Surgery is indicated in all patients in order to achieve a histologic diagnosis, allow for cyst or hydrocephalus decompression, and to minimize disease burden (Van Effenterre and Boch 2002). Some surgeons believe that an aggressive GTR is required for cure, while others believe the morbidity associated with that approach is too great and opt for cyst drainage and a subtotal resection (STR) with planned postoperative radiation therapy (Aggarwal et al. 2013; Fahlbusch et al. 1999; Merchant et al. 2002; Sanford 1994; Stripp et al. 2004; Weiner et al. 1994; Yasargil et al. 1990; Clark et al. 2013; Schoenfeld et al. 2012).

Initial surgical intervention is focused on relieving acute symptoms related to these tumors. In patients presenting with hydrocephalus, decompression of the lesion itself is the favored treatment approach to restore CSF flow (Fahlbusch et al. 1999; Choux and Lena 1979). However in severe cases, an external drain may be required to relieve pressure prior to tumor resection.

The approach to surgical resection is dependent on the location and makeup of the tumor. Historically, a common surgical approach included a right frontotemporal incision (Fahlbusch et al. 1999). Also, suprasellar tumors were also resected using a transcranial approach, while prechiasmatic tumors may be best visualized via a supraorbital craniotomy (Moore 2000). For tumors that are primarily intra-sellar, an endoscopic, transsphenoidal approach may provide optimal visualization while remaining minimally invasive (Elliott et al. 2011; Zona and Spaziante 2006). More commonly, newer techniques including microsurgery, endoscopic assistance, and minimally invasive approaches have allowed neurosurgeons to improve the quality of their resections while minimizing morbidity (Sughrue et al. 2010; Muller 2008, 2013). Ultimately, the intraoperative findings will dictate the degree of resection.

Factors, which most impact the choice of whether to proceed with adjuvant radiotherapy or observation, are most often the presence of residual disease and the age of the patient. For tumors arising in very young patients, a discussion is warranted regarding the appropriateness of observation since very young children are particularly susceptible to adverse radiation effects. However, because incompletely resected lesions have a high rate of local relapse without adjuvant therapy (71–90%), postoperative radiotherapy should be recommended in patients with subtotal resection (Clark et al. 2013; Becker et al. 1999).

Numerous radiation therapy techniques exist and have been utilized with excellent outcomes in terms of disease control for these tumors. It is, however, worth noting that no prospective studies evaluating the benefits of advanced technology exist, in terms of reduced radiation adverse effects, either for photon or proton therapy. In the majority of published literature, photon-based therapies, both traditionally fractionated and hypo-fractionated stereotactic techniques have been employed (Merchant et al. 2002, 2006; Becker et al. 1999; Habrand et al. 1999, 2006; Minniti et al. 2009). Merchant and colleagues conducted a single arm prospective study of reduced margin 3D conformal radiation therapy for pediatric patients with craniopharyngioma (Merchant et al. 2006). In a total of 28 patients, the solid and cystic tumor components along with a 1 cm expansion were targeted to a dose between 54 and 55.8 Gy with acceptable rates of disease control. Greenfield and colleagues have also retrospectively evaluated the use of intensity modulated radiation therapy techniques (Greenfield et al. 2015). These investigators documented high rates of disease control but also noted a high burden of pre-radiotherapy comorbidities including endocrinopathies. More recently, dosimetric studies of volumetric arc therapy (VMAT) have been conducted and indicate that close attention to beam angles is important in order to minimize hippocampal exposure (Uto et al. 2016).

In addition to studies using common 1.8 Gy fractionation schemes, given the relatively well-demarcated tumor boundaries on imaging, radiosurgery, or hypo-fractionated stereotactic radiotherapy have also been explored in the treatment of these tumors. For radiosurgery, the most commonly employed platform has been gamma knife. Patients selected for gamma knife radiotherapy typically have small tumors, measuring less than 3 cm; also, there must be a safe separation (generally 3–5 mm depending on stereotactic delivery technique) between the target and nearby critical structures such as the optic chiasm in order to allow for adequate target coverage while respecting normal tissue constraints. Reported doses delivered in single fractions using gamma knife range from 12 to 14 Gy (Amendola et al. 2003; Kobayashi 2009; Mokry 1999; Chung et al. 2000). In these select patients, reported disease control rates seem acceptable and toxicities minimal (Lee et al. 2014; Park et al. 2011; Xu et al. 2011). However, patient numbers included are small and outcomes for adult and pediatric patients are frequently not separated. Given various differences in patient selection (i.e., tumor volumes, etc.), it is difficult to compare these results to traditional fractionated regimens.

As for many pediatric brain tumors, proton therapy is increasingly used in the treatment of craniopharyngiomas. Boehling and colleagues described in detail the dosimetric advantages of proton therapy including the potential for sparing of vascular structures and the hippocampus, especially with advanced proton therapy delivery modalities, namely intensity modulated proton therapy (IMPT) (Boehling et al. 2012). One of the preliminary reports by Luu and colleagues on the use of proton therapy for patients with craniopharyngiomas showed equivalent local control rates (88%) as compared to previous photon-based studies (Luu et al. 2006). More recently, a multi-institutional study comparing proton therapy to photons, again reported comparable outcomes (Bishop et al. 2014).

