“The most formidable of intracranial tumors”
Harvey Cushing (1932)
The introduction of the transsphenoidal approach to the pituitary fossa by Schloffer in 1907 revolutionized the management of pituitary pathology. Although the employment of this approach was initially slow, the advantages of a direct midline access negating brain retraction enabling rapid and safe decompression of the optic apparatus—was quickly appreciated. Over the ensuing decades many important technical, methodological, and pharmacological advances were introduced—significantly enhancing the scope of this approach. Concurrently, it became apparent that the transsphenoidal approach also had an important role in the management of certain craniopharyngiomas. This role has, of late, been significantly advanced by the introduction and pioneering of extended transsphenoidal approaches and endoscopic techniques, the latter approaches evolving specifically in the process of attempting to achieve gross total removal of craniopharyngiomas-previously considered too hazardous to resect.
This chapter outlines the evolution of the transsphenoidal approach to craniopharyngiomas and describes the rationale, technique, advantages, complications, and variations to the transsphenoidal procedure in the management of these tumors.
Epidemiology and Controversy
Craniopharyngioma definition: A benign, partly cystic epithelial tumor of the sellar region presumably derived from Rathke pouch epithelium. Two clinicopathological forms are distinguished, the adamantinomatous and the papillary craniopharyngioma.
Craniopharyngiomas are uncommon, histologically benign developmental tumors that account for 1.2% to 4% of all intracranial neoplasms. They exhibit a bimodal age distribution with a peak incidence in childhood and another in middle age. Despite craniopharyngiomas being benign lesions, their anatomical location, consistency, potential for adherence to vital brain structures, and propensity for recurrence complicates their management, making complete resection without iatrogenic deficit extremely difficult. Despite their relatively low incidence, there is perhaps no other intracranial tumor that attracts as much controversy as craniopharyngiomas. Their origin, natural history, surgical ressectability, response to radiation, and optimal management have all been topics of intense debate. In 1993, this culminated in a dedicated conference hosted in New York to address these controversies. The subsequent release of a conference consensus statement on management guidelines still left many questions unresolved and the neurosurgical community divided between aggressive complete resection and limited resection followed by radiotherapy.
It is nevertheless universally accepted that the goal of craniopharyngioma surgery—as stated by Van Effentere—should be “complete removal with improved visual function, minimal deterioration of endocrine function and no neuropsychological impairment. ” The frequent anatomical location of craniopharyngiomas in the “deep-seated hypothalamic area paramount for vegetative, emotional, endocrine function and for maintaining worthwhile life” makes this objective challenging.
In determining an individualized management strategy for patients with craniopharyngioma, one has thus to achieve this balance between extent of resection, degree of neuroendocrine deficit, risk of neuropsychological deficit and risk of recurrence. It is also only recently, with the development of complete hormone replacement therapy, that more extensive tumor resection has become possible.
More so than in any other neurosurgical field, personal experience on the part of the surgeon plays a major role in determining the indications for surgery, limitations of surgery, contraindications for surgery and management outcome. It is therefore perhaps misleading for many surgeons to assume that the morbidity and mortality rates for craniopharyngioma surgery of eminent published international experts in quaternary referral centers, will apply to their own practice.
When comparing different series, it is evident that there is considerable variation to the preferred approach and to how complete the surgical resection should be.
As a result of the embryogenesis of these lesions and in approximately 20% to 40% of cases, craniopharyngiomas arise in an anatomical location below the diaphragma sella, for which the transsphenoidal approach is considered the most favorable. Recently, these anatomical and embryological prerequisites for the transsphenoidal approach have to a certain extent been negated by the development of extended and endoscopic transsphenoidal approaches—especially for predominantly cystic suprasellar tumors.
In 1838 Martin Heindrich Rathke (1793-1860) first described the entodermal evagination of the anterior end of the foregut. In 1860 Hubert von Luschka (1820-1875) noted “rests” of squamous epithelium in close relation to the pituitary. In 1900 and 1901 respectively, Joseph Babinski (1857-1932) and Alfred Frohlich (1871-1953) began reporting cases of hitherto undescribed unusual epithelial suprasellar tumors that were occurring without “acromegaly”. In 1899 Mott and Barrett first postulated that they arose from the hypophyseal duct or the “Rathke’s pouch. ” The first accurate histological interpretation and adequate description of these lesions was by Jakob Erdheim (1874-1937) in 1904. He postulated that they arose from embryonic squamous cell rests of an incompletely involuted hypophyseal/pharyngeal duct – a view still largely held to this day, although not completely confirmed.
