Introduction

Craniopharyngiomas arise from remnants of embryonic epithelial cells of the craniopharyngeal duct or from metaplasia of the pituitary stalk [120, 145]. They are histologically benign but locally aggressive in the sellar region and are associated with a high rate of recurrence and morbidity [157]. Indeed, they have the highest mortality among sellar tumors [14, 172] and the 5-year overall survival depends on the age at diagnosis, tumor size, extension, and recurrence rate [15, 44]. More than a third of craniopharyngiomas occur in the pediatric population where they account for 10% of all intracranial tumors, [15] while they account for about 3% of intracranial tumors for all age groups, [122] with a second peak of incidence at 50–75 years [49].

The majority of studies from literature describe the management of craniopharyngioma in children, while publications related to adult-onset tumors are scarce. There has not been as yet any consensus on the gold standard of treatment with therapeutic options tending to differ between centers, with variability in the choice of the therapeutic approach between surgery (trancranial and/or endoscopic/microscopic endonasal) [2, 17, 25, 57, 60, 62, 87, 116, 131], radiotherapy [5, 117] and intracystic chemotherapy [9, 21, 32, 72, 132]. Also, surgical strategy may vary from an aggressive gross total resection or a maximal safe resection with stalk preservation to biopsy/cyst decompression followed by radiation. The overall survival rate in adults is 89–94% at 5 years and 85–90% at 10 years follow-up [92, 175] but this may vary based on the therapy applied.

The aim of this paper is to summarize the most recent evidence on the surgical treatment of adult patients with craniopharyngiomas, while addressing the more controversial points of the management that exist in current neurosurgical practice.

Methods

The systematic review was conducted according to the PRISMA criteria [112]. The literature search was performed using the MedLine database, including literature from January 1970 to February 2019. The search was conducted using the terms “craniopharyngioma” combined with “epidemiology,” “radiology,” “ophthalmology,” “neuropsychology,” “surgery,” “endoscopy,” “microscopy,” “resection,” “radiation,” “recurrence,” “survival,” and “outcome.” Our search was limited to studies conducted in adults. Additional relevant studies were searched in the reference list of identified studies manually, and through the use of the “related article” tool in PubMed. Duplicate studies were eliminated. Three authors (GC, MM, and RTD) independently reviewed abstracts, full-text articles, and citations to select pertinent studies. A PICO question was formulated to guide the selection process: the population was defined as adult patients with craniopharyngiomas, the intervention was any type of surgery performed and outcomes included endocrinological, visual and clinical outcomes, the extent of resection, recurrence rate and overall survival, early and long-term morbidity, and quality of life. Case reports, preclinical studies, and pediatric trials were excluded.

Only studies in English were included. The methodological quality of selected articles was evaluated using the GRADE system [6] without masking the authorship of the article.

A task force composed of the EANS skull base section along with renowned international experts was constituted to formulate evidence-based recommendations. Consensus was elaborated after a systematic review of literature and direct discussion among the experts. If randomized blinded trials or prospective matched-pair cohort studies were identified, the recommendations were Level A or B. For controlled non-randomized trials or uncontrolled studies the recommendations were Level C or “expert opinion,” respectively (Table 1) [67]. If unanimous responses were recorded, we used the sentence: “we recommend.” Divergent opinions were discussed until a consensus was reached and we used the terms: “we suggest.” After each recommendation or suggestion, the literature supporting the assumption was reported, and in some cases, a remarks section that specified some details or technical/practical issues was included [67].

Table 1 Recommendation levels on the basis of the category of the studies
Full size table

At the end of this process, we applied the AGREE Reporting Checklist to review and improve the quality and transparency of the manuscript [13].

The authors used the term gross total resection (GTR) to define a macroscopically complete resection, with no residual tumor visible at the postoperative MRI. When residual tumor is present, the term subtotal resection (STR) is used.

Results

We formulated 17 recommendations in total: 13 recommendations and 4 suggestions.

  1. 1.

    Preoperative management: How to evaluate a craniopharyngioma?

    1. 1.1

      We recommend that all patients with a suspected craniopharyngioma undergo cerebral MRI to evaluate the extension and radiological features of the tumor and a cerebral CT scan and/or CT angiogram to determine the presence of calcification, bone erosion, hyperostosis, anterior skull base anatomy and vasculature pertinent to the tumor and thereby to define the appropriate surgical technique and approach (Level C).

  1. 1.1.1

    Evidence

Craniopharyngiomas may present heterogeneous radiological features in the adult population and cerebral MRI with gadolinium enhancement and dedicated pituitary protocols should be performed. They may appear most frequently as sellar and suprasellar lesions with a variable growth pattern within the parasellar and perichiasmatic spaces. They may present as entirely solid or with an associated cystic component and may frequently engulf neurovascular structures in the interpeduncular and suprasellar cisterns. In adults, they arise most commonly posterior to the chiasm and extend posteriorly into the third ventricle [138]. The papillary histopathological type account for at least one-quarter of cases in the adult age group and this finding translates into more solid lesions with fewer calcifications [4, 198], a homogeneous contrast enhancement and a thickened pituitary stalk [193]. Purely intrasellar lesions or isolated intraventricular tumors are rare and generally associated with the papillary histopathological subtype [33, 83]. The differential diagnosis with a pituitary adenoma presenting with a suprasellar extension and a cystic portion remains open till surgery or after the pathological analysis. However, the solid portion of a craniopharyngioma generally enhances on the contrast MRI, more intensely than a classic pituitary adenoma and the normal gland. The MRI provides useful information regarding the tumor characteristics and the relationship with nearby structures, in particular T2-weighted sequences should be performed in the coronal and sagittal planes to study the relationship with the optic chiasm [179] and the floor of the third ventricle and [94] the mammillary body angle, as well as to determine the extension of the tumor and perilesional edema [151,153,154,154]. The localization of the hypothalamo-hypophyseal tract, as well as its displacement and infiltration, might be also be predicted through the use of a diffusion tensor imaging tractography [180].

