In contrast, positron emission tomography scan determinations of glucose metabolism in 5 adult patients and a 16-year-old girl with pilocytic astrocytomas were similar to measurements from patients with anaplastic astrocytomas, and were consistently higher than measurements from patients with low-grade diffuse astrocytomas (54; 166). These findings suggest an aggressive biological potential in pilocytic tumors that is not expressed clinically. Alternatively, the elevated glucose metabolism may be a reflection of the dense vascularity noted in these tumors and may be caused by endothelial cell activity (166). Silver staining of nucleolar organizing regions correlates with protein synthesis and is an indirect measure of proliferative activity. In a study of 20 patients with pilocytic astrocytomas, Dirven and coworkers noted that measures of silver staining below the group median were associated with indolent clinical behavior (45). They concluded that silver staining of nucleolar organizing regions may be helpful for predicting behavior of incompletely resected pilocytic tumors.

Cytogenetic and chromosomal studies of pilocytic astrocytomas reveal frequent abnormalities (161; 84; 210). In an analysis of 3 adult pilocytic astrocytoma patients, Jenkins and colleagues found all 3 to have heterogeneous chromosomal aberrations (84). Two patients had tumors with complex clonal karyotypes containing 27% and 100% abnormal metaphases, respectively. In the third patient, the tumor contained near-tetraploid cells that demonstrated extra copies of chromosomes 6 and 8 and lacked chromosomes 15 and 22. Two other pilocytic tumors have been cytogenetically analyzed, the first of which was normal, whereas the other lacked an X chromosome (161). In a study utilizing fluorescence in situ hybridization, 6 of 18 pilocytic astrocytomas were found to have chromosomal gains, most commonly chromosomes 7 and 8 (202). These gains were not present in a similarly studied group of low-grade fibrillary astrocytomas (p less than 0.05). Zattara-Cannoni and colleagues cytogenetically evaluated 24 pilocytic astrocytoma samples and found abnormalities in 50% of the cases (210). Chromosomes 7, 8, and 11 were most frequently involved.

In a series of 44 patients with pilocytic astrocytoma that included 12 patients 16 years of age or older, whole genomic analysis was performed to evaluate chromosomal changes (88). The most frequently affected chromosomes were 5 and 7, followed by 6, 11, 15, and 20. Chromosomal gains were more frequent in tumors from patients over 15 years of age (p = 0.03). In addition, older patients were noted to have a greater number of chromosomes involved.

