Skull Base Tumors
Introduction to Skull Base Tumors
The glomus tumor has been the focal point, owing to its precarious location, around which the field of neurotologic skull base surgery has evolved. The past 50 years have seen the development of microsurgical techniques, advances in anesthesia and reconstructive capabilities to a degree which currently allow the heretofore unresectable to be reliably considered resectable. Surgery has emerged as the undisputed definitive management preference for glomus tumors. This innovative surgery has been predominantly stimulated by developments in our ability to diagnose and delineate type and extent of disease radiologically. Consistently outpaced by diagnostic progress in the past, surgery has finally evolved to a point of compatibility with our ability to image these tumors. Treatment planning is fortified by contemporary diagnostic capability to describe the disease as well as to predetermine treatment 'pitfalls.' It is this concept of informed pretreatment investigation which will be discussed in this section.
Differential Diagnosis
When the clinical circumstances suggest the diagnosis of lateral skull base disease, in addition to craniocervical paraganglioma, the neurotologic skull base surgeon must produce a differential diagnosis which could foil his treatment plan for glomus tumors. In addition to glomus tumors, the most common lesions are similarly "benign" lesions such as neuroma or meningioma. Rarer lesions must be expected and their possibility investigated as their management may differ significantly. This is particularly true of carcinoma. A pragmatic differential diagnosis list for glomus tumors includes: glomus tumor, meningioma, neural lesions (neurilemmoma, neurofibroma, and chordoma), carcinoma (primary, metastatic, and nasopharyngeal), primary cholesteatoma, and rare lesions (rhabdomyosarcoma, plasmacytoma, melanoma, giant cell tumor, osteoblastoma, lipoma, chondrosarcoma, petrositis, and histiocystosis). A glomus tumor confined only to the middle ear is a lomus tympanicum. One involving the jugular bulb and cranial base is a glomus jugulare.
Tumor Classification
In an effort to communicate, compare results, adjudicate outcome and plan treatment, updated schemes to classify this "new breed" of glomus tumor were needed. Fisch led the way in 1978. The Otology Group functionally upgraded in 1982. Both systems are erratically applied. No uniform nomenclature exists for this purpose. Table I presents the Fisch "A, B, C, D" system while Table II exhibits the Glasscock-Jackson method.
Associated Lesions
Lesions occurring in association with glomus tumors are important to preoperatively identify to the extent that they may complicate the treatment plan for the primary lesion. The diagnostic inquiry must be developed to incorporate the search for these occurrences. Glomus tumors exhibit a tendency toward multiplicity estimated to occur with an incidence of approximately 10% in the nonfamilial situation, while an incidence of 25% to 50% of synchronicity is suggested in familial situations. Glomus tumors can occur in any permutation arising from any of the craniocervical paraganglia stations. Glomus tumors may also arise below the clavicle, but are rare in association with craniocervical lesions. For familial disease, the pattern of inheritance is not well established. The chief cells of paraganglia are one of 40 distinct cell types which comprise the diffuse neuroendocrine system (DNES) and, as neural-crest derivatives, are grouped with cells of similar derivation (adrenal medulla, thyroid "C" cells, etc.). In this regard, glomus tumors have been associated with thyroid carcinoma, visceral neoplasms, parathyroid adenoma, pheochromocytoma, other neurogenic disorders and multiple endocrine neoplasia (MEN) syndromes. Spector and Sobol identified second primary malignancy in 7% of 95 glomus tumor patients.
Glomus tumors possess the histochemical machinery to produce catecholamines and, in so doing, mimic pheochromocytoma from which it must be differentiated. "Functional" glomus tumors occur in 1% to 3%. Serum elevation of catecholamines 3 to 5 times normal is required for symptoms to be elaborated. Glomus tumors have been associated with 5-hydroxyindoleacetic acid (5-HIAA) and carcinoid syndromes.6 Consistent with their roles as interneurons, paraganglia and their tumors may secrete an array of neuropeptides which may be chemically significant. Cholecystokinin (CCK) serum levels have been shown to be elevated in glomus tumors. A role in postoperative gastrointestinal effects may be related. Clinical consequences of these observations are apparent. Multiplicity of lesions may complicate surgical planning and must be prospectively identified. Familial tumor patterns are particularly suspect. Synchronous lesions must be differentiated from metastatic disease, which is usually identified by occurring in locations other than those from which paragangliomas usually arise. Anticipation of a secretory potential permits the identification of endocrine syndromes so that their efficient perioperative management limits anesthetic morbidity. Pheochomocytoma must be differentiated. Associated nonglomus tumors are not routinely sought unless clinical indicators suggest synchronous thyroid, central nervous system (CNS), parathyroid, or visceral disease.