A complementary cerebral CT scan with or without angiographic sequences with fine cuts through the sella should be performed. This is the gold standard for intralesional calcification and may be helpful in defining cystic portions of the tumor and bony details such as bone erosion, hyperostosis, sphenoid sinus pneumatization, and septations to plan an endoscopic approach. CT angiogram can be added in cases where there is a high suspicion for arterial or venous involvement or in selecting an appropriate surgical approach where there are potential issues with cerebral or cranial base vasculature.

Remarks:

Several surgical classifications have been described based on the relationship of the tumor with the surrounding anatomical structures [74, 94, 142, 143, 155, 156, 168, 191].

Craniopharyngiomas may be classified according to their site of origin and competence of the diaphragm sella. Yasargil et al. classified craniopharyngiomas as purely intrasellar, intra- and suprasellar, supradiaphragmatic parachiasmatic extraventricular, intra- and extraventricular, paraventricular, and purely intraventricular lesions [186]. Wang et al. simplified this classification dividing craniopharyngiomas into subdiaphragmatic with competent diaphragm sellae, subdiaphragmatic with incompetent diaphragm sellae, and supradiaphragmatic [179, 191]. Hoffman et al. also classified craniopharyngiomas on the basis of their relationships with diaphragma sellae, optic chiasm, and third ventricle into subdiaphragmatic, subchiasmatic, prechiasmatic, retrochiasmatic, and intraventricular [75]. These classifications consider the growth pattern of the lesion and the displacement of the optic apparatus, essential factors in the choice of the surgical strategy.

Kassam et al. classified craniopharyngiomas in relation to the infundibulum to plan their surgical access: the preinfundibular, trans-infundibular, and retroinfundibular types can be accessed through the endonasal route while the purely intraventricular subtype requires a transcranial access [94].

The invasiveness and adherence to the hypothalamus should also be carefully analyzed at the preoperative MRI. This was the basis for Puget’s classification, [155] as it represents the most important factor determining the extension of resection and it helps in predicting and limiting the postoperative hypothalamic morbidity [154, 155].

Although these classifications are useful in planning surgical approaches and proposed extent of resection, they are difficult to apply when the tumors are very large. Furthermore, to the best of our knowledge, there is no classification that considers all these factors globally to orient surgeons in choosing between endonasal and trancranial approaches.

  1. 1.2

    We recommend that patients with craniopharyngiomas undergo a complete history and physical examination including the search for hypopituitarism and hypothalamic compromise. Complementary blood tests should be performed to rule out any endocrinological deficit (Level C).

  1. 1.2.1

    Evidence

A complete physical examination should be performed for clinical signs of hypopituitarism or a hypersecretion syndrome that would orient the diagnosis toward a different pituitary pathology. Pituitary function should be evaluated in all sellar/suprasellar lesions even in the absence of clinical manifestations. Anterior panhypopituitarism is recorded in the majority of craniopharyngioma patients [44, 93, 147]. Hypersecretory syndromes should be systematically searched for all sellar lesions through the assessment of serum TSH, free T4, ACTH and morning cortisol, HGH and IGF-1, prolactin, LH, FSH and progesterone, estradiol, and testosterone levels. Diabetes insipidus should also be excluded through a thorough clinical history to search for polydipsia, polyuria, and nocturia, as well as a complete physical evaluation to look for signs of dehydration and biological examination to evaluate if hypernatremia and a low urinary specific gravity and osmolarity are present. The prevalence of diabetes insipidus may be higher than 50% [29].

Furthermore, all patients with craniopharyngiomas should undergo evaluation to determine if hypothalamic functions such as body weight, temperature regulation, and sleep-wake cycles are preserved [55, 199]. The preoperative evaluation of hypothalamic functions deserves increased attention in future studies to evaluate the morbidity and impact of any therapy we apply.

  1. 1.3

    We recommend that all patients with craniopharyngioma undergo an initial ophthalmological evaluation with visual acuity, visual fields and optical coherence tomography (Level C).

  1. 1.3.1

    Evidence

Examination of the visual function is an important element in the diagnosis, monitoring, and prognosis for craniopharyngiomas. Visual decline is one of the most frequent initial manifestations in patients with craniopharyngiomas [150], accounting for two-thirds of clinical presentation in adult-onset tumors [31]. As many as 80% of patients have a visual deficit at the preoperative ophthalmologic evaluation [159, 183]. Visual acuity and visual field defects as well as optic nerve atrophy and papilledema should be carefully evaluated by a neuro-ophthalmologist at diagnosis, because in addition to direct involvement, associated conditions such as hydrocephalus and intracranial hypertension can secondarily impact visual function [31]. Visual loss may be rapidly progressive, and if present, it should motivate an emergent surgery, as it can rapidly lead to permanent blindness [84]. A younger age at diagnosis, edema of the optic nerve, and tumor recurrence are risk factors associated with visual decline in the pediatric population [178]. Data on the adult population are scarce but we can assume that the risk factors are similar.

Optical coherence tomography may represent a valuable tool to evaluate visual damage and predict visual recovery, especially in non-compliant patients [129]. Its utility was widely described in the pediatric population [11, 123] and these findings could also be applied to adults.