A few studies have examined the role of oncogenes and tumor suppressor genes in the genesis of pilocytic astrocytomas (17). In loss of heterozygosity studies evaluating evidence for a tumor suppressor gene on chromosome 19q, pilocytic astrocytomas consistently retained heterozygosity, unlike astrocytomas of higher grade (196). In a general screen for loss of heterozygosity, James and colleagues were unable to demonstrate any chromosomal abnormalities in 3 pilocytic astrocytomas (82). Lang and colleagues, in an evaluation of 7 pilocytic astrocytomas, were unable to demonstrate loss of heterozygosity affecting either chromosomes 17p or 10, and found no evidence for amplification of the epidermal growth factor receptor gene (105). In related studies, several authors have found frequent mutations of the PTEN gene (tumor suppressor gene located on chromosome 10q23) in high-grade astrocytomas but not in pilocytic or low-grade fibrillary astrocytomas (159; 46; 41; 212). Mutations of the p53 tumor suppressor gene are common in higher-grade astrocytomas, but they are infrequent in pilocytic astrocytomas (35; 113; 78; 64; 189; 17). In a study by Lang and colleagues, p53 protein accumulation was noted in 5 of 7 pilocytic astrocytomas examined immunohistochemically, whereas only 1 of these tumors demonstrated a genomic mutation of p53 within exons 2-11 (106). Using single-strand conformation polymorphism analysis and DNA sequencing of the p53 gene, Patt and colleagues noted a genomic mutation in only 1 of 7 pilocytic tumors (151). The mutation was silent and did not lead to any amino acid changes or protein alterations. Other investigators have not found immunohistochemical evidence for p53 protein accumulation in these tumors (35; 99). However, a study by Hayes and colleagues used denaturing gradient gel electrophoresis of the entire p53 coding region of 20 pilocytic astrocytomas and noted that 7 tumors (35%) had causative mutations (68). Similarly, a report by Wemmert and colleagues noted frequent deletions of the p53 gene using fluorescent in situ hybridization techniques (200). Because of its ability to bind to p53 and reduce p53 inhibition of the cell cycle, MDM2 has also been evaluated in astrocytic tumors (124). Amplification of MDM2 appears to play a role in transformation of malignant gliomas, but is rare in pilocytic tumors. Immunohistochemical analysis also failed to demonstrate amplification of the epidermal growth factor receptor in a large series of pilocytic astrocytomas (99). Reduced expression or loss of CDKN2/p16 can lead to upregulation of the cell cycle due to increased activity of CDK4. In an analysis of 78 gliomas (7 pilocytic tumors), Miettinen and coworkers found that expression of CDKN2/p16 correlated inversely with glioma grade (131). Tumors with a loss, or reduced expression, of CDKN2/p16 tended to have a higher grade, higher proliferative indices, and cause shorter survival intervals. Expression of CDKN2/p16 was retained in all 7 pilocytic astrocytomas. Using immunohistochemical techniques, van der Valk and colleagues screened 4 pilocytic tumors for EGFR, TGF-alpha, PDGF-A, PDGF-B, and bFGF (194). All 4 tumors stained for TGF-alpha, but had negligible or weak staining for the other growth factors. Several authors have used immunohistochemical techniques to assess the labeling index of cyclin D1 in pilocytic astrocytomas (32; 121). In an evaluation of 10 pilocytic tumors by Chakrabarty and colleagues, the mean LI was 3.0% in the pilocytic samples and 14.0% to 15.2% in malignant astrocytomas (p equal to 0.04). The authors concluded that cyclin D1 is overexpressed in astrocytic tumors and is involved as an oncogene in astrocyte transformation. In contrast, Machen and Prayson studied 48 pilocytic tumors and found much lower amounts of cyclin D1 labeling, with a mean index of 0.1% (121). Chakrabarty and Bridges evaluated the role of cyclin A in a series of astrocytic tumors (31). The cyclin A labeling index was low (mean 0.12%) for pilocytic tumors and significantly higher (mean 5.95% to 6.7%) in malignant astrocytomas. In loss of heterozygosity studies evaluating chromosome 17q, 4 of 20 (20%) patients with pilocytic astrocytomas demonstrated allelic loss (195). Eight of the 20 patients were greater than 18 years of age, and 2 of these adult patients demonstrated allelic loss. These data suggest that the presence of a tumor suppressor gene on the long arm of chromosome 17 may be involved in the transformation of pilocytic astrocytomas. In an analysis of 9 radiographically progressive neurofibromatosis type 1-associated pilocytic astrocytomas, Li and colleagues were unable to demonstrate abnormalities of p53, p16, RB, EGFR, CDK4, PDGF-A, or PDGFR (112). The authors concluded that different genetic pathways were active in these tumors versus more aggressive fibrillary astrocytomas. Similar conclusions were reached by Cheng and coworkers in a molecular and immunohistochemical analysis (p53, p16, CDK4, PTEN) of 29 pediatric pilocytic astrocytomas (34); however, in this series, the authors did find evidence of reduced protein expression of the PTEN tumor suppressor gene (62% of tumors), which is in contrast to PTEN expression in low-grade fibrillary astrocytomas. The neurofibromatosis-related NF-1 gene was considered a potential candidate because it is located on 17q and may function as a tumor suppressor gene due to its ability to convert activated Ras-GTP into the inactive form, Ras-GDP (195). However, data suggest that NF-1 may not function as a simple tumor suppressor gene in pilocytic astrocytomas (156). An evaluation of 6 pilocytic tumors demonstrated upregulation of NF-1 gene expression instead of reduced or absent expression that would be expected with a tumor suppressor gene and protein product. Further analysis suggested that the expressed transcript and protein were full-length and probably wild type. The authors theorized that the increase in NF-1 expression may represent a normal regulatory response to excessive Ras-GTP-mediated proliferative signals. Kluwe and associates evaluated sporadic and NF1-associated pilocytic tumors for allelic loss of chromosome 17, with emphasis on the NF-1 region (98). There was a significant difference between the groups (p< 0.0001), with loss of heterozygosity noted in 11 of 12 NF1-associated tumors and only 1 of 24 sporadic tumors. The authors conclude that different genetic pathways exist leading to sporadic and NF1-associated pilocytic astrocytomas. In an analysis of 10 sporadic pilocytic astrocytomas, Wimmer and coworkers were unable to discern any inactivating mutations or altered gene expression of the NF1 gene (204). In light of reports suggesting responsiveness of pilocytic astrocytoma cell cultures to gefitinib (ZD1839, Iressa), an inhibitor of EGFR, Addo-Yobo and colleagues investigated the expression of the entire family of ErbB receptors in pilocytic and other astrocytomas (01). They noted that ErbB3 was highly overexpressed in pilocytic tumors in comparison to other pediatric gliomas. In addition, overexpression of ErbB3 was often present in association with overexpression of Sox10, which is known to positively regulate the expression of ErbB3. It was theorized that Sox10-regulated overexpression of ErbB3 might be driving growth in pilocytic astrocytomas. Other authors have attempted to correlate p53 and MGMT expression, MIB-1 labeling, morphological features, and tumor location with patient survival and outcome (73). MIB-1 proliferation index, as well as expression of p53 and MGMT, did not correlate with outcome. Oligodendroglial morphology and the absence of leptomeningeal invasion were adverse histologic features but only in cerebellar tumors. The authors concluded that morphologic biomarkers were available for pilocytic tumors but had variable utility according to tumor location. Data have implicated the Sonic Hedgehog pathway in the tumorogenesis of pilocytic astrocytomas (171). In a cohort of 20 patients (5 over 18 years of age), they noted elevated expression levels of PTCH in 45% of tumor samples. The expression levels were inversely correlated with patient age. Immunohistochemical analysis of PTCH, Gli-1, and Ki-67 labeling indices demonstrated higher expression in patients younger than 10 years of age, in comparison to those diagnosed after 10 years of age. A molecular evaluation of the PI3K/Akt signaling pathway by Rodriguez and coworkers studied 3 subsets of pilocytic astrocytomas: conventional (N = 43), clinically aggressive/recurrent (N = 24), and anaplastic (N = 25) (164). Cytogenetic studies noted more frequent heterozygous PTEN and homozygous p16 deletions in the recurrent and anaplastic subgroups. A reduced level of PTEN expression, along with an increase in cytoplasmic pAkt, was noted in the more aggressive tumors. The authors concluded that activation of the PI3K/Akt pathway may underlie biological aggressiveness in pilocytic astrocytomas. A study has evaluated and compared the Wnt/beta-catenin/Tcf pathway in pilocytic astrocytomas and diffuse grade II astrocytomas (175). The expression of beta-catenin, Tcf4, Lef1, and c-Myc was increased in pilocytic tumors in comparison to grade II tumors, suggesting differential activation of this pathway.