The diagnostic protocol must take the concept of associated lesions into account.
Diagnosis
The ultimate treatment of the glomus tumor depends on tumor type and extent. Each lesion is considered unique.
The clinical features of the disease serve to alert the neurotologic skull base surgeon to the involvement of the ear and lateral skull base. The glomus tumor patient usually complains of pulsatile tinnitus and hearing loss. Tumor growth into the ear presents a conductive loss, the sensorineural component being determined by the degree of labyrinthine involvement. Cranial neuropathy suggests a more extensive process. Voice changes, dysphagia, aspiration, shoulder weakness, or tongue deformity point toward the jugular and/or hypoglossal foremen. "Idiopathic" cranial neuropathy should be considered unacceptable, and an extensive evaluation of the lateral skull base planned. A significant incidence of our glomus tumor population, in retrospect, originated with a diagnosis of "idiopathic" vocal cord or twelfth nerve palsies. Facial paralysis is an ominous sign prognostically, but should not confuse the differential diagnosis. Palpitations, labile blood pressure, flushing, cephalgia, etc. must be sought as indicators of catecholamine secretion.
The presence of a hypotympanic vascular mass is characteristic on physical exam, but not uniform. Visible margins 360 degrees around the tumor permit the diagnosis of a tympanicum-type tumor. Without the security of circumferential visualization, the differentiation of tympanic from jugulare lesion is not possible by examination alone. A jugulare lesion should be expected until proven otherwise. The absence of a hypotympanic mass does not exclude glomus tumor from the diagnosis, but should heighten differential curiosity. Epitympanic glomus tumors are rare but do occur. Myringotomy for biopsy is mentioned only to condemn it. Biopsy of an aberrant internal carotid artery (ICA) is not only dramatic, but potentially catastrophic. Bleeding can be controlled only by packing, a maneuver which risks damage to hidden regional anatomy. Furthermore, myringotomy unnecessarily extends the limits of definitive resection. When tissue diagnosis is unavoidably necessary, a postauricular transmastoid approach in which all vital anatomy is exposed and in which bleeding is readily controllable is indicated. Even in the frailest of patients, this can be done under local anesthesia.
Data for any neurotologic skull base surgical plan must include: 1. determination of tumor type, size, and extent; 2. evaluation for synchronous lesions or histochemical associated phenomenon; 3. assessment of major vascular involvement; 4. assessment of intracranial collateral circulation; 5. assessment for/of intracranial extension.
Most objectives are attainable by the identification of a soft-tissue mass and/or its associated bone destruction. Initial-phase assessment seeks to differentiate tympanicum from jugulare lesions. Disease extent is then assessed. It should be readily apparent that none of these data essential to surgical planning are safely derivable from the clinical evaluation, i.e., the patient history and physical examination. The diagnostic mainstay of glomus tumor diagnosis is medical imaging.
The identification of either air and/or bone between a tympanic cavity mass and the jugular bulb characterizes the mass as a tympanicum tumor. Computed tomography (CT) of the temporal bone is the best imaging format for this purpose. CT contrast enhancement and bone detail best allow the identification of small middle ear masses and their relationship to the dome of the jugular bulb. Bone and softtissue windows in the axial and coronal planes are used. The role of CT in differential diagnosis is useful. The use of CT in large-lesion assessment is in the determination of the degree of bony destruction as it relates to the vital anatomy of the temporal bone. Tumor extent, intracranial extension (ICE), and vascular involvement are, at present, better evaluated by MRI.
Magnetic resonance imaging provides unsurpassed data regarding extent of tumor and tumor relationship to both neural and vascular regional anatomy. Furthermore, MRI offers the best data differentiating glomus tumors from other possibilities. It is used preferentially in identifying synchronous lesions as well as ICE. The quantitative (differential diagnostic) capabilities of MRI remain, as yet, untapped. MRI angiography may well supplant traditional angiography.