We suggest that all patients with craniopharyngiomas undergo a neuropsychological assessment (Expert opinion)

  1. 1.4.1

    Evidence

The impact of craniopharyngiomas on neurocognitive functions in pediatric patients has been studied. Executive functions, memory and learning and fine-motor coordination were impaired in a proportion of pediatric cohorts, where presurgical hypothalamic involvement and impaired visual status were identified as poor prognostic factors [56, 140]. Data in the adult population are limited but it might be of interest to assess neuropsychological functions in the preoperative period and to compare these data with those obtained in the postoperative period. Furthermore, since a significant proportion of patients are subjected to adjuvant radiotherapy a baseline assessment enables tracking neurocognitive changes that may occur after radiation. Some authors reported no changes in neuropsychological performance after surgical procedures [78] while others describe improvement of preoperative cognitive dysfunction and memory [37]. However, reported postoperative morbidity includes cognitive dysfunction with attention deficit, impaired episodic memory, and processing speed [50, 140], which were associated with microstructural alterations of the white matter tracts detected with the use of DTI [51]. A baseline cognitive assessment would thus be helpful in the long-term follow-up of these patients.

  1. 2

    Surgical approach: endoscopic or transcranial?

The choice of the approach will depend to a large degree on the anatomy of the tumor, its extensions, its consistency and last, but not least, the preferences and experience of the surgical team. In general, the best approach should be able to provide the maximal surgical exposure and the most direct trajectory to the tumor limiting brain retraction and minimizing manipulation of neurovascular structures. The best results of surgical treatment depend on the quality of the first surgery and a careful consideration of the ideal approach needs to be taken at initial presentation in order to obtain the best oncological result while minimizing the potential complications associated with craniopharyngioma surgery.

It should be kept in mind that craniopharyngiomas do not always have an identifiable arachnoidal layer separating the tumor from surrounding neurovascular structures unlike meningiomas, schwannomas, and pituitary adenomas that usually do. Moreover, especially at the level of the floor of the 3rd ventricle, craniopharyngiomas may grow in a true subpial fashion making safe GTR of the tumor impossible. These two surgical facts make surgery of craniopharyngiomas extremely challenging if preservation of function is sought and this should be preeminent in the surgeons’ mind while operating on these tumors.

We recommend the use of transcranial skull base approaches for craniopharyngiomas presenting an extension lateral to the internal carotid artery (Level C)

Evidence:

Multiple skull base transcranial approaches have been described to safely resect craniopharyngiomas [116]. These may vary from midline approaches (subfrontal, anterior interhemispheric), anterolateral approaches (pterional, orbitozygomatic, and suprabrow craniotomy) and lateral approaches (subtemporal, transcavernous petrosectomy) [3, 46, 47, 62, 115, 160]. Furthermore, the transcallosal (transcortical/interhemispheric) approaches are used for tumors with a significant intraventricular component [45]. Combined approaches in the same sitting or staged fashion have also been described [191].

The medial subfrontal approach, through the use of a unilateral frontal [160] or a bifrontal transbasal approach [47], offers a direct view of the prechiasmatic space and of the optico-carotid cisterns bilaterally [115] and may be used to perform a trans-lamina terminalis approach to resect craniopharyngiomas with a sellar and prechiasmatic origin and anterior third ventricular extension [115]. In the large series of Du et al., GTR was achieved in 94% of cases treated with this approach with a 12.7% of recurrence and > 90% of surviving patients were living independently at the end of the study [39].

The anterior interhemispheric approach offers not only a direct view of the prechiasmatic space and of the optico-carotid cisterns bilaterally but also to the third ventricular extension via a trans-lamina terminalis approach. The narrow space between the bridging veins (usually < 20 mm in width) is generally sufficient for tumor dissection and this access allows the control and preservation of the anterior communicating complex, mammillary body and fornix, and midline vessels of the interpeduncular cistern [80]. Indeed, small feeding vessels entering the tumor can be identified, coagulated, and cut under direct visual control at the beginning of tumor resection [80].

The lateral subfrontal (pterional) approach allows a large exposure of the suprasellar area through a subfrontal and transsylvian route [62, 191], where the prechiasmatic, optico-carotid, and carotid-oculomotor triangles are used for tumor access [190]. The addition of an orbital or orbitozygomatic osteotomy may increase the access to the suprasellar area and interpeduncular cistern [63].

Complete resection is reported in a variable percentage of cases: Van effenterre et al. described GTR in 59% of cases (operated through a frontolateral approach in 92% of cases) [175], while Gerganov et al. reported GTR in 87.5% of patients with extensive craniopharyngiomas treated through a pterional approach, with no significant postoperative morbidity [62]. Yasargil et al. reported GTR in 90% of cases of their surgical cohort of 144 patients mainly treated through different transcranial approaches (90%) chosen according to tumor’s characteristics [191].

However, traditional transcranial approaches have well-described constraints such as the need for a variable amount of brain retraction and dissection through multiple, long, narrow corridors across major neurovascular structures [165]. The optic apparatus is positioned between the surgeon and the target and the superior pole of the tumor is often situated in the operative blind spot. Also, the vascular supply from the superior hypophyseal arteries to the optic apparatus lies relatively hidden during a lateral approach. Endoscopic assistance can partially solve this problem with angled keyhole endoscopes.

For craniopharyngiomas with a lateral extension beyond the supraclinoid carotid arteries and into the middle cranial fossa, transcranial approaches remain the first choice [165] The technical difficulties in performing very lateral endoscopic approaches associated with the risk of damage to the carotid, posterior communicating artery or thalamoperforators that lie just behind Liliquist’s membrane, make expanded endonasal approaches (EEA) less appealing in these specific cases. Intraoperative vascular injury can be more easily managed through a craniotomy as opposed to an endoscopic approach.