Rickman and associates reported an analysis of molecular profiling of glioblastoma versus pilocytic astrocytoma using oligonucleotide microarray techniques (163). They compared the expression pattern of approximately 6800 genes and identified a group of 360 genes that were significantly different (p< 0.01) between the 2 tumor groups. Enhanced expression of genes involved in cell proliferation and cell migration was noted in the subset of glioblastomas. In contrast, the pilocytic tumors had reduced expression of genes involved in cell motility and enhanced expression of genes that suppress migration (eg, TIMP3, TIMP4). Similar results were noted by Hunter and colleagues in a differential expression analysis comparing pilocytic astrocytomas to anaplastic astrocytomas (77). In addition, they found enhanced expression (6.5- to 8.5-fold) of apolipoprotein D in the pilocytic tumors. Apolipoprotein D has been proposed as a marker of cell cycle arrest. Apolipoprotein D was also found to be elevated in a similar study comparing pilocytic astrocytomas to oligodendrogliomas and normal white matter (63). Another gene, EF-1alpha2, was noted to be increased in sporadic pilocytic astrocytomas relative to NF1-associated pilocytic tumors. EF-1alpha2 may determine the susceptibility of cells to transformation or apoptosis. An oligonucleotide microarray study evaluated the gene expression of 21 pilocytic astrocytomas and compared the results to expression data from normal cerebellar tissue (205). An unsupervised hierarchical analysis of the results suggested 2 potential subgroups of pilocytic astrocytoma, which differed in their expression of numerous genes, including those involved in cell adhesion, growth, motility, and angiogenesis (eg, vascular endothelial growth factor). A microarray analysis comparing pilocytic astrocytomas to normal cerebellum, grade 2 astrocytomas, and oligodendrogliomas noted increased expression of immune defense genes, including HLA-DR-alpha, HLA-DPA1, HLA-DPB1, and A2M (76). The authors suggest that higher expression of these genes may be related to the more benign biological behavior of pilocytic astrocytomas. Studies have also investigated gene expression differences between pilocytic and higher-grade astrocytomas. Rorive and colleagues used microarray analysis and noted increased expression of TIMP4, CINH, CHAD, THBS4, IGFBP2, and TLE2 in the pilocytic tumor group, in comparison to the more malignant tumors (168). All of these genes produce proteins involved in the inhibition of migration, which is consistent with the less invasive phenotype of pilocytic tumors. A similar study by Colin and colleagues, using suppression subtractive hybridization, documented elevated expression of several proteins involved in invasion and angiogenesis, including fibronectin, osteopontin, chitinase-3-like-1, keratoepithelin, and fibromodulin (39). Takei and colleagues examined 64 cases of pilocytic astrocytoma for expression of oligodendroglial differentiation markers, including myelin basic protein, PDGFR-alpha, Olig-1, and Olig-2, using immunohistochemical techniques (186). There was a significant inverse correlation between myelin basic protein expression and Ki-67 labeling index (p=0.014), as well as a positive correlation between PDGFR-alpha and Ki-67 labeling (p=0.011). No correlations were noted between Olig-1 and Olig-2 expression and the Ki-67 labeling index. Progression-free survival was maximized in tumors with low expression of PDGFR-alpha and high expression of myelin basic protein.

Several investigators have determined that adult and pediatric pilocytic astrocytomas often demonstrate gains in chromosome 7q34, which contains the known or predicted sequence of up to 17 genes, including BRAF (12; 89; 100; 17). In one study, activation of downstream pathways (MEK and ERK) was demonstrated in the majority of tumor specimens (12). Another study revealed that in many pilocytic tumors, a tandem duplication can occur at 7q34, which results in the production of an in-frame fusion gene that incorporates the kinase domain of the BRAF oncogene (89). The fusion protein results in constitutive BRAF activity, with the ability to transform NIH-3T3 cells. A report from Korshunov and colleagues showed that the BRAF-KIAA1549 fusion gene was present in 70% of pilocytic astrocytomas, but not in grade 2 diffuse astrocytomas (100). However, the diffuse astrocytomas contained mutations in the IDH1 and IDH2 genes, which were not present in the pilocytic tumors. Studies have corroborated the high frequency of BRAF gene fusions in pilocytic tumors, and the lack of these molecular abnormalities in higher grade astrocytomas (74; 108; 150). In addition, the BRAF gene rearrangements were more common in cerebellar tumors than noncerebellar tumors and were more often associated with classic biphasic histology (74). The BRAF fusion protein mechanism for activation of the MAPK pathway has been corroborated and extended by several new studies. The KIAA1549-BRAF fusion protein has been noted in up to 80% of pediatric pilocytic astrocytomas, as well as in some nonpilocytic astrocytomas (85; 67). In both tumor types, the presence of the fusion protein denotes a more favorable prognosis. A similar deletion mechanism affecting chromosome 7q34 can result in a FAM131B-BRAF fusion gene and protein (36). The deletion removes the N-terminal inhibitory domains of BRAF, leading to constitutive activity of the BRAF kinase. A study using proteomics analysis has corroborated the activation and dominance of the MAPK signaling pathway in pediatric pilocytic astrocytoma (07). Using comparative microarray technology, Tchoghandjian and colleagues have discovered differences between hypothalamic-chiasmatic and cerebellar pilocytic astrocytomas (187). Several genes were found to be upregulated only in the hypothalamic-chiasmatic cohort, including NOTCH2, LHX2, SIX6, CDK5, and PTEN. LHX2 and SIX6 are genes involved in normal optic nerve development, whereas the NOTCH2 gene is important for promoting normal gliogenesis during development and maintaining glial function in adults. The authors speculate that hypothalamic-chiasmatic pilocytic astrocytomas may derive from the remnants of radial glial cells that normally are important for guiding the path of retinal ganglion cell axons. A similar study by Reis and colleagues attempted to differentiate pilocytic astrocytomas of the optic nerve from those arising in the posterior fossa (160). They found that BRAF mutations and aberrations were more frequent in posterior fossa pilocytic tumors. In particular, BRAF duplication was more frequently noted in posterior fossa tumors in comparison to those from the optic nerves (p=0.011). In contrast, expression of phospho-MAPK1 and CDKN2A were high in tumors from both regions. In a review of 96 pilocytic astrocytomas that underwent whole-genome sequencing, Jones and coworkers also verified the presence of new BRAF-activating mutations (87). In addition, they noted recurrent activating mutations in FGFR1 and PTPN11, as well as new NTRK2 fusion genes, in noncerebellar pilocytic tumors. MAPK pathway activation was noted in all of the tumors analyzed.

