Department of Neurological Surgery, University of California, San Francisco 94143
Address correspondence and requests for reprints to: Charles B. Wilson, Department of Neurosurgery, University of California-San Francisco, 533 Parnassus Avenue, Room U125, Box 0350, San Francisco, California 94143-0350.
![]() |
Introduction |
---|
![]() |
The procedure |
---|
I subscribe to the principle of specialized care, and if it makes sense for interested internists to acquire special knowledge and experience in endocrinology, it seems reasonable to apply the same rationale to specialization in pituitary surgery. Particularly in neurosurgery practice makes perfect, and I advocate concentrating, rather than diffusing, surgical referralsnot in the sense of creating "centers", but by encouraging one of several neurosurgeons in larger communities to become the local expert in pituitary surgery. There is no question about it: such a plan for referrals for pituitary surgery can provide improved outcomes for your patients. Has managed care, including capitated care, complicated the referral process? Unquestionably it has, but I hardly need to remind you that inexpert pituitary surgery can be very expensive, in both the short term and the long term. Care of high quality is rarely cost-ineffective, particularly in the care of children and patients with Cushings disease. Endocrinologists have a critical role in the diagnoses, preoperative preparation, and postoperative management of all patients with pituitary adenomas. The short- and long-term followup of every patient treated by surgery is best placed in the hands of an endocrinologist because critical decisions regarding recurrence of endocrine-active adenomas and the subtleties of anterior pituitary insufficiency require continuous monitoring by an expert for optimal outcomes.
![]() |
Pathology |
---|
In Table 1 is shown the distribution of tumor cell types
in my own series of adenomas. Tumors are classified as either
endocrine-active or endocrine-inactive based on their clinical
presentation. For several reasons, the distribution of cell types does
not reflect the true proportion of cell types in the universe of
pituitary adenomas. As this is a series of tumors removed surgically,
it does not include recognized but asymptomatic ("incidental" or
"silent") adenomas, or PRL-secreting adenomas that are managed
satisfactorily by medical means and therefore not referred for surgery.
Persistent or recurrent adenomas operated on after failed initial
surgical procedures are not distinguished in the table, although the
relative proportions of tumor types are no different for the secondary
than for the primary procedures. Most adenomas of mixed
endocrine-active cell types involved PRL and another cell typeeither
adrenocorticotrophin (ACTH), growth hormone (GH), or thyrotrophin
(TSH); and in each instance the tumor was classified according to the
other cell type, rather than as a PRL-secreting adenoma. Tumor type for
the adenomas in the other categories was determined by the presenting
endocrinopathy. Finally, because of an institutional interest in
Cushings disease and a large pediatric endocrinology group at UCSF,
the proportion of ACTH-secreting adenomas in the series, particularly
those in children, is misleadingly high.
|
![]() |
Prolactin-secreting adenomas |
---|
In Table 2 are shown the indications I follow in
recommending surgical removal of a microadenoma, and in Table 3
are shown the indications for removal of a
macroadenoma. These indications for both small and large adenomas
reflect my experience as well as my biases. Based on the basal PRL
value and a high-resolution magnetic resonance MR image I can predict
the likelihood of a surgical cure. If cure is not possible because of
extrasellar spread, almost always on the basis of cavernous sinus
invasion, surgery is not advised unless the patient fails to respond to
medical management. When there is no likelihood of a surgical cure,
surgery has the focused objective of reducing the adenomas mass,
either to reduce the production of PRL to a level that can be further
reduced into a desirable range by a tolerated dose of a dopamine
agonist or to relieve symptoms of compression and reduce the bulk of
the adenoma before irradiation. Because prolactinomas continue to
secrete PRL for several years after being rendered sterile by radiation
therapy, irradiation is a poor and much delayed means of treating the
hyperprolactinemic syndrome.