The first successful removal of a craniopharyngioma was performed transsphenoidally in 1909 in Chicago by William Stewart Halstead (1852-1922). Harvey Cushing’s (1869-1939) first craniopharyngioma operation was in 1912 on a child who subsequently survived for more than 50 years. Over a 20-year period he subsequently operated on 92 patients harboring craniopharyngiomas. Writing in 1932—by which stage he had all but abandoned the transsphenoidal approach, Cushing’s was almost apologetic when he noted that 14 of the 92 craniopharyngiomas had been resected transsphenoidally. He went on to explain that many of these cases had mistakenly been diagnosed preoperatively as pituitary adenomas. These tumors were first called “craniopharyngeal pouch tumors ” or “ craniopharyngeal tasche tumoren ” until Cushing’s coined the term “craniopharyngioma. ” Although the accuracy of the term has been questioned—Rathke’s pouch being an evagination of the primitive stomodeum and not the pharynx, its usage has nevertheless subsequently become universal.
Before the development of steroid therapy, the mortality and morbidity rates for the surgical treatment of craniopharyngiomas were prohibitive. With the advent of steroid therapy in the 1950s, Matson and Crigler were, for the first time, able to remove these lesions safely. The transnasal approach was then repopularized in 1963 by Hardy and Lalonde. Subsequently, many series adopting an endonasal approach have been reported.
Surgical Anatomy and Embryology
Before considering this approach, it is necessary to review both the embryology of the pituitary gland and aspects of the surgical anatomy specific to craniopharyngiomas. It is important to appreciate that craniopharyngiomas can be considered part of a wide spectrum with regard to a number of characteristics. They can vary in size from microscopic to extensive lesions comprising a large percentage of the intracranial volume ( Figure 18-1 ). Almost all have both cystic and solid components in varying combinations. Calcification is found on microscopic examination in approximately half of adult craniopharyngiomas and in almost all pediatric cases ( Figure 18-2 ).
Craniopharyngiomas vary from being well circumscribed to being grossly invasive and can cause an intense glial reaction in the surrounding brain. This reaction is particularly dense around small papillary projections towards the hypothalamus. Some surgeons maintain that traction on this glial attachment will invariably lead to hypothalamic injury and preclude safe total resection of the tumor, while others maintain that this “glial envelope” often provides a safe plane facilitating cleavage without damaging neural tissue.
Craniopharyngiomas are usually adherent to major arteries at the skull base, including small perforators originating from the anterior communicating vessels, the posterior communicating artery, branches from the anterior choroidal artery, and thalamo-perforating vessels. Tumor adhesion to vessels is one of the primary reasons for incomplete tumor resection. Attempted radical dissection has been associated with weakening of the adventitia of blood vessels by injury to the vasa vasorum. This may result in fusiform dilatation of the internal carotid artery and the potential for catastrophic delayed hemorrhage. The previous factors, tumor within the third ventricle and the presence of hydrocephalus, all play an important role in determining extent of resection.
The blood supply of the anterior part of the tumor is supplied by perforators from the anterior communicating artery and proximal anterior cerebral artery. The lateral part of the tumor receives branches from the posterior communicating artery. The intrasellar component is supplied by the intracavernous meningohypophyseal arteries. Of great surgical significance is that craniopharyngiomas never receive blood supply from the posterior cerebral and basilar arteries unless the anterior blood supply from the lower hypothalamus and floor of the third ventricle is absent.
From a histological point of view, craniopharyngiomas can vary from having columnar to cuboidal to respiratory to squamous epithelium. Their cyst contents vary from being clear fluid to viscid then “purulent” with cholesterol crystals, flakes, and calcific and keratinized debris. Their anatomical location may be third ventricular, hypothalamic, infundibular, pars distalis, pars intermedia, neurohypophyseal, or pharyngeal. The direction of growth may be prechiasmatic, suprasellar, intraventricular, or retrochiasmatic and their microanatomical disposition might be either extraarachnoid or subarachnoid and extrapial or intrapial. This wide variation in pathology and behavior once again underscores the fact that management of craniopharyngiomas needs to be very much individualized for each patient and is not suited to the occasional surgeon.
Importantly, the position of the optic chiasm can usually be extrapolated from the position of the anterior communicating artery on preoperative midsagittal MRI. This is critical in determining the best approach. Appreciating the size of the lateral ventricles usually enables one to determine the anatomical relationship of a craniopharyngioma to the third ventricle or its remnants and to plan an appropriate surgical approach to prevent transgressing the hypothalamus.