As noted above, activation of the MAPK pathway is common in sporadic pilocytic astrocytomas, and it often involves abnormalities in Raf signaling (17). A review of a large series of adult patients with pilocytic astrocytomas described molecular data screening for abnormalities in the Ras/Raf/MAPK pathway (188). In a series of 127 adult patients, available tumors were screened for BRAF-KIAA1549 fusion/duplication and BRAF V600E mutations by fluorescence in situ hybridization. B-K fusion was noted in 20% (9 of 45) of patients but did not correlate with outcome. No BRAF V600E mutations were found in the cohort (0 of 40). In contrast, a report by Bannykh and colleagues noted a relatively high number of BRAF V600E mutations in a series of 51 adult patients with pilocytic astrocytomas (11). The BRAF V600E containing tumors appeared to be more infiltrative, but limited follow up did not detect any negative prognostic significance. The detection of BRAF mutations is of clinical importance as they are potentially targetable, with agents approved for the treatment of other diseases such as melanoma. In a study of 59 adult patients with pilocytic astrocytomas, Pathak and colleagues noted that 36% of the cohort harbored BRAF and/or FGFR genetic alterations (150). The FGFR related genetic alterations were most often noted in tumor from the supratentorial region (32%; P = 0.01). There was widespread activation of the MAPK/ERK/mTOR signaling pathway, with immunopositivity of p-MAPK and p-MEKq present in all tumors examined.

Advances into the understanding of angiogenesis and the process of malignant neovascularization have been applied to brain tumors (201). Jallo and colleagues examined the concentration of tenascin-C (extracellular matrix protein) in the cyst fluid of various brain tumors and correlated its level with tumor grade and degree of angiogenesis (81). In general, the level of tenascin-C was 5- to 10-fold less in pilocytic than malignant astrocytomas. Tenascin-C immunoreactivity was found to correlate with tumor grade and the degree of vascular hyperplasia. Similar results have been noted by Kim and colleagues, with tenascin-C levels being lowest in the cells and neoplastic vessels of pilocytic astrocytomas, compared to higher grade fibrillary astrocytomas (97). Several authors have evaluated the expression of vascular endothelial growth factor (VEGF) in pilocytic astrocytomas (111; 155; 120). Leung and colleagues evaluated the expression of VEGF and its receptors (KDR and Flt-1) in 14 pilocytic tumors using in situ hybridization techniques (111). High levels of VEGF transcripts were found in the tumor cyst walls, regions of cystic degeneration, in stellate reticulated astrocytes, and in tumor cells with degenerative pleomorphic nuclei. Both receptors were highly expressed in the tumor vasculature. In the study by Pietsch and colleagues, immunohistochemical techniques were used to analyze the degree and distribution of VEGF expression in 16 pilocytic astrocytomas (155). Significant immunoreactivity for VEGF was noted in 7 of 16 tumors (44%); however, there was a poor correlation between the degree and pattern of VEGF staining and the extent of vascularization. In general, the amount of VEGF immunoreactivity was significantly higher in malignant tumors. Machein and associates found a correlation between VEGF mRNA expression and vascular permeability (120). In addition, pilocytic astrocytoma VEGF mRNA levels were substantially higher than measurements in low-grade fibrillary astrocytomas, consistent with the robust vascularity noted in pilocytic tumors. In a study comparing the vasculature of pilocytic and anaplastic astrocytomas, Gesundheit and coworkers noted that the anaplastic tumors had smaller and more immature vessels, with a more prominent concentration of VEGFR-1 receptor expression (59).

Studies have discovered the presence of somatic mutations within the DNA of mitochondria (mtDNA) in various forms of cancer, including brain tumors. Lueth and colleagues have studied the mtDNA from a series of 19 patients with pilocytic astrocytomas, including several adults (118). Unexpectedly, 84% of the samples demonstrated somatic mutations in tumor mtDNA, which is much higher than prior analyses of more malignant tumors, such as glioblastoma multiforme (41%) and medulloblastoma (40%). Several of the mutations were present in genomic regions involved in pathways of oxidative phosphorylation. Some of these mutations may enhance the generation of reactive oxygen species and contribute to oncogenic transformation; however, the exact mechanisms remain to be elucidated.

Micro-RNA (miRNA) are small noncoding RNA sequences that are able to modify the activity of mRNA and subsequent protein expression, which have been implicated in the transformation and pathogenesis of many solid tumors, including brain tumors. A study by Ho and colleagues examined a series of 43 patients with pilocytic astrocytomas (35 sporadic) for evidence of differentially expressed miRNA in comparison to normal brain tissue controls, using miRNA microarray and quantitative real-time PCR techniques (71). They noted a subset of miRNAs that were differentially expressed in the tumors: 13 that were underexpressed and 20 that were overexpressed. miR-124 and miR-129 were underexpressed 17-fold and 15-fold, respectively, whereas miR-21 was overexpressed by 19.8-fold. Predicted targets of mi-RNA with increased expression at the mRNA and/or protein level included PBX3, METAP2, and NFIB.

Gene methylation patterns are now known to influence gene activity and protein expression in tumors, and are involved in transformation and therapeutic resistance. Lambert and colleagues analyzed the global DNA methylation profiles of a series of 62 pilocytic astrocytomas, along with normal cerebellar controls (104). They noted 2 subgroups of pilocytic tumors that separated according to methylation pattern and tumor location: infratentorial versus supratentorial. They also identified several important neural developmental genes that were differentially methylated between the 2 groups, including NR2E1 and EN2. Another study looked at differential gene expression and differential methylation patterns in pilocytic astrocytomas and noted 1 set of genes that were positively related, whereas another set were negatively related (eg, DOCK2 [dedicator of cytokinesis 2], DOCK8, FCGR2A [Fc fragment of IgG, low affinity IIa]) (211). Many of the differentially expressed genes were strongly methylated and were considered to be potential therapeutic targets. Another case report of an adult with a pilocytic astrocytoma, in which array comparative genomic hybridization analysis was performed, also showed enrichment of genes involved in Fc gamma R-mediated phagocytosis, as well as olfactory transduction and p53 signaling (152).

Mutations of the isocitrate dehydrogenase 1 and 2 genes (IDH-1/IDH-2) are very common in grade II and III astrocytomas but have been generally absent in most studies in pilocytic astrocytomas. However, several reports do suggest that they can occur in rare, sporadic pilocytic astrocytomas (13; 129). The 2 case reports have been in older patients with tumors in the cerebellum.