Remarks:

The situation where this limitation of the EEA can be overcome is when the lateral extension of a midline craniopharyngioma is cystic. Decompression of a small cystic component may bring its lateral wall into the view afforded by the endoscope enabling removal and limiting the chances of recurrence.

We recommend performing transcranial approaches for tumors primarily arising from the floor of the 3rd ventricle with an intraventricular location and not extending to the pituitary stalk or the suprasellar space, where the endonasal route can also be performed (Level C)

Evidence

Intrinsic intraventricular craniopharyngiomas account for 3–11% of all craniopharyngiomas [192]. Although the transventricular corridor through either a transcallosal or trancortical-transforaminal approach is generally considered as first surgical options for craniopharyngiomas arising from the floor of the third ventricle and not extending to the pituitary stalk or to the suprasellar space, the interhemispheric transcallosal approach might be associated with less tissue damage and incidence of seizures [105, 192]. From the lateral ventricle, entry into the third ventricle may be performed through the interforniceal, transforaminal, or subchoroidal routes [2]. However, with these approaches, the visualization of the underlying optic apparatus might be limited in some cases and there might be an increased risk of visual deterioration [192]. Alternatively, the subfrontal trans-lamina terminalis approach can be performed to directly access lesions situated in the lower half of the third ventricle, either through a medial subfrontal or a lateral subfrontal approach with a pterional craniotomy [105, 192].

Remarks:

  • The presence of a tumor projecting through the foramen of Monro leads preference to a transcortical/trans callosal-transforaminal approach [45].

  • For purely intraventricular craniopharyngiomas the endoscopic endonasal route has been described in some selected cases with encouraging results [24, 52, 135]. The more suitable cases are those with protrusion of the floor of the third ventricle over the suprasellar space and who already have hypopituitarism. The pituitary stalk and gland function will be most likely compromised through the endonasal route. The preoperative presence of empty and deep sella may also favor the endonasal route.

We suggest performing an expanded endonasal transsphenoidal approach as first-line surgical approach for midline and retrochiasmatic craniopharyngiomas without lateral extension (Expert opinion)

Evidence

Expanded endonasal approaches (EEA), through a transtubercular approach (with or without a transplanum route), allow an excellent exposure of the tumor without crossing crucial neurovascular structures [17, 25, 26, 34, 37, 57, 60, 61, 94, 107]. Furthermore, the direct access through the endonasal route has the advantage of early identification of the superior hypophyseal arteries, which are important for the optic apparatus and also of early identification of the chiasm and optic nerves thus avoiding retraction not only of the nerves but also of the basal frontal lobes. The EEA provides an excellent access from the sellar component of the tumor till the floor of the third ventricle allowing resection of the suprasellar, ventricular, and interpeduncular extensions [25, 26, 60, 94, 107, 165].

Since the 1980s, the surgical excision of supradiaphragmatic craniopharyngiomas by using an extended transsphenoidal microsurgical approach has been reported [40, 47, 90, 95, 97, 106], there is no doubt that the use of the endoscope has permitted to overcome most of the limitations of this route in terms of visualization and surgical maneuverability

Several studies have compared the results of transcranial microsurgical versus endoscopic transsphenoidal resections for midline craniopharyngiomas.

Jeswani et al described a similar extent of resection in the 2 subgroups (86% with endoscopic EEA versus 91% with transcranial approaches, p = 0.77) as well as a similar PFS and recurrence rate. Although the rate of CSF leakage was higher with EEA, the rate of cranial nerve injury was higher with transcranial approaches [87]. Moussazadeh et al. reported a higher rate of GTR after endoscopic EEA than transcranial approaches (90 vs 40% respectively, p = 0.009). Endoscopic EEA was also associated with an improved visual outcome (p < 0.05), fewer recurrences and complications (p < 0.001) [131]. In both these studies, patients presented with lesions of similar radiological characteristics [87, 131]. Wannemuehler et al. reported a similar extent of resection between endoscopic EEA and transcranial approaches, with a higher rate of visual improvement in the EEA group (89% vs 25%, p = 0.0075) [182]. Komotar et al. reviewed the surgical series of pediatric and adult craniopharyngiomas and they showed that a greater rate of GTR (67% vs 48%, p < 0.003) and an improved visual outcome (56% vs 33%, p < 0.003) was achieved with endoscopic EEA compared to transcranial approaches [103].

Recent literature is replete with numerous studies that strongly support the use of endoscopic techniques as the approach of choice for suprasellar and retrochiasmatic craniopharyngiomas because of the ability to achieve a high percentage of patients with complete resection (similar to transcranial approaches) while limiting the incidence of neurological and vascular complications [17, 22, 23, 25, 26, 35, 47, 57, 60, 61, 79, 85, 94, 103, 107, 118, 189]. Due to the direct access obtained with ventral approaches, the postoperative morbidity in terms of cranial nerves palsy and postoperative seizures is lower after endoscopic EEA [87, 131]. Through EEA, a careful dissection is possible aided by a direct visualization of the capsule, major vessels, and perforators and also of the walls of the third ventricle (hypothalamus). It is possibly due to this that the preservation of hypothalamic functions is superior with endoscopic EEA when compared to the transcranial approaches [4, 65, 189]. A direct decompression of the optic apparatus may also be performed with a direct visualization and preservation of the superior hypophyseal artery [37, 164]. This may translate into a greater improvement of the postoperative visual outcome [65, 103, 131, 182]. In fact, between 75 and 89% of patients having a preoperative visual impairment showed a recovery in endoscopic series [16, 107, 110, 182]. A prefixed chiasm and the presence of a large tumor extending upward and behind the optic chiasm have been considered for long as relative contraindications to achieve GTR through endoscopic EEA. Nevertheless, recent literature indicates that a narrow corridor between the top of the pituitary gland and the bottom of the chiasm has no relationship with the extent of resection obtained endoscopically [138] and that endoscopic EEA is an effective approach also for retrochiasmatic craniopharyngiomas, even in cases with a low-lying chiasm [99].