Rationale for the Transsphenoidal Approach
The decision and rationale regarding the optimal microsurgical approach to a craniopharyngioma should be based on an interpretation of the clinical findings and radiological investigations in conjunction with an understanding of the pathological anatomy and embryology of these lesions. Up until very recently, most patients with craniopharyngiomas were thought to require a transcranial resection of one form or another. There is, now, an expanding proportion of patients in which the transsphenoidal approach represents an alternative if not superior option. When deciding on an optimal surgical approach, eight factors need to be taken into consideration:
Is the sella enlarged and where did the lesion arise?
Is the lesion predominantly cystic or solid?
Is there a major calcified component?
Can the position of the chiasm be determined?
Is hypopituitarism present?
How large is the tumor?
What degree of lateral extension is there?
Has there been prior surgical therapy or irradiation?
What is the relationship of the third ventricle and hypothalamus to the tumor?
The prevailing theory is that craniopharyngiomas arise from lingering ectoblastic cell rests or areas of squamous metaplasia related to the embryogenesis of the pituitary. They can therefore occur anywhere along the path of the incompletely involuted hypophyseal-pharyngeal duct. The migrational axis of this structure extends from the tuber cinereum rostrally to the sphenoid bone and pharynx caudally with the infundibulum, the pars intermedia, and the pars tuberalis of the hypophysis interposed. Anywhere along this tract, ectoblastic cell rests may persist, giving rise to craniopharyngiomas.
When craniopharyngiomas arise above the diaphragma, the sella turcica will appear normal with regard to its size and contents ( Figure 18-3 ). Approximately one third of craniopharyngiomas arise from below the diaphragma sella. When they arise here, they produce progressive enlargement of the sella—a radiological feature seen in 30% to 60% of previously reported series and presumptive evidence of an intrasellar origin. This anatomical origin has important implications regarding the optimal surgical strategy and is by far the most important consideration when selecting an approach. As the tumor enlarges the diaphragma stretches over it. This occurs even with considerable suprasellar expansion. The advantage of this pattern of growth is that although the dorsal tumor capsule may virtually fuse with the diaphragma, the craniopharyngioma does not extend through it. The diaphragma therefore remains an effective barrier against pial invasion. This prevents the craniopharyngioma arising in the sella from having intimate attachments to rostral intracranial contents above the diaphragma—in particular the optic chiasm, basal vasculature, and hypothalamus. It also prevents mutual vascularization of the tumor and the optic apparatus. Accordingly, tumors of intrasellar origin remain almost invariably extraarachnoid and extrapial in disposition. These features of the pathologic anatomy of infradiaphragmatic craniopharyngiomas theoretically permit complete resection transsphenoidally if the surgeon is prepared to resect the diaphragma and risk the potential of a postoperative cerebrospinal fluid (CSF) fistula and meningitis ( Figure 18-4 ).
It is nevertheless also critical to be aware that the same principles do not necessarily apply to the lateral capsule of the intrasellar component of a craniopharyngioma. This may be densely adherent to the medial wall of the cavernous sinus, which may itself also be deficient—adversely affecting resectability. Dissection between the lateral wall of the tumor capsule and the cavernous sinus runs the risk of precipitating hemorrhage from the sinus, injuring the cavernous internal carotid artery, false aneurysm formation, or traumatizing cranial nerves within the sinus.
An expanded sella is also extremely advantageous in allowing the surgeon sufficient room and maneuverability to effectively manipulate transsphenoidal instrumentation during a resection. Nowadays, however, sella enlargement should not be regarded as an absolute prerequisite for transsphenoidal surgery. Until only very recently and before the advent of extended transsphenoidal approaches, it was considered at best challenging and at worst inadvisable, to proceed in the presence of a normal sized sella. From important craniopharyngioma surgical series reported over the last 20 years, the use of the transsphenoidal approach has varied between 10% to 79%. To a certain extent, this variation is explained by individual surgical preferences, experience and training – particularly with respect to extended transsphenoidal approaches.
