PNETs often seed within the central nervous system (CNS) via cerebrospinal fluid (CSF). These tumors may be multicentric at the time of diagnosis. In the Children’s Cancer Group study, metastatic spread at the time of diagnosis was found in 19% of supratentorial PNETs (Cohen et al. 1995). Dirks et al. (1996) reported a 38.9% rate of intracranial or spinal dissemination at the time of diagnosis. Other investigators reported a rate ranging from 7 to 12% for spinal seeding in supratentorial PNETs (Yang et al. 1999). Accordingly, CSF analysis can be helpful, if not mandatory, in the primary assessment; in order to discover any metastatic dissemination of these tumors into CNS. CSF analysis is reported to be positive for malignant cells in 9% of the cases (Johnston et al. 2008).

Supratentorial PNETs may even metastasize beyond neuraxis. However, systemic metastases are not frequent, and usually occur at the time of recurrence (Jakacki 1999; Yang et al. 1999). Thus, for the newly diagnosed cases of supratentorial PNETs, preoperative bone marrow analysis and bone scans may be unnecessary and are no longer recommended as part of the staging process, unless clear symptoms and signs of involvement of other locations exist (Johnston et al. 2008).

Surgery is a crucial part of treatment in these tumors and is beneficial for both diagnosis and decompression. Nevertheless, total resection is not always possible in these cases. The preoperative administration of corticosteroids appears to help the surgeon by decreasing the amount of peritumoral edema and ICP. Their preoperative use for at least 24–48 h appears to improve the whole clinical and neurological status of the patient. Similarly, pre-and intraoperative administration of anticonvulsant agents can be helpful for these tumors.

Standard neurosurgical and anesthetic techniques should be used in the removal of these tumors. Careful monitoring of urine output through a Foley catheter is recommended. Since these can be vascularized tumors, transfusion may be necessary. Accordingly, adequate venous access as well as arterial pressure monitoring is important. This is particularly required in young children in whom the amount of blood loss can be significant and may limit the extent of resection. Doppler cardiac monitoring is also required when the head is positioned significantly higher than the heart. Hence, placement of a CV line is strongly proposed in such cases. Air embolism must be considered if there is a sudden decrease of end-tidal PCO₂ or a drop in O₂ saturation.

These tumors are generally purplish gray or pinkish lobulated neoplasms with well demarcated borders along the adjacent normal brain in the majority of cases (Kim et al. 2002). Debulking of the tumor can allow the surgeon to bring in the edges of the lesion to better identify the margins of it. Gross total resection of the tumor is the optimal goal of surgery, but it must be balanced against potential neurological deficits particularly when there is adjacency to eloquent locations. Complete resection can be achieved in only 33–58% of the cases (Johnston et al. 2008; Kim et al. 2002; Timmermann et al. 2002) and as usual, surgeons most familiar with these tumors achieve the greatest amount of resection. The wide involvement of the brain and increased vascularity of these tumors are the main reasons precluding total or radical excision. Use of an ultrasonic aspirator greatly enhances the speed of tumor removal and assists with reducing blood loss. When the tumors are in the vicinity of eloquent regions, performing awake surgeries is another option. Also, surgical approach can be tailored according to preoperative diffusion tensor (DTI) images and concomitant use of neuronavigation. Additional surgical interventions may be needed when there is local recurrence. Stereotactic radiosurgery is another option when the local recurrence is not voluminous (Kim et al. 2002).

Hemorrhage in the tumor bed, brain edema, and new neurological deficits such as hemiparesis are some of the reported complications of surgery (Kim et al. 2002). Cerebrospinal dissemination can be a potential complication of surgical manipulation of these tumors but the reported incidence of this phenomenon is negligible (Kim et al. 2002). In the modern era, the rate of operative mortality is very low (Jakacki 1999; Kim et al. 2002).

Determination of residual disease is best done by MRI performed within 24–48 h after surgery, before any enhancement attributable to postoperative inflammation or gliosis can cloud the imaging of residual neoplasm. Cytologic examination of CSF can also be performed by lumbar puncture about 2 weeks after surgery, after sufficient time has passed to avoid contamination by operative debris. Similarly, intraoperative samples may be taken at the beginning of the procedure. A CSF cytology determination that is positive for tumor cells, either preoperatively or postoperatively, predicts a poor outcome. A negative cytology test, however, does not preclude advanced disease. Spinal MRI should be done after surgery if it was not already performed. Postoperative imaging can occasionally be difficult because of the presence of postoperative debris and blood.

Despite the concomitant use of hyperfractionated craniospinal radiation and adjuvant chemotherapy in the postoperative period, survival is generally poor in supratentorial PNETs particularly in children (Johnston et al. 2008). Local recurrence and CSF dissemination are chief causes of such a poor outcome. Current data suggest that children diagnosed with supratentorial PNETs have a poorer median survival interval than children with infratentorial medulloblastomas (Nishio et al. 1998a; Packer and Finlay 1996; Paulino and Melian 1999; Reddy et al. 2000). It is possibly because of younger age at time of diagnosis and special considerations for radiotherapy or frequent dissemination of such tumors (especially pineoblastomas) in the earlier phase of the disease. The 4-year survival of these tumors is also reported to be 37.7% in children (Johnston et al. 2008). Accordingly, most of the management protocols have included these patients within treatment regimens designed for children with poor-risk medulloblastoma. The prognosis is usually better for adults (Paulino and Melian 1999; Terheggen et al. 2007), in whom the mean survival is as high as 86 months and 3-years survival is 75% (Kim et al. 2002). Another important related issue in the management of PNETs is the quality of life among long-term survivors. It is now well recognized that a significant number of long-term surviving children have noticeable neurocognitive, endocrinologic, and psychological sequelae (Packer and Finlay 1996). However, this does not seem to be true among adults. The karnofsky performance scales (KPS) in the adult survivors is generally more than 70 with the mean follow-up duration of 49 months (Kim et al. 2002).

There is a paucity of articles discussing the prognostic factors of supratentorial PNETs.

Some factors have been proposed as being consistently important for outcome including: adjuvant therapy, age at the time of diagnosis, extent of initial resection, evidence of metastatic disease, histopathology, proliferation markers, and site of tumor. The impact of most of these factors on the outcome is highly controversial and the true prognostic effect of each is yet to be recognized: