The Journal of Clinical Endocrinology & Metabolism Vol. 82, No. 8 2381-2385
Copyright © 1997 by The Endocrine Society
Surgical Management of Pituitary Tumors
Charles B. Wilson
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
|
|---|
Having performed 3000 transsphenoidal
procedures during the past 27 years, almost all of them for pituitary
adenomas, I have been asked to give a surgical perspective on this
small but important aspect of endocrine practice. In preparing this
retrospective, I was freed from the customary journalistic format to
present a personal perspective on surgical management that may prove
helpful to endocrinologists in making decisions about the treatment of
patients for whom surgery is a consideration.
|
The procedure
|
|---|
Transsphenoidal surgery has become an operation with remarkably
little morbidity and exceptionally low mortality rates. Hospital stays
of 2 days are standard, and selectively patients are being discharged
even earlier. Patients whose jobs are not physically demanding can
usually return to work within 1 to 2 weeks after surgery. To be sure,
complications, both minor and major, occur even in the most experienced
hands. As one example postoperative loss of pituitary function, whether
categorized as a side effect or a complication, is a serious concern in
a patient of any age, but most critically in children and young adults.
Considering the benign behavior of most pituitary adenomas, the
preservation and possible improvement of anterior pituitary function
assumes a priority equal to that of avoiding injury to critical
parasellar structures.
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 referrals—not 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 Cushing’s 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
|
|---|
I am an advocate of immunostaining. At the University of
California, San Francisco (UCSF) we do immunostaining routinely on all
pituitary adenomas, even though it provides critically important
information in only a minority of cases. Without immunostaining, for
example, nodular corticotropic hyperplasia might go undetected; and in
the case of a large nonsecreting adenoma associated with
hyperprolactinemia as a nonspecific effect of the compression and
distortion of dopaminergic vascular pathways, the adenoma might be
mistaken for a prolactin (PRL)-secreting adenoma and could be managed
incorrectly. Only immunostaining differentiates an endocrine-inactive
adenoma from a prolactinoma, each of which is managed quite
differently.
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 type—either
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
Cushing’s 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.
Categorized under endocrine-inactive adenomas are those adenomas that
were recognized by their clinical presentation as a mass compressing
either the anterior pituitary gland or extrasellar structures,
presentation only with headache being one characteristic of the latter
group. Although immunostaining revealed that many tumors contained one
or several cell types, as determined by the presence of secretory
granules or their appropriate ultrastructural characteristics, in none
was a secretory syndrome associated with the mass effect. In contrast
to immunostaining, in my experience routine 
-subunit determination
has been expensive and valueless in practice.
|
Prolactin-secreting adenomas
|
|---|
The prolactinoma is the only pituitary adenoma for which medical
management in the long term is fully satisfactory, and for that reason
the proportion of patients with prolactinomas referred for surgical
consultation varies widely in different geographic regions. I assume
that the proportion of patients referred for surgical consideration in
Northern California represents an approximation of general practice.
The referred group in my practice includes patients who have