The surgical technique is similar to that used for the transsphenoidal management of pituitary adenomas. We use frameless stereotaxy routinely finding that it has a negligible effect on operating time while conferring considerable benefit with respect to intraoperative navigation—particularly with redo procedures in which anatomical landmarks may be absent or distorted. This may be used in combination with fluoroscopy. For patients with large tumors and for tumors with suprasellar extension, a lumbar catheter is inserted preoperatively. This serves two purposes: It allows the introduction of gas or saline intrathecally to occasionally assist in reduction of the suprasellar extension/component of the tumor while the diaphragma is still intact. It also permits drainage of CSF postoperatively to assist the creation of a watertight seal following the repair of often large CSF fistulas created intraoperatively. An area over the left abdominal flank is therefore prepared to allow harvesting of a fat graft to assist in the skull base repair and to seal off any fistula into the subarachnoid space.
While the original sublabial midline rhinoseptal approach pioneered by Cushing’s has of late been largely abandoned in favor of endonasal approaches, it remains advantageous for transsphenoidal craniopharyngioma resection because it confers a wider exposure than endonasal approaches. The sella floor is then carefully and widely exposed through the transsphenoidal approach to gain maximal exposure of the sella region and the greatest possible working area before being opened ( Figure 18-5 and 18-6 ). This is a critical factor for the success of the procedure. We prefer making an aperture in the sellar floor using the small diamond drill bit on the Anspach Micro© drill or the curved TAC attachment for the Midas Rex© drill. The bayonet dural elevator/dissector is then introduced to create a plane between the sella floor and the dura before using the 2-mm Kerrison punch to complete the removal of the sellar floor. The lateral extent of the dural exposure should be from one cavernous sinus to the other. The vertical exposure should extend from the junction of the clivus with the sellar floor inferiorly to the dural reflection at the tuberculum superiorly. The dura should then be opened with caution, initially developing a subdural plane to prevent injury to the capsule of the tumor. Most craniopharyngiomas that expand the sella compress the normal anterior pituitary gland so that hypopituitarism is common. If normal pituitary tissue is present, it is usually just behind the dura, anteroinferiorly presenting as an attenuated membrane. It needs to be displaced laterally or incised via a vertical midline incision and carefully separated in order to be preserved. The dura is then opened either using an inverted “U”-shaped flap incision based superiorly with relieving incisions as required or a stellate incision. The dural edges can then be cauterized using bipolar diathermy causing them to retract back even further thereby maximizing exposure. This may risk precipitating further extradural bleeding. Once the normal pituitary—if present, has been identified, an extracapsular plane of cleavage between the tumor and the pituitary and the tumor and the sella boundaries is then developed. We find it advantageous at this stage to have the benefit of being able to use the microsucker for both aspiration and as a retractor in combination with a malleable ball end microdissector to develop the plane around the tumor. The dissection required to resect a craniopharyngioma is considerably more complex than that required for reducing a pituitary adenoma—particularly if faced with adherence and/or bleeding. We find bimanual instrumentation under an operating microscope an advantage over endoscopy despite the latter being our standard approach for pituitary pathology. This can be either augmented with endoscopic assistance or be undertaken as a totally endoscopic approach.
Care should be taken at this stage with attachment/invasion into vascular structures within the cavernous sinuses. Gentle dissection should take place under direct vision with the aid of the operating microscope or endoscope. If present, the cystic component of the lesion is aspirated and tissue sent for frozen section. It is helpful to try and keep the capsule in continuity, as it gives the surgeon a handle to apply traction to the remaining tumor helping secure total removal. Having dissected and mobilized the anterior, inferior, and lateral aspects of the tumor capsule; having drained any cystic component; and having debulked the intrasellar component; one might be fortunate in being able to reduce the suprasellar component of the tumor into the sella. With some craniopharyngiomas, various techniques or combinations thereof can be employed to facilitate this, namely intrathecal infusion of saline or nitrous oxide, jugular venous compression, or positive end expiratory pressure ventilation. The advantage of using gas is that one is able to obtain an intraoperative pneumoencephalogram outlining the descent of the diaphragma and extent of one’s resection. Attention is then directed towards the dorsal component of the tumor capsule. Inspection usually reveals that it is fused to the diaphragma and one is very rarely able to separate the two. To complete the resection, the diaphragma needs to be detached and resected together with the capsule. This is usually achieved with a BD Beaver® sickle blade on a long bayonet mount. The incision into the diaphragma is initiated at the anterolateral margin opening up into the subarachnoid space ( Figure 18-5, D ) and continued medially and inferiorly. This mobilizes the anterior attachment of the tumor capsule which is fused with the diaphragma from the tuberculum. By gently depressing the tumor, one is able to visualize the pituitary stalk ( Figure 18-5, E ). A combination of further careful bipolar diathermy and predominantly sharp dissection detaches the tumor from the stalk with minimal trauma. Every attempt should be made to preserve the stalk without excessive manipulation. Limiting manipulation to sharp dissection prevents damage to the stalk and secondary traction injury to the hypothalamus. Any further intracranial dissection required can be performed under direct microscopic or endoscopic control. If the tumor is densely adherent to the stalk and cannot be removed without excessive manipulation—then the stalk should be transected as low as possible, often enabling preservation of neurohypophyseal function. The membrane of Liliequist is always initially intact when removing a craniopharyngioma and provides a useful protective barrier between the tumor, the basilar artery, and the brainstem.