Mutations in the H3 K27M gene have been reported to be associated with diffuse astrocytomas developing in the midline regions of the brain, as reported in the 2016 World Health Organization classification of tumors of the CNS, and are now considered a new distinct clinical entity -- diffuse midline glioma, WHO grade 4 (117). Patients with these tumors typically have a poor prognosis and short survival time. However, this mutation has now been reported in several cases of pilocytic astrocytoma, including an adult case (134). In the adult case described by Morita and colleagues, the tumor was completely resected and the patient did not receive any further treatment. The most recent follow-up MRI scans at 8 months did not demonstrate any recurrence of tumor. The authors concluded that in selected rare cases, the H3 K27M mutation might not be associated with grade 4 behavior and a malignant course.

Epidemiology

The exact incidence and prevalence of pilocytic astrocytomas in adults are unknown. Overall, pilocytic astrocytomas have accounted for 1% to 6% of all brain tumors in several large series (149; 215). The majority of these neoplasms develop in children or adolescents, with approximately 20% to 25% occurring in patients older than 18 years of age (215; 198). In describing a series of 7 patients with cerebral pilocytic astrocytomas, Garcia and Fulling estimated an incidence of 7% from a population of adult patients with cerebral low-grade and anaplastic astrocytomas (56). Cerebellar pilocytic astrocytomas constitute 4% of all intracranial gliomas, and approximately 25% of these tumors occur in adults (215). In a population-based study from Zurich, Switzerland, Burkhard and coworkers reviewed the regional incidence of pilocytic astrocytomas from 1980 to 1994 (26). In a series of 987 astrocytic and oligodendroglial neoplasms, pilocytic astrocytomas had an adjusted incidence rate of 4.8 per 1 million per year.

Prevention

No known measures exist to prevent the development of a pilocytic astrocytoma in de novo patients or in those at risk from associated disorders (eg, neurofibromatosis).

Differential diagnosis

The differential diagnosis of pilocytic astrocytoma consists of other mass lesions or disease processes that can cause elevations of intracranial pressure, seizure activity, or focal neurologic signs and symptoms. Due to the typical longstanding nature of the symptoms and the slowly progressive pace of the illness, a low-grade brain neoplasm should be suspected. The most common preoperative diagnoses in 30 adult patients with cerebral pilocytic astrocytomas were avascular mass (67%), calcified mass (23%), meningioma (13%), cystic mass (10%), astrocytoma (10%), and ependymoma (10%) (37). Other masses to consider for a cerebral lesion include abscess, ganglioglioma, anaplastic astrocytoma, and oligodendroglioma. For pilocytic astrocytomas of the suprasellar or chiasmal region, other considerations consist of craniopharyngioma, germinoma, loculated leptomeningeal metastases, and invasive pituitary macroadenoma (109; 181). For example, a report describes 3 patients with pilocytic astrocytomas that presented in the intra- or suprasellar region, and were thought to be craniopharyngiomas (181). Pilocytic tumors of the cerebellum must be differentiated from medulloblastomas, ependymomas, and higher-grade astrocytomas. In the periventricular region, subependymal oligodendrogliomas and ependymomas can appear similar to pilocytic astrocytomas (109).

Diagnostic workup

In those patients with neurologic signs and symptoms suggestive of a pilocytic astrocytoma, neuroimaging with CT or MRI is the most critical diagnostic test. These studies should be performed with and without contrast media for the most accurate visualization of the mass and to assist in differential diagnosis. Pilocytic astrocytomas appear as discrete, well-circumscribed, sharply demarcated masses on CT and MRI (109; 49; 52; 54; 91; 38).

Infiltration into surrounding structures, although rare, is most commonly noted with tumors involving the optic pathways. The mass usually appears hypodense or isodense compared to brain on CT (109). On MRI T1-weighted images, the tumor is isointense or slightly hyperintense relative to brain, whereas on T2-weighted images, it is always hyperintense (109; 38). Calcification is noted in 11% to 40% of tumors with CT but is more poorly visualized by MRI (109; 183). Cysts of various sizes are present in many tumors, most commonly those of the cerebellum and cerebrum.

With large cystic tumors, a solid mural nodule is usually evident somewhere along the cyst wall. The cyst fluid is hyperintense on T2-weighted and hypointense on T1-weighted MRI scans.

Rarely, the tumor can appear as a multicystic mass in the midline cerebellum, without the classic mural nodule (137). On CT, the cyst fluid appears hypodense, similar to cerebrospinal fluid, but with slightly higher density (109).

Peritumoral edema is typically mild or nonexistent on CT, even with large tumors. Although MRI T2-weighted images are more sensitive than CT to the presence of increased water content, most pilocytic astrocytomas demonstrate little edema. Contrast enhancement is common in these tumors on both CT and MRI (109; 49; 52; 54; 91; 102).

The enhancement is usually homogeneous and is noted uniformly with tumors in all locations except the brainstem, where enhancement is less frequent.

The presence of enhancement is a characteristic feature that differentiates pilocytic astrocytomas from other low-grade tumors (eg, low-grade diffuse fibrillary astrocytomas), which typically do not enhance (109; 52; 54; 38; 102). The explanation for this enhancement is a more dense tumor vasculature and the presence within these vessels of open tight junctions and fenestrae that are more likely to leak contrast material (109; 52; 54). Enhancement of the cyst wall does not appear to affect prognosis (14). Three patients with enhancement of the cyst wall underwent tumor resection and biopsy of the cyst. No tumor tissue was noted in the cyst wall material, and the patients have remained free from recurrence. In a study attempting to correlate MRI and CT imaging features of typical pilocytic astrocytomas with a subgroup of tumors displaying more aggressive clinical behavior, no differentiating characteristics could be determined (183). Overall, MRI with gadolinium is considered more sensitive than CT for detecting small tumors and those located in the posterior fossa. Cerebral angiography will generally show an avascular mass, although occasional tumors may demonstrate hypervascularity (52). One study attempted to define MRI characteristics to differentiate between cerebellar pilocytic astrocytomas and medulloblastomas (08). In this study, Arai and colleagues focused on the signal intensity of the solid portion of each type of tumor. They noted that in 50% of pilocytic tumors, the T2-weighted images had signal intensity equivalent to cerebrospinal fluid (ie, hyperintense). None of the medulloblastomas had similar T2-weighted signaling features. A study by Grand and colleagues evaluated a series of 9 patients with histologically proven pilocytic astrocytomas and studied the utility of perfusion MR imaging (62). In all cases, the relative maximum cerebral blood volume (rCBVmax) was less than 1.5 (mean 1) with a signal intensity curve that overshot the baseline. It appeared the rCBVmax underestimated the true perfusion of the tumors, and the signal intensity curve was likely caused by massive leakage of contrast media into the interstitial space of the tumor. One study has reviewed how often typical pilocytic astrocytomas may present with aggressive features on MRI that can mimic higher-grade gliomas (102). They reviewed the neuroimaging studies of 100 patients with histologically proven pilocytic astrocytomas; 24 of the patients were older than 18 years of age (range 19 to 45 years). The imaging appearance was typical and nonaggressive in 71 cases, and had aggressive features in 29 cases. Features that were considered more typical for a higher-grade glioma included: irregular, thick peripheral enhancement; irregular inhomogeneously enhancing tumor with infiltrative margins; ill-defined, nonenhancing tumor; complex cystic tumor with a peripheral rim of enhancement and hemorrhage; and rapid growth of tumor on serial follow-up MRI studies. Another report analyzed 6 cases of pilocytic astrocytomas with atypical features on MRI that were more typical of high-grade gliomas (139). In comparison to typical pilocytic astrocytomas on MRI, this group had small cyst formation, heterogeneously enhancing tumor nodules, irregularly enhancing tumor nodules, and enhancing tumor nodules with regions of hemorrhage. In addition, several tumors had evidence of leptomeningeal spread in the nearby meninges or ventricles.