Progressive debulking through piecemeal removal before embarking on arachnoid dissection remains an important principle as in other microsurgical procedures. Some skull base surgeons affirm that the visualization of certain portions of retrochiasmatic craniopharyngiomas situated in the interpeduncular cistern and retroinfundibular space may be better visualized and removed through transcranial approaches. However, transcranial approaches may present some disadvantages: through the pterional approach, the contralateral optico-carotid and the hypothalamic surfaces are poorly visualized. The subfrontal approach allows a direct visualization of the third ventricle and hypothalamus but the region beneath the optic chiasm may not be properly visualized [102], and thus, the sellar portion of craniopharyngiomas may not be properly visualized. The petrosal approach was also proposed for large retrochiasmatic craniopharyngiomas [3] but it is technically demanding and time-consuming, not to mention the necessity of retracting the temporal lobe and the difficulty in accessing the most upper portion of the third ventricle.

All these pitfalls of transcranial approaches are avoided with EEA, which provides a ventral approach with direct visualization of the optic chiasm, third ventricle ,and hypothalamus and is thus suitable for the resection of retrochiasmatic craniopharyngiomas [94, 113]. When the craniopharyngioma presents an extension into the interpeduncular fossa, a superior clivectomy with or without posterior clinoidectomy and temporary displacement of the pituitary might be necessary and may help in avoiding petrosectomy.

Remarks:

  • Craniopharyngiomas should be preferentially treated in tertiary referral centers [20]. The choice between transcranial or endoscopic approaches should be based only on the tumor anatomy. Ideally all tertiary referral centers should have equivalent expertise with transcranial and endonasal approaches which will allow the choice of the approach to be independent of the personal preferences of any given surgeon. Size, location, relation to vascular and nervous structures, and tumor consistency should be carefully evaluated preoperatively to choose the best surgical approach.

  • The presence of extensive peripheral calcification may favor the use of transcranial approaches. The performance of an endoscopic dissection of a large calcified craniopharyngioma may be risky, as the basilar artery with the posterior cerebral arteries and its perforators are all situated posterior to the tumor and a blind posterior dissection will need to be performed in this area with endonasal approaches. This is associated with a dramatic increase in the risk of vascular injuries. The transcranial approaches seem to allow superior vascular control with large calcified tumors [41, 197].

  • The accurate knowledge of endonasal anatomy is a fundamental step to safely perform the procedure and avoid serious neurological morbidities.

  • The four-hands technique is used in many centers but the endoscope-holder may represent a valid alternative for bi-manual endoscopic EEA [25, 57, 110, 141, 189].

  • Another key point of EEA is the correct management of perioperative CSF leakage and the prevention of a postoperative one. Multiple skull base reconstruction techniques have been described to address this point [1, 28, 69, 71, 82, 109, 110, 144, 171, 177].

  • The use of multiple working corridors should be considered for the resection of giant craniopharyngiomas extending to multiple anatomical compartments. They should be addressed through the use of combined endonasal and transcranial approaches to obtain a maximal resection while limiting the complication rate. The medial and retrochiasmatic portion of the tumor should be addressed through endonasal approaches, while the portion lateral to ICA bifurcation should be approached through standard transcranial approaches.

We recommend the use of traditional endonasal transsphenoidal approaches for purely intrasellar craniopharyngiomas (Level C)

  1. 2.4.1

    Evidence

The endonasal approach can be tailored according to the surgeon’s needs and tumor characteristics. A standard endonasal endoscopic or microscopic transsphenoidal approach (similar to pituitary adenoma surgery) may be used in selected sellar lesions with a limited and well-defined suprasellar and retrosellar extensions [18, 26, 61]. An enlarged pituitary fossa and a cystic extra-arachnoidal infradiaphragmatic component favor the use of a standard transsphenoidal approach [26].

We recommend performing a careful closure with a nasoseptal flap to limit the risk of postoperative CSF leakage when an extended endonasal approach is performed (Level C)

Evidence

performed for craniopharyngiomas often requires a large bone and dural exposure and the presence of a postoperative CSF leakage is common after a large arachnoid dissection and in cases where a wide opening of cisterns and third ventricle is necessary for tumor resection. CSF leakage has limited the widespread application of endonasal approaches for skull base tumors for a long time [104, 177]. The repair failure rate and postoperative meningitis rates in endoscopic skull base tumor surgery have varied over the last 15 years, but the trend has been encouragingly downward [28].

A multilayer reconstruction with the use of abdominal fat and or a fascia lata patch (if possible around a bone buttress according to the gasket seal closure) [109], combined with a vascularized flap [69] and intercalated by a sealant materials, is the most common technique used to limit CSF leakage and it has been largely described in literature [28, 69, 71, 171]. Surgical series using this technique report a rate of postoperative CSF leakage variable from 23% to less than 4% [25, 28, 42, 107, 110, 121, 141].

Remarks:

  • The nasoseptal flap is the most used and preferred option because of its large surface and excellent vascular supply with a long pedicle, which allows the correct positioning. Careful placement of the flap is very important, to avoid dead space behind the flap and torsion.