Traditional, bilateral carotid angiography is still employed to specifically determine ICA tumor involvement and, more importantly, collateral blood flow to the brain. The predictability of successful ICA interruption is the most controversial topic in skull base surgery. Our position is that the success of ICA interruption is not uniformly predictable preoperatively. The difference in hemodynamic circumstances surrounding tests occluding the awake patient in the radiology suite and that attending the anesthetized intraoperative patient when ICA interruption actually occurs is central to this assessment.
Owing to its unpredictable outcome, ICA sacrifice is not taken lightly. Whenever the ICA is sacrificed a reconstruction strategy is adapted.
All glomus tumor patients undergo preoperative catecholamine screening. Significant elevations (three times normal and higher) require pheochromocytoma differentiation from endogenous glomus production by standard urologic protocols.
Treatment Options
Treatment is regarded as palliative or definitive, and is based on data generated by the diagnostic protocol. Emphasizing individuality, a therapeutic plan is tailored to each patient's unique circumstances. The modern definitive management of glomus tumors is surgical. No lesion is "unresectable" by traditional concepts. The issue to be reconciled is: in the natural course of the patient's remaining years is this tumor likely to cause serious morbidity or mortality?
In general, the elderly are not candidates for neurotologic skull base surgery. "Elderly" is defined physiologically, but hovers around age 65. For these slow-growing lesions, it is unlikely that a small, asymptomatic lesion in a 72-year-old would cause any concern in the remaining years. "Slow tumor growth rate" logic, however, is not applicable to the 30-yearold woman customarily encountered. Although surgical morbidity for the elderly likely exceeds potential surgical benefit, as well as tumor risk itself, this cannot be said for the situation in a 30-year-old. Surgery is the only option that offers the opportunity for definitive cure.
In the asymptomatic patient who is not a surgical candidate, the lesion is intelligently followed by serial imaging only. The symptomatic patient is irradiated, since this modality is not regarded as benign. In combination with radiation therapy, subtotal resection to control adverse local tumor effects or if neurosurgically indicated, may enhance palliative improvement of quality of life.
Simultaneous, bilateral skull base lesions are complicated. Curative surgery risks laryngeal denervation as well as pharyngeal deafferentiation. Although laryngology is better prepared to deal with bilateral denervation than ever before, its occurrence in the face of sensory and coordinated swallowing deficits mandates serious assaults on quality of life by tracheostomy, permanent tracheal diversion, and/or alternate means of alimentary support. In this circumstance, definitive management of the most imminent life (style)-threatening lesion is entertained, while the subordinate lesion is palliated when outcome is clear. The quality of life should not be disregarded. Often, no "right" answer is possible.
As a generalization, combination forms of therapy have been forsaken. Radiation therapy complicates surgery, either from the standpoint of wound healing and/or neural regeneration, and is avoided preoperatively or postoperatively. Its advantages are unclear. On the other hand, the surgical advantages of embolization for the experienced neurotologic skull base surgeon have been proven.
Internal Cartoid Artery
Every glomus tumor, in variable degrees, will relate itself to the ICA, the fundamental key to successful lateral cranial base surgery. The rate-limiting step of tumor resection is its dissection from the ICA. The basic principles of vascular surgery, proximal and distal control, must be rigidly applied. Proximal control in the neck is familiar. Exposure must warrant distal control in the tympanic, petrous, or intracranial segments. Purposeful control exceeds mere visualization. The ICA should be dissected over 180 degrees of its circumference and mobilized to allow instrumentation in the event of planned or other losses of ICA integrity. Dissection of tumor from the ICA occurs in a subadventitial plan and underscores the desirability of exquisite control. Extrication of the ICA from the tumor is most always possible, except in extreme cases.
Guidelines for ICA sacrifice are not secure and when indicated, is always reconstructed. The involved risks, always individualized, seem validated by the extreme circumstances in which this option is entertained. ICA manipulation risks ICA spasm: our management preference is topical or injected pharmacologic vasodilation. Irreversible ICA spasm is one of few indications for abortion of the procedure.