MRI studies are attempting to discern between enhancing pilocytic astrocytomas and the more common enhancing high-grade astrocytomas (42). In this study, 16 patients with pilocytic astrocytoma and 22 patients with high-grade astrocytomas were studied using MR imaging, diffusion-weighted imaging, perfusion-weighted imaging, and MR spectroscopy. The relative cerebral blood volume values were significantly lower in the pilocytic astrocytoma cohort (1.4 vs. 3.3; p = 0.0008). ADC values were higher in the pilocytic astrocytoma group (1.5 x 10(-3) vs. 1.2 x 10(-3); p = 0.01). In addition, the lipid-lactate in tumor/creatine in tumor ratios was significantly lower in the pilocytic cohort (8.3 vs. 43.3; p = 0.03). The presence of necrosis favored the presence of a high-grade astrocytoma.

Spinal pilocytic astrocytomas are uncommon but are typically iso- to hypointense on T1 images and hyperintense on T2 and FLAIR images and often demonstrate the presence of a cyst (75). Most of these tumors enhance, but to a lesser degree than similar tumors in the brain. The classic enhancing mural nodule may not be present.

Management

The initial management of newly diagnosed adult patients with pilocytic astrocytomas is similar to that of patients with other brain tumors (142). Seizures must be adequately controlled, preferably with first-line agents like phenytoin or carbamazepine. Monotherapy should always be attempted, maximizing a single anticonvulsant until seizure control is achieved or toxicity occurs, before adding a second agent. Second-line agents include valproic acid, gabapentin, and phenobarbital. If possible, alternative anticonvulsants should also undergo a trial of monotherapy. For patients with elevated intracranial pressure, dexamethasone should be prescribed in the lowest dose that will relieve symptoms. Following stabilization of pressure-related symptoms and initiation of therapy, dexamethasone should be carefully tapered to minimize potential complications (142).

For the majority of patients who develop a pilocytic astrocytoma, the mainstay of therapy will be an aggressive surgical resection (Table 1). Patients with accessible tumors (ie, cerebellar hemisphere, cerebrum, optic nerve) that are completely resected have the potential for surgical cure and typically have prolonged recurrence-free survival approaching 100% (04; 37; 56; 02; 198; 179; 49; 52; 69; 180; 72; 83; 193; 93; 25; 207; 20). Many of these accessible tumors have discrete cleavage planes and tumor margins that are amenable to complete or aggressive subtotal resection. Less often, a more infiltrative or diffuse margin is noted that does not allow complete extirpation. In the report by Forsyth and colleagues, in which a multivariate analysis of prognostic factors was performed, the only significant factor (p less than 01) was extent of surgical resection (52). For patients who underwent complete or radical subtotal resection, survival at 5 years and 10 years was 100%. For those who underwent a subtotal resection, survival was 92% and 84%, respectively; for patients who only had a biopsy, survival was 44% at both intervals. The importance of extent of resection as an independent prognostic factor has been verified by a more recent analysis reported by Bond and colleagues from the Mayo Clinic (20). In a metaanalysis of 7 appropriate studies, gross total resection versus subtotal resection had an odds ratio of 3.46 (P < 0.001) for recurrence in the tumors that were only partially resected. For cystic tumors, the most important aspect of complete resection is removal of the enhancing mural nodule, whether or not the cyst wall can be totally removed (149). Patients with subtotal removal of the mural nodule have more frequent tumor recurrences and often require multiple operative resections. After surgical resection, some authors recommend early follow-up with contrast enhanced CT to assess the amount of residual tumor, which is often more extensive than what is predicted by the surgeon (72). Data suggest that early postoperative MRI is not as accurate as CT in differentiating residual tumor from postoperative changes (167). Because of their location (eg, thalamus, basal ganglia, optic chiasm and tracts, brainstem), many pilocytic astrocytomas cannot be completely resected without severe morbidity. Using advanced stereotactic computer-assisted techniques that are not available in many medical centers, some smaller diencephalic tumors can be radically removed with excellent results (127; 119; 135); however, many of these deep-seated tumors can only undergo limited resection or biopsy. A report by Moshel and coworkers describes a series of 72 patients with thalamic pilocytic astrocytomas who underwent stereotactic volumetric resection (135). A gross total resection was accomplished in 58 patients (80.5%), with worsening neurologic function noted in only 6 patients (8.3%). On long-term follow-up (mean 8 to 13 years), patients with complete tumor resection were likely to remain free of progression. For tumors that are located in the optic chiasm or tract, hypothalamus, or brainstem and that are not diffusely infiltrative, some authors report that subtotal or partial resection can be achieved with acceptable morbidity and potential long-term survival (04; 193; 93; 133). Well-circumscribed optic pathway tumors that do not infiltrate into the optic tracts are considered excellent candidates for surgical resection (193). Potential complications of surgery in this region include endocrine dysfunction, infarction, and further deterioration of vision. Patients with pilocytic astrocytomas of the brainstem that remain focal and noninfiltrative are also candidates for attempted subtotal removal, and may have excellent survival even without radiation therapy (04; 133). A description of 2 cases by Miyamoto and coworkers suggests that in selected patients, pilocytic astrocytomas of the midbrain can be completely resected with minimal neurologic injury (133). Although long-term survival is compromised following a less than complete resection, because of the indolent nature of most pilocytic astrocytomas, overall survival remains excellent when compared to low-grade and high-grade diffuse fibrillary astrocytomas. Pilocytic astrocytomas of the intra- and suprasellar region are excellent candidates for aggressive surgical resection (181). In their description of 3 patients with sellar region tumors, Skipworth and colleagues noted an excellent response to complete resection, with no significant surgical morbidity. A review of surgical results in 20 adults with pilocytic astrocytoma verifies the importance of gross total resection of accessible tumors (207). Complete tumor removal was curative, whereas subtotal removal was associated with higher rates of recurrence. Overall survival and progression-free survival rates were 87% and 60%, respectively.