It should be of adequate size and it should be positioned over a bare bone [1] to prevent a postoperative mucocele formation and also to diminish the risk of flap dehiscence [42].

  • The middle or inferior turbinate flap may represent a valid alternative when the nasoseptal flap fails or cannot be performed [28]. Other pediculated flaps, such as the transpterygoid, temporo-parietal fascia, transfrontal pericranial, and Oliver palatal flaps, were described [53, 54, 68, 137, 196]. The Janus flap with a bilateral nasoseptal flap was also described as a safe alternative to cover large bone and dural defects, which are not completely sealed by a unilateral nasoseptal flap [136].

  • Even with revision surgeries for recurrence, the nasoseptal flap should be spared and reused if possible. The use of acoustic Doppler ultrasonography and indocyanine green fluorescence may represent good tools to assess the presence of a viable vascular pedicle [96, 148]

  • The factors correlated with failure of the skull base reconstruction are obesity, lack of buttress and postoperative Valsalva manoeuver [28].

  • The systematic use of a lumbar CSF drain after a multilayer skull base reconstruction is controversial. Some authors affirm that its use may facilitate the healing of the skull base reconstruction [25, 110], while others claim that it carries an infectious risk [28] and may predispose to intracranial hypotension and pneumocephalus, thus limiting its use to persistent CSF leakage after reconstruction [42]. We would suggest their usage based on surgeon experience/preference for primary repair and for all cases following a secondary repair after a persistent postoperative CSF leak.

Extent of resection and hypothalamic involvement: GTR or STR plus radiotherapy?

Craniopharyngiomas have a locally aggressive behavior and the stalk and hypothalamus often have difficult dissection planes. The aim of surgical management is to obtain a safe maximal resection while limiting postoperative morbidity [25, 107, 110].

We recommend performing a GTR when there is no infiltration of the hypothalamus (Level C)

Evidence

A complete resection at first surgical attempt is described as the most effective treatment from an oncological perspective as the treatment of a recurrent lesion may be more complicated [47, 191]. The priority of surgery is to maximize resection while preserving the patient’s long-term functional outcome and quality of life. A balance should be found between tumor removal and damage to nearby critical neurovascular structures [7, 92]. When a dissection plane between the tumor and the hypothalamus is present, GTR should be attempted to limit the long-term risk or recurrence [92, 167, 191] but surgeons should be mindful about tumors with microscopic subpial invasions of the hypothalamus and to perform GTR in such cases may risk injury to the hypothalamus.

We recommend performing STR coupled with adjuvant radiotherapy (STR + XRT) when hypothalamic infiltration is confirmed (hypothalamic-sparing resection) (Level C)

Evidence

The main limitation in performing GTR is the presence of hypothalamic invasion [151], defined as the absence of a surgical plane between the tumor and the hypothalamus. This is the most important predictor of postoperative morbidity and mortality, [151] as the postoperative quality of life should remain a priority.

The handling of the pituitary stalk during surgery is controversial, as its preservation can limit the risk of postoperative endocrine deficits and diabetes insipidus, but this is known to increase the risk of craniopharyngioma recurrence [81, 88, 89, 167, 186, 187, 191]. Patients with hypothalamic disturbances and hypopituitarism have an elevated prevalence of metabolic syndrome and the mortality rate from cardiovascular causes is increased [86, 125]. According to Sughrue et al, patients with GTR experience a 2.5-fold increased risk of developing at least one endocrinopathy compared with patients undergoing STR + XRT and over 10% with GTR had panhypopituitarism [170]. Intentional STR + XRT has thus gained some favor in recent years due to a reduction in morbidity and possibly an equivalent progression-free and overall survival compared to GTR [162, 170, 188] A recent meta-analysis on 759 cases of adult craniopharyngioma showed that despite the recurrence rates favoring GTR, the difference between GTR and STR + XRT did not reach statistical significance (p = 0.18) [31]. The same findings were reported in Zacharia’s meta-analysis on 644 patients, where no survival advantage was associated with GTR [195]. The absence of a clear superiority of GTR in terms of outcome improvement and the higher complication rate of GTR in terms of endocrinological and hypothalamic dysfunctions, associated with greater attention to the quality of life of patients, are therefore changing the paradigm of treatment [181]. The increased use of EEA seems to have improved the rates of achieving hypothalamic preservation regardless of the degree of involvement by the tumor, principally due to the increased visibility of the hypothalamic dissection plane with the tumor [189].

Remarks:

  • Although the choice to perform intentional STR + XRT is based on the preoperative analysis of radiological features, involvement of neurovascular structures, surgeon’s preference, and experience, the final decision should be made upon surgical exploration and intraoperative findings [181].

  • Aside from fractionated adjuvant radiotherapy, radiosurgery can also represent an attractive option as an adjuvant therapy after STR [27, 64, 100, 101, 108, 127, 163, 174].

How to treat recurrences and residual tumor progression?

An individualized approach is highly recommended for recurrent tumors and residual tumor after STR + XRT (Level C)

Evidence

Multiple options have been described in the literature to manage recurrent tumors and residual tumors after STR + XRT. A watch and wait strategy, second surgery, radiotherapy, intracystic chemotherapy, and even immunotherapy or target therapies have been described in patients with craniopharyngiomas harboring BRAF V600E mutation [12, 48, 66, 72, 73, 158]. The timing to perform these different options is, however, a matter of debate [12, 48, 72, 73, 158] The tumor progression history is the key factor to consider the best tailored appropriate treatment option for the patient, and the management plan should be based on a multidisciplinary discussion. If a small calcified residual tumor is present, showing no growth at follow-up images, a watchful wait and scan strategy can be adopted, while with rapidly growing residual tumors, an early surgical procedure may be preferred. [25, 61, 110, 173, 181].