At this stage a complete resection should have been achieved. Refractory bleeding can usually be controlled with the aid of Surgicel Fibrillar and/or FloSeal hemostatic sealant, particularly if the bleeding point is not easily visible. Attention is now directed toward a sella floor reconstruction to seal the CSF fistula created.
Extended Transsphenoidal Surgery
The transsphenoidal approach was initially advocated primarily for lesions arising in the sella, with or without suprasellar extension, with the resection technique predicated on an intracapsular removal. Not infrequently, total tumor resection is prevented by invasion into adjacent structures, fibrous consistency or the adherence of the tumor to structures such as the hypothalamus, the optic nerves, and chiasm. Maneuvers designed to increase intracranial pressure in order to reduce suprasellar lesions into the operative field are thus not universally effective.
Early experience with extended transsphenoidal surgical approaches evolved in the process of attempting to achieve gross total removal of craniopharyngiomas. The simultaneous development of modern microinstrumentation, microsurgery air drills, frameless stereotaxy, the operating microscope, precision ultrasonic aspirators, and the endoscope has extended the role of this approach still further.
During the past 2 decades, largely inspired by the work of minimally invasive endoscopic surgeons such as Jho and Cappabianca, and pioneered by pituitary surgeons such as Oldfield, Weiss, Couldwell, and Laws, numerous other innovative methods of exposing the anterior fossa cranial base have been described. In this respect, the transsphenoidal approach is “extended” to gain access to lesions of the parasellar and clival regions. In the hands of experienced transsphenoidal surgeons, craniopharyngiomas arising above the diaphragma, of predominantly solid composition and even with significant calcification, are now being successfully resected transsphenoidally, providing an alternative to craniotomy.
This approach has numerous advantages, although it often it requires the expertise of an experienced transsphenoidal surgeon. Its main advantage is that midline exposure is achieved without requiring any brain retraction, enabling one to expose the tumor face-on ( Figure 18-6 ). The exposure provided, albeit via a narrower corridor, is often wider than the exposure provided via pterional and frontal craniotomy approaches using the surgical corridors in the region of the optic chiasm, nerves, and internal carotid arteries ( Figures 18-12, A, B and C ). It has the added advantage of facilitating a dissection along the axis of the tumor and avoids dissecting across the long axis of the optic nerves. It does not require an enlarged sella and is not restricted by the position of the chiasm in the same way as a transcranial approach. See Table 18-1 for a review of the literature.
|Publication||Year||Number of Cases||Adults (%)||Preferred Approach||Total Removal %||Early Mortality %||Follow-up %||Recurrence %|
|Baskin and Wilson||1986||74||62||Subfrontal>transsphenoidal||9.5||3||48||14|
|Yasargil et al||1990||144||51||Pterional||90||9||NA||7|
|Symon et al||1991||50||80||Transtemporal||60||4||30||10|
|Maira et al||1995||57||88||Transsphenoidal>pterional||75||0||78||0|
|Fahlbusch et al||1999||168||80||Pterional>transsphenoidal>bifrontal||49 prim||0.7 prim, 10 rec||65||11|
|Van Effenterre and Boch||2002||122||76||Pterional||59||2.5||84||13|
|Chakrabarti et al||2005||86||NA||Transsphenoidal||84||0||>60||7|
|Di Rocci et al||2006||54||0||Pterional||78||3.7||104||7|
|Shi et al||2006||284||80||Pterional>bifrontal||84||4.2||25||14|
|Zuccaro||2005||153||15 day-21 yr||69||NA||192||0|
|Gardner et al||2008||16||100||Endoscopic endonasal||50||0||34||31|
|Shi et al||2008||309||83.8||Pterional 68.3%||89.3||3.9||25||13.7|