The role of conventional external beam radiation therapy remains unclear for patients with pilocytic astrocytomas. Although no large controlled clinical trials are available on which to base definitive conclusions, most authors agree that radiation is not indicated for patients after a complete surgical resection (04; 37; 56; 02; 198; 179; 49; 52; 69; 180; 72; 83; 193). These patients should be followed long-term with neurologic exams and CT and MRI for evidence of recurrence. However, whether or not radiation therapy should be administered following a biopsy or subtotal or partial resection remains controversial. Some authors have been unable to demonstrate a significant difference in survival between groups of patients that have undergone radiation therapy versus surgical resection alone (56; 57; 02; 179; 52; 69). In a study of pilocytic astrocytomas of the cerebellum, Hayostek and colleagues found a 10-year survival rate of 80% for 55 patients who had received postoperative radiation (27 of 55 greater than 4500 cGy) versus 82% for 50 patients who had undergone surgical resection alone (69). Similarly, radiation therapy, after incomplete resection of supratentorial pilocytic astrocytomas, failed to demonstrate improved survival over incomplete resection alone in a number of studies, many of which contained adult patients (56; 57; 02; 52). Conversely, other authors suggest that radiation therapy administered to a total dose of 4500 to 6000 cGy in 180 to 200 cGy daily fractions should be considered following subtotal removal of tumor or at the time of recurrence (198; 179; 180; 83). In the study by Shaw and colleagues that evaluated 41 patients with pilocytic tumors, the overall survival was not significantly different between irradiated and unirradiated groups (179); however, in an analysis of the subgroup of 31 patients that received a subtotal resection or biopsy, the trend in survival seemed longer in the 27 patients who were irradiated. Wallner and colleagues state that radiation therapy should be applied to all patients with incomplete resections, based on the concept that irradiation will be more efficacious when administered to a smaller residual tumor volume (198). In a study of 19 patients with pilocytic astrocytomas, all of whom had undergone incomplete resection or biopsy, radiation therapy was considered more effective when used immediately versus after disease progression (96). The 5-year progression-free survival rates were 77% versus 50% (p=o.013), respectively. For those patients in whom irradiation is deemed appropriate, some authors would recommend using stereotactically guided conformal techniques to minimize radiation exposure to normal surrounding brain structures (174). This technique can also be used effectively to treat tumors located in eloquent regions of the brain with minimal risk of morbidity (143). A retrospective study evaluated patients, following biopsy or surgical resection, who had received radiation therapy (N = 11; 5000 cGy) or were followed off-treatment with serial MRI scans (N = 19) (79). The overall survival between groups was similar. However, there was a significant improvement in progression-free survival at 5 and 10 years for the group that received radiotherapy. The 5- and 10-year progression-free survival for the radiation therapy and observation groups were 91% versus 42% and 60% versus 17%, respectively (p = 0.005).

Stereotactic radiosurgery with a linear accelerator or Gamma Knife is another option for selected patients, including those with tumors in the brainstem (18; 65; 66; 114; 192). In 2 series totaling 56 patients, including many adults and teenagers, growth control rates following radiosurgery ranged from 70% to 90%. Tumor shrinkage was noted in many tumors on follow-up MRI imaging. Several reports suggest that stereotactic radiotherapy is a viable option in selected patients with residual or recurrent pilocytic astrocytomas (66; 114). In a study from the Mayo Clinic by Hallemeier and colleagues, 18 patients with a median age of 23 years were treated with stereotactic radiosurgery (median margin dose of 15 Gy) (66). The 5- and 10-year overall survival rates were both 71%, with 5- and 10-year progression-free survival rates of 41% and 17%, respectively. Similar results were noted by Lizarraga and colleagues, with actuarial long-term progression-free and disease-specific survival probabilities of 73.3% and 91.7%, respectively (114). A more recent series of patients reported by Trifiletti and coworkers used Gamma Knife radiosurgery on 28 consecutive patients with progressive pilocytic astrocytomas (192). The median age of the cohort was 17.4 years, with the most common locations being the cerebellum, brainstem, and basal ganglia. The median dose for the group was 16 Gy (range 4-20 Gy). The actuarial progression-free survival at 1, 3, 6, and 12 years was 96%, 96%, 96%, and 80%, respectively. The overall local control rate for the cohort was 93%.

Conventional radiation therapy should also be considered for the uncommon pilocytic astrocytoma patient who develops leptomeningeal metastases (123). The patient with symptomatic disease affecting the cerebrum or cranial nerves might benefit from whole-brain treatment, whereas meningeal tumor affecting the lumbosacral nerve roots or cauda equina might respond to focal spinal irradiation.

In a study of incompletely resected or progressive pilocytic astrocytomas treated with interstitial brachytherapy, Kreth and colleagues report survival similar to more conventional radiation therapy (101). Temporary or permanent low-activity iodine-125 seeds were implanted into 97 patients, 27 of whom were over 18 years of age. The 5- and 10-year survival percentages were 84.9% and 83%, respectively.