Long-term clinico-radiological follow-up is mandatory to evaluate the evolution of residual disease.

We suggest applying the similar surgical principles detailed for primary lesions to the treatment of recurrent craniopharyngiomas that require surgical treatment (Expert opinion)

Evidence

Recurrences are reported in a variable percentage of cases even after GTR (0-62%) [91,93,93] and the surgical approach does not seem to have an impact on the recurrence rate [31]. In addition, the initial tumor size has not been correlated to the recurrence rate [146]. Surgery for recurrent lesions is generally considered to be more difficult than primary surgeries because of the loss of the arachnoid planes and the iatrogenic creation of scars and adhesions [173]. The size and the location of the tumor residue are the main determinants for the choice of the surgical approach [114, 116].

The rates of GTR are significantly inferior to results obtained after primary surgery [23, 92, 185, 191] while the morbidity and mortality rate is considerably higher [23, 92, 126, 130, 185]. GTR for recurrent tumors may vary from 0 to 56% according to the different series considered [47, 91, 92, 169, 173] and only one study reported a GTR rate as high as 78% after EEA for recurrence [37]. The postoperative mortality may vary from 11 to 24% [47, 91, 92, 169]. and is higher in the adult population than in children [92].

The same principles detailed for primary tumors should be applied for recurrent/progressive craniopharyngiomas:

  • EEA should be strongly considered (with the exceptions previously described) for all midline lesions not crossing the lateral margin of the carotid arteries and optic nerves.

  • Endonasal approaches should be considered when the first surgery was performed transcranially [23, 36, 98, 173].

  • Transcranial approaches should be preferentially considered for recurrences in the middle and posterior fossa and lesions limited to the third ventricle [173].

Remarks:

  • The first surgical attempt remains the best option to obtain a surgical cure.

  • The use of intraoperative technologies such as image guidance may help in redoing endonasal approaches where the classical landmarks have been removed.

  • The use of radiotherapy or radiosurgery [117, 134] and of intracystic agents [132] should be discussed in a multidisciplinary meeting and should be considered with cystic lesions, progression of the residual tumor, or, in cases, of tumor recurrence if no further surgery is possible [59, 173].

Postoperative management

The different potential complications of craniopharyngioma treatments necessitate each a specific follow-up.

We recommend a close clinical and endocrinological follow-up in the management of patients with treated craniopharyngiomas (Level C)

Evidence

A clinical follow-up should detect early and late complications such as seizures, hydrocephalus, cerebrovascular injuries, and radiation-induced complications [43].

The development of endocrinological dysfunction is very likely to occur in the early and late postoperative period [16, 25, 57, 107, 110]. Aggressive resections, particularly in the setting of stalk invasion, may increase the rate of postoperative endocrinological deficits [37, 107, 111]. About three-quarters of patients will need a long-term hormonal replacement therapy with growth hormone (GH) and thyroid hormone deficiencies being most frequently observed [92, 158]. GH deficiency is associated with an increased cardiovascular risk [128] and a physiologic substitution seems to have beneficial effects on body fat mass, cholesterol profile and blood pressure [119, 124]. The impact of thyroid substitution is less known but subclinical hypothyroidism may increase the cardiovascular risk [139]. A physiologic substitution of adrenocortical deficit is an important determinant in lowering mortality [10, 166]. Diabetes insipidus is another frequent complication and is permanent in more than half of the cases after surgery [133, 149]. Correct management of dysnatremia represents a priority during the immediate postoperative course as ADH secretion may follow a triphasic course.

However, the complication most difficult to treat remains hypothalamic damage, which may strongly impact the quality of life of these patients. This may present as an obesity-hyperphagia syndrome, sleep cycle disturbances, and temperature dysregulation or behavioral abnormalities [30, 76, 77, 133].

Lifestyle and dietary modifications, antihyperlipemic agents, psychotherapeutic, and bariatric surgery have been used to treat hypothalamic obesity with controversial results. Of late, new agents targeting the hypothalamus (if partially damaged) or other brain or peripheral receptors (if the hypothalamus is completely destroyed) are being tested to improve the management of these patients [176].

Remarks:

  • We recommend a clinical and endocrinological checkup every 3 months during the first postoperative year to test the different pituitary axes and then a 6-monthly follow-up during the first 5 years postoperatively. After this period, if the patient is stable, an annual follow-up should be performed.

  • We recommend regular checking of body weight, beginning in the immediate postoperative period, as in some cases, weight gain may be rapid and difficult to reverse.

  • We recommend a regular check of cardiovascular and cerebrovascular risk factors, as they are strong determinants of the increased mortality in craniopharyngioma patients [14, 147, 172, 184, 194].

We recommend a close ophthalmological follow-up in the management of patients with treated craniopharyngiomas (Level C)

Evidence

Visual complications may strongly impact the quality of life of patients with craniopharyngiomas. Eighty percent of patients may present with a preoperative visual deficit [159, 183] and an improvement is described in 40–60% of patients in the postoperative period, while a visual worsening is described in 5–30% of cases [60, 61, 116, 191]. A regular ophthalmological follow-up should be performed to evaluate the evolution of the visual status as well as for early detection of a tumor recurrence.

Remarks:

We recommend an ophthalmological checkup every 6 months during the first year postoperatively, and annually thereafter.