While it has long been known that craniopharyngiomas may contain cystic components, only recently has it been appreciated the dynamic changes that occur within these structures. Investigators at the Massachusetts General Hospital were among the first to describe in detail the potential for rapid cystic changes in these tumors (Winkfield et al. 2009). A group of 24 pediatric patients, 19 with a cystic component, underwent treatment. Seventeen of the 19 patients with cysts received surveillance imaging during the course of radiation therapy. Six of these patients had imaging evidence of cyst enlargement with 4 requiring revision of the radiation plan in order to ensure target coverage. The authors recommended routine imaging during the course of radiation therapy in patients with tumors containing cystic components (Winkfield et al. 2009). Following this initial observation, a subsequent report from the University of Texas MD Anderson Cancer Center confirmed relatively frequent cystic changes during irradiation (40%), with some patients requiring alteration to the treatment plan (20%) (Bishop et al. 2014). Therefore, when small CTV and PTV expansions are used, weekly or biweekly imaging during treatment is recommended in patients with cystic components to their tumors in order to ensure adequate tumor coverage. Additionally, close surveillance imaging is of greater importance in the setting of reduced CTV margins and/or the use of proton therapy.

A less commonly employed treatment approach is the intracavitary injection of either β-emitting radioisotopes or sclerosing substances. The radioisotopes (Schoenfeld et al. 2012) phosphorus or ⁹⁰yttrium have been used both as a primary and salvage treatment for cystic craniopharyngiomas, in which they are injected into the cysts via a catheter. Once within the cyst, photons are emitted by beta decay allowing for delivery of high doses of radiation to the epithelial lining of the cyst. Several studies have reported cyst regression in 81–88% of patients (Blackburn et al. 1999; Leksell 1952; Pollock et al. 1995). Alternatively, intracystic catheters can be used to instill chemotherapeutic agents (i.e., bleomycin or interferon α) with some efficacy (Cavalheiro et al. 2010; Schubert et al. 2009).

While disease control rates are high, craniopharyngiomas are associated with decreased overall survival compared to the normal population; the 10-year survival outcomes commonly reported are between 80 and 91% (Schoenfeld et al. 2012; Bishop et al. 2014; Merchant et al. 2013; Karavitaki et al. 2006; Pereira et al. 2005; Fernandez-Miranda et al. 2012). In addition to the deaths attributable to disease progression or surgical mortality, these patients are at higher risk of cardiovascular and cerebrovascular death (Pereira et al. 2005). Notably, based on the current evidence, there does not seem to be a difference in long-term survival between patients treated with GTR versus STR and postoperative radiation therapy. However, a recent SEER study did suggest a short-term survival benefit independently associated with STR (HR 0.39) and RT (HR 0.45) on multivariate analysis (Zacharia et al. 2012).

In terms of surgical resection, GTR rates range widely (between about 30 and 75%), with recurrences most commonly reported to occur in about 8–30% of patients following a GTR (Caldarelli et al. 2005; Fahlbusch et al. 1999; Sanford 1994; Stripp et al. 2004; Shi et al. 2008; Gardner et al. 2008). Conversely, the rate of recurrence following a STR alone is unacceptably high (40–75%) without RT (Fahlbusch et al. 1999; Karavitaki et al. 2005; Villani et al. 1997). However, direct outcome comparisons are difficult given the variable median follow-up and differing definitions of disease progression.

Patients that receive STR and postoperative radiation therapy appear to have similar control rates to GTR with disease relapse occurring in 0–12% of patients (Schoenfeld et al. 2012; Bishop et al. 2014; Merchant et al. 2013). Several studies have also differentiated solid from cystic tumor progression. Immediately following radiotherapy many patients experience transient cyst growth before they regress (Bishop et al. 2014; Shi et al. 2008). Shi and colleagues reported cyst enlargement in 11 of 21 patients following radiation therapy. Similarly, Bishop and colleagues observed a third of patients to have evidence of cyst growth within 3 months of completing treatment. However, in the majority of cases, growth was transient and followed by subsequent shrinkage of the cystic component. Reported 3-year recurrence rates of the solid tumor component were 5% versus 24% for cysts (Bishop et al. 2014). Cyst growth is appreciably more challenging to control, but its biologic significance is undefined.

If tumors recur, management is increasingly difficult primarily because surgical resection of recurrent tumors is more challenging. The recurrent tumors are commonly more adherent to surrounding tissue with poorly defined tissue planes, which is evidenced by the lower rate of resection (13–53%) achieved at the time of recurrence (Fahlbusch et al. 1999; Villani et al. 1997; Duff et al. 2000). For patients that recur, surgery alone is unlikely to provide durable control. In a study by Kalapurakal and colleagues, the 10-year progression-free survival after surgery and radiation therapy for recurrent tumors was 82–83% compared to 0% for surgery alone (Kalapurakal et al. 2000). However, withholding radiation in the definitive treatment setting in order to reserve it for use at the time of salvage is not justified based on available evidence. In fact, a recently published study by Bishop and colleagues showed that recurrent cyst growth was associated with poorer visual outcomes and hypothalamic obesity; furthermore, radiation therapy as salvage therapy negatively affected endocrine function (Bishop et al. 2014). Therefore, reserving radiation for salvage in the setting of STR is not advisable.