|
|
Dopamine agonists, such as Parlodel (Sandoz Pharmaceuticals, Hanover, NJ), inhibit tumor-cell replication in the great majority of prolactinomas, and for that reason patients known or proven to have residual or recurrent adenomas should be treated indefinitely with Parlodel. Often a dose as low as 1.25 mg, taken at bedtime, is sufficient to maintain or restore normal PRL levels and prevent tumor regrowth if the residual volume of adenoma is small or if the adenoma is highly dopamine-sensitive. In all cases involving incompletely removed or recurrent PRL-secreting tumors, long-term administration of a dopamine agonist is the first line of treatment. Irradiation is reserved for those few patients who have adenomas refractory to medical therapy and for the larger number of individuals who cannot tolerate the medication.
![]() |
Growth hormone-secreting (somatotrophic) adenomas |
---|
Technically, surgery for acromegaly is often difficult. Because of the greater distance from the patients lip to the sphenoid sinus, a longer speculum is required in many cases. The anesthesiologist must be forewarned of the difficulty often encountered in introducing an endotracheal tube because the enlarged tongue, uvula, and epiglottis of patients with acromegaly and their thickened vocal cords may catch the anesthesiologist unawares, resulting in crisis if pharmacological apnea rapidly produces hypoxia when intubation proves to be difficult.
Many patients with acromegaly have chronic sinusitis, and as a consequence the nasal and sinus mucosa is thickened and difficult to manage. However, the surgeons biggest concern is the cavernous segment of the carotid artery. Typically, the dolichoectatic artery deviates medially and at times may be found, literally, protruding into the sella and compressing the sellar contents. The size and course of both carotid arteries can be determined exactly on MR coronal images, which is a compelling argument for obtaining an MR image even if doing so requires general anesthesiathe exception being a patient with a cardiac pacemaker.
Following selective transsphenoidal adenomectomy, more than 80% of patients with acromegaly have a sustained remission. Incompletely removed adenomas should be treated. Although some patients will do well with medical management using octreotide, most will be treated postoperatively with irradiation.
![]() |
Corticotrophic adenomas (Cushings disease) |
---|
The patient with Cushings disease has friable tissue, soft bone, and capillary fragility. Moreover, in many cases, obesity and other factors conspire to produce venous hypertension, a major complicating factor during the transsphenoidal exploration of a normal-size sella surrounded, literally, by a moat of turgid, confluent dural venous sinuses. Neuroanesthesiologists at UCSF have developed a protocol for coping with elevated venous hypertension, and in the majority of cases the venous pressure can be reduced to normal or near-normal levels. However, these pharmacologic maneuvers are not always effective, and maintaining a lowered venous pressure requires sustained effort. In some cases, the surgeon must deal with persistently elevated venous pressure or, very rarely, abandon the procedure altogether. Even after extensive experience, I am wary, even apprehensive, as I prepare to operate on any patient who has Cushings disease, and this is especially true if the patient is a young child.
In adults with suspected Cushings disease, unless MR imaging indicates a tumor larger than 5 mm in diameter, we proceed to preliminary sampling of the cavernous sinuses. This means that the majority of such patients, more than 75%, undergo venous sampling and concurrent cavernous sinus venography. The tumors of Cushings disease are tiny, the typical adenoma having a diameter of considerably less than 5 mm. Their minute size, coupled with the generally accepted likelihood of a false-positive MR image in 15% of the normal population, argues convincingly for preoperative venous sampling in almost all cases.
Venous sampling is invaluable in surgical decision-making. It helps to determine which patients to operate on, how to conduct the intrasellar exploration, and what to do in the case of a negative exploration. I have several caveats. The cavernous sinuses, rather than the downstream inferior petrosal sinuses, should be sampled bilaterally after the venographic anatomy is defined, and venograms of both cavernous sinuses should be obtained at the same time (2, 3). If venous drainage from the sella is dominantly unilateral, then simultaneously obtained left-right values may not be reliable in indicating lateralization. A cavernous sinus venogram should be obtained because a filling defect in a cavernous sinus may disclose the presence of an intracavernous adenoma (4). A 2:1 or larger cephalic-to-peripheral ACTH gradient establishes the diagnosis of Cushings disease, and in our experience, the additional testing with corticotropin-releasing factor adds nothing to the value of samples taken from the cavernous sinus. If a patient with no anomalous venous drainage patterns exhibits a lateralizing ACTH gradient of 2:1 or greater, then removal of the appropriate half of the anterior pituitary gland will be curative in 80% of cases in which hemihypophysectomy is performed. Pediatric endocrinologists at UCSF are not enthusiastic about venous sampling because children rarely have an ectopic source of ACTH, venous sampling is technically difficult in children, and the risk of incurring morbidity from either local or general anesthesia is not entirely insignificant.
In contrast, I anticipate a difficult exploration whenever I treat a child, and I welcome a guide to the adenomas probable location. Assuming a normal venographic anatomy, I would like to offer the option of hemihypophysectomy in the event of a negative exploration. The complexity of interpreting the results of venous sampling underscores the need for caution in performing hemihypophysectomy under any circumstances. Nonetheless, under the right circumstances, hemihypophysectomy is appropriate and, in most cases, will be successful. My clinical observations support the view that one third, or even slightly less, of a normal anterior pituitary gland provides normal anterior pituitary function.
The adenoma of Cushings disease has distinctive gross characteristics. It is dirty-white and soft, bordering on semiliquid, and it is often freckled by petechial hemorrhage. Because it may be small, the adenoma can go undetected entirely or can suddenly disappear into the tip of a sucker. The adenoma often has a diameter of 2 mm or less. It can reach the surface of the anterior lobe, can be exophytic into the subarachnoid space, can invade or even lie entirely within the cavernous sinus, as it did in one recent case, and, rarely, it can originate in the posterior lobe. Furthermore, in a small number of cases, Cushings disease is the consequence of non-neoplastic hyperplasia. In none of the other endocrine-active tumors is expert knowledge of the gross and microscopic pathology so critical.
The surgical exploration must be meticulous and must include the posterior lobe. If total hypophysectomy is indicated after a negative exploration (that is, in an adult after nonlateralized venous sampling), then the sella must be exenterated, the stalk carefully inspected for involvement of the pars tuberalis, and the pituitary stalk sectioned as close as possible to the anterior lobe to avoid permanent diabetes insipidus. When hemihypophysectomy is indicated after a negative exploration (that is, in a child or adult who has a clearly lateralizing venous gradient and a normal venous anatomy and who has given an informed consent), the stalk must not be injured. Before removing the half of the anterior lobe suspected of harboring the adenoma, I explore it once again and, in addition, explore thoroughly the "normal" half of the anterior lobe that is not to be removed. I have operated on several patients in whom the normal half of the anterior lobe had previously been removed elsewhere based on an assumed infallibility of inferior petrosal venous sampling.
Cushings disease is not a common disease; by one estimate, there are approximately 200 newly recognized cases annually. If left untreated, the disease is fatal. With no other tumor is the surgeons experience more critical in determining outcome, and for this reason I suggest that inexperienced pituitary surgeons must gain experience with less complex tumor types, initially referring patients with Cushings diseaseparticularly pediatric patientsto an experienced colleague. In my view, these rare and difficult cases should be concentrated in the capable hands of experienced pituitary surgeons throughout the United States. Requests for referral to an "out-of-plan" surgeon may be denied because of concern for cost, but a comparison of the long-term costs of a curative operation to those accrued from a failed first attempt creates a compelling argument for referral to an experienced surgeon for economic considerations alone, even without invoking the equally valid issue of appropriate patient care.