We suggest assessing neurocognitive functions in the postoperative period (Expert opinion)

Evidence:

Hypothalamic damage is known to be associated with a neurocognitive decline [58, 149] and this impairment, along with memory deficits, increases the postoperative morbidity [19, 38, 147] with consequent reduction in the quality of life [149]. Up to 50% of patients present psychosocial impairment at long-term follow-up due to problems with concentration, memory. and executive function [147]. Recent studies affirm that neurocognitive functions are not impaired when a careful removal of the tumor is performed [78, 93] but these studies had a heterogeneous population in terms of hypothalamic involvement and neuropsychological tests applied and therefore these results are still a matter of debate [50, 51, 140].

Remarks:

A postoperative evaluation and follow-up would help in supporting patients in their daily life activities and in improving their capacity for adaptation.

We recommend an early postoperative MRI to assess the extent of resection and to plan the follow-up. We recommend a close long-term radiological follow-up (Level C)

Evidence

A postoperative MRI may help in evaluating the extent of resection, the decompression of the optic nerves/chiasm and the presence of postoperative complications, such as the presence of intraventricular blood and the ventricular size, to detect the development of early hydrocephalus. This imaging would also detect early ischemic accidents, secondary to vasospasm or intraoperative arterial occlusions.

Remarks:

  • An early MRI performed during the first 48–72 h after surgery represents a baseline study to compare future exams and to evaluate the radiological evolution of postoperative events [70].

  • Patients should be further carefully monitored for recurrence or growth of residual disease. A close radiological follow-up with cerebral MRI every 3 months should be performed in the first year after surgery and then every 6 months for the first 5 years. After this period an annual follow-up is recommended for at least 10 years, as craniopharyngiomas are associated with a high risk of local recurrence [8, 23, 157] and less frequently presenting with postoperative intracranial seeding [161]. Most recurrences occur in the first 5 years after treatment [175, 183, 188, 191] but also delayed recurrences were described [31, 92].

  • The images should be evaluated with an experienced team of neuroradiologists.

Summary

1.1 We recommend that all patients with a suspected craniopharyngioma undergo cerebral MRI to evaluate the extension and radiological features of the tumor and a cerebral CT scan and/or CT angiogram to determine the presence of calcification, bone erosion, hyperostosis, anterior skull base anatomy, and vasculature pertinent to the tumor and thereby to define the appropriate surgical technique and approach (Level C).

1.2 We recommend that patients with craniopharyngiomas undergo a complete history and physical examination including the search for hypopituitarism and hypothalamic compromise. Complementary blood tests should be performed to rule out any endocrinological deficit (Level C).

1.3 We recommend that all patients with craniopharyngioma undergo an initial ophthalmological evaluation with visual acuity, visual fields and optical coherence tomography (Level C).

1.4 We suggest that all patients with craniopharyngiomas undergo a neuropsychological assessment (Expert opinion).

2.1 We recommend the use of transcranial skull base approaches for craniopharyngiomas presenting an extension lateral to the internal carotid artery (Level C).

2.2 We recommend performing transcranial approaches for tumors primarily arising from the floor of the 3rd ventricle with an intraventricular location and not extending to the pituitary stalk or the suprasellar space, where the endonasal route can also be performed (Level C).

2.3 We suggest performing an expanded endonasal transsphenoidal approach as first-line surgical approach for midline and retrochiasmatic craniopharyngiomas without lateral extension (Expert opinion).

2.4 We recommend the use of traditional endonasal transsphenoidal approaches for purely intrasellar craniopharyngiomas (Level C).

2.5 We recommend performing a careful closure with a nasoseptal flap to limit the risk of postoperative CSF leakage when an extended endonasal approach is performed (Level C).

3.1 We recommend performing a GTR when there is no infiltration of the hypothalamus (Level C).

3.2 We recommend performing STR coupled with adjuvant radiotherapy (STR + XRT) when hypothalamic infiltration is confirmed (hypothalamic-sparing resection) (Level C).

4.1 An individualized approach is highly recommended for recurrent tumors and residual tumor after STR + XRT (Level C).

4.2 We suggest applying the similar surgical principles detailed for primary lesions to the treatment of recurrent craniopharyngiomas that require surgical treatment (Expert opinion).

5.1 We recommend a close clinical and endocrinological follow-up in the management of patients with treated craniopharyngiomas (Level C).

5.2 We recommend a close ophthalmological follow-up in the management of patients with treated craniopharyngiomas (Level C).

5.3 We suggest assessing neurocognitive functions in the postoperative period (Expert opinion)

5.4 We recommend an early postoperative MRI to assess the extent of resection and to plan the follow-up. We recommend a close long-term radiological follow-up (Level C).

Conclusion

The initial evaluation of adult patients with craniopharyngiomas should include a clinical, endocrinological, ophthalmological, radiological and neuropsychological assessment. Treatment of these patients should be performed in tertiary referral centers. Based on data from the literature, the endoscopic approach has gained supremacy in the treatment of midline craniopharyngiomas in terms of improved GTR, endocrinological and visual outcomes when compared to standard transcranial approaches. The latter are recommended in cases with lateral extension or with purely intraventricular tumors. Independent of the technique, a safe maximal but hypothalamic-sparing resection should be performed to limit the occurrence of postoperative hypothalamic syndromes and metabolic complications. A close multidisciplinary evaluation is necessary for endocrine, hypothalamic and oncological outcomes to define a long-term treatment plan, tailored to the requirements of each patient. Further clinical studies focused on present-day treatment outcomes of adult-onset craniopharyngiomas would help to better define the optimal management and thereby improve outcomes for these challenging tumors.