On rare occasions, chemotherapy has been used for patients with aggressive, multifocal, or inaccessible pilocytic astrocytomas (147; 145; 154; 24; 144; 110). The majority of experience has been in children with chiasmatic or hypothalamic tumors. Packer and colleagues treated 13 patients less than 5 years of age with progressive disease using 6 8-week cycles of actinomycin D and vincristine (147). After a median follow-up of 4.3 years, 9 of 13 patients had no evidence of progressive disease. In a similar study of older children, 6 patients were treated with a regimen of 6-thioguanine, procarbazine, dibromodulcitol, lomustine, and vincristine (154). Four of 6 patients had partial responses (mean time to progression, 122 or more weeks), whereas another had stable disease (mean time to progression, 252.3 or more weeks). In a study of 7 older children, including 4 with progressive pilocytic astrocytomas, using either carboplatin plus vincristine or thioguanine, procarbazine, lomustine, and vincristine, Heath and coworkers noted objective responses in 2 (1 partial response, 1 minor response, 2 stable disease) (70). The progression-free survival was 71%, with a median follow-up duration of 32 months. In 4 children with chiasmo-hypothalamic tumors, Nishio and associates report responses in 3 with various nitrosoureas (mean time to progression 15.5 or more months), whereas the fourth child improved with a combination of carboplatin and etoposide (time to progression 4 or more months) (144). Brown and colleagues used various chemotherapy regimens to stabilize 11 patients with clinically aggressive pilocytic astrocytomas that had progressed after surgical resection, and in 7 patients they used radiation therapy (24). Five of 11 patients had radiographic improvement after therapy, 3 had stable disease, and another 3 progressed despite therapy. The overall response rate was 75% for either regression or stabilization of disease. However, the median time to progression following chemotherapy, regardless of response, was only 7.5 months. In a study of 12 children with progressive cerebellar pilocytic tumors, Chamberlain used daily oral etoposide (33). Six of the patients responded with tumor shrinkage or stabilization with a median duration of 7 months. Chemotherapy has also been used for children with leptomeningeal spread of pilocytic astrocytomas (126). High-dose cyclophosphamide (4 to 5 gm/m2 per cycle), used every 4 weeks, was able to stabilize 1 patient and reduce tumor burden in 3 patients. In a series of 46 patients (median age 41 years) with progressive low-grade gliomas, Quinn and coworkers included 5 patients with pilocytic astrocytomas in a phase II trial of temozolomide (200 mg/m2 per day for 5 days every 28 days) (158). Three patients had partial responses and 2 had stable disease, with a median progression-free survival of 14 months. Khaw and associates have also used temozolomide (200 mg/m2/day for 5 days every 28 days) in a series of 13 patients with progressive low-grade glioma (9 with pilocytic astrocytoma) (95). Eight patients completed a 12-month course of treatment. There were 2 complete responses, 3 partial responses, and 3 minor responses; the median time to best response was 7.7 months. Median time to progression was 6.7 months (range 1.5 to 41.8 months), with an event-free survival rate at 3 years of 57%. In 1 case report, McLaughlin and colleagues treated a patient with a progressive, refractory pilocytic astrocytoma with imatinib mesylate (STI-571, Gleevec), a PDGFR tyrosine kinase inhibitor (128). The patient had a transient but significant regression of the tumor. It is interesting to note that on pathologic examination, the tumor did not express a significant concentration of PDGFR. An inhibitor of the EGFR, gefitinib, has also been tested against pilocytic astrocytoma cells generated in short-term primary culture (51). All 5 pilocytic astrocytoma cell lines were inhibited by gefitinib. However, the activity of gefitinib did not correlate with the expression of EGFR. Another group has used glioblastoma cell lines and 3 pilocytic astrocytoma explant cultures to study the activity of fluvastatin and/or celecoxib (130). Both drugs alone and in combination were active against the glioblastoma cell lines but were able to induce massive cell death in the pilocytic astrocytoma explant cultures. Based on these preliminary data, a patient with a refractory multifocal pilocytic astrocytoma was treated with fluvastatin and celecoxib, which led to a durable partial response over 18 months.

As noted above, pilocytic astrocytomas are very vascular tumors that demonstrate high levels of transcription and expression of VEGF and VEGFRs. Based on this vascular dependence on VEGF and VEGFR, some researchers have theorized that recurrent and progressive enhancing pilocytic astrocytomas may be good candidates for treatment with bevacizumab -- a monoclonal antibody that targets VEGFA, and has shown benefit in patients with glioblastoma (141). In a report from Wasilewski and Mohile, a series of 4 consecutive patients with recurrent pilocytic astrocyomas were treated with bevacizumab (10 mg/kg every 2 weeks x 6 cycles) (199). The patients ranged in age from 23 to 66 years (median 34 years); 3 of the tumors were infratentorial whereas 1 was in the lateral ventricle. After the completion of the course of bevacizumab, all 4 patients were noted to have significant clinical and radiographic improvement. All of the patients had tumor shrinkage on follow-up MRI scan, including 2 complete responses. In addition, all of the patients were able to tolerate tapering and discontinuation of steroids.

With the presence of the V600E BRAF mutation in a subset of pilocytic astrocytoma, BRAF inhibitors such as vemurafenib and dabrafenib have been utilized with some success.

Special considerations

Pregnancy

Pregnancy does not affect the clinical behavior of pilocytic astrocytomas.

Anesthesia

No anesthetic concerns are specific to pilocytic astrocytomas; rather, concerns are regarding the presence or absence of elevated intracranial pressure during the induction, maintenance, and emergence from anesthesia that are common to surgical treatment of any brain neoplasm (107). In patients with pilocytic astrocytomas and elevated intracranial pressure, care should be taken with premedications to avoid agents that produce excessive sedation and ventilatory depression, as this could exacerbate intracranial pressure. Hypotonic fluids should also be avoided whenever possible. During the induction and maintenance of anesthesia, agents should be chosen that minimize hypertension, cerebral vasodilation, chest wall rigidity, and hypercapnia (107).

Acknowledgement

The author would like to thank Sandra Cottingham MD PhD for her efforts in preparing the neuropathologic slides, and also Nicole Ghaffari and Shawna Huckell for their research assistance.