![]() |
Endocrine inactive adenomas |
---|
With infrequent exceptions, symptomatic endocrine-inactive adenomas are large, even when headache is the only presenting symptom. Neither suprasellar extension nor focal perforation of the sellar floor, by itself, precludes complete removal. Nearly all symptomatic, endocrine-inactive macroadenomas have extended beyond the confines of an expanded sella turcica, the most common pattern of extrasellar growth being directly upward into the suprasellar space. With infrequent exceptions, the only patterns of growth that preclude total removal are lateral extension into the cavernous sinus or middle fossa and diffuse destruction of the sellar floor with extension of tumor into the sphenoid sinus. Invasion of the dura separating the sella from the cavernous sinus, even without extension into the sinus itself, renders the tumor surgically unresectable unless the invasion is focal, in which case the involved dura can sometimes be removed. A large tumor often displaces the intact dural wall of the cavernous sinus very far laterally without actually invading the dura, and unless coronal MR images display extreme lateral extension, the wall of the sinus may be stretched rather than invaded. The caveat: preoperative radiographic prediction of invasion into the cavernous sinus may be, and often is, incorrect.
For many years, my surgical objective in treating large endocrine-inactive adenomas was decompression of the optic nerves and chiasm with the assumption that irradiation was required afterwards to prevent regrowth. I did not attempt a total removal of large tumors, very possibly as a carryover from the established practice when these tumors were treated by craniotomy. However, as I gained experience and discovered that it was possible to achieve total removal of intrasellar macroadenomas while preserving compressed anterior lobe tissue, I changed the surgical objective from decompression to total removal. In the majority of large adenomas, a clean surgical plane separates the surface of the tumor from the compressed normal structures. By finding and developing this plane early in the process of removal, a surprisingly large proportion of large tumors can be removed completely. Subsequent follow-up of these adenomas with serial MR images at 6-month intervals for the first 2 years and then yearly has shown a 5% rate of recurrence after 5 years. If tumors that invade the cavernous sinus are excluded, I believe that as many as 4050% of endocrine-inactive adenomas can be cured by surgery alone, and for this reason the surgeon should go into the operation with complete removal, rather than decompression, as the goal. Loss of anterior pituitary function that existed before surgery has been an infrequent complication; more often, function that was lost preoperatively later returned spontaneously. To borrow from the lexicon of the Olympics, the bar has been raised for pituitary surgeons.
![]() |
Incidental (coincidental) pituitary adenomas |
---|
The defining features of a presumed incidental pituitary adenoma are
these: (1) it is asymptomatic; (2) there is normal anterior pituitary
function, including a normal GH response to provocative testing; (3)
calcification is absent; and (4) the lesion is confined to the sella
(Table 4). Of the anterior pituitary cell types,
somatotrophs are the most vulnerable to compression, followed by
gonadotrophs as a distant second. Although GH has a legitimate
physiologic role throughout life, GH secretion is essential for the
normal development of children. If GH secretion is normal, all other
anterior pituitary functions are almost certainly normal as well.
Because most craniopharyngiomas are calcified, and because there is a
compelling argument for early intervention, the presence of calcium
should be excluded by computed tomography scanning before the surgeon
assumes that a mixed solid and cystic mass is an incidental
adenoma.
|
What are the indications for surgical intervention? The indications
presented in Table 5 are not absolute, and a decision to
operate that is appropriate for an adolescent may be inappropriate for
an elderly patient. Moreover, an adenoma that has doubled in size
within 1 year, although still asymptomatic, should be viewed
differently from an adenoma that has shown slight but unquestioned
growth (expansion) over the course of 5 years. The listed indications
for operation are only guidelines to be used in conjunction with
clinical judgment. A pure cyst that, by its size, has qualified for
inclusion with incidental adenomas has a likely potential for further
growth; and, in contrast to the common and innocuous small cysts seen
so often in the pars intermedia, the cyst that is 1 cm or larger is, in
my opinion, certain to have further expansion at some unpredictable
rate of growth.
|
![]() |
Closing thoughts |
---|
Received January 8, 1997.
Revised May 6, 1997.
Accepted May 21, 1997.
![]() |
References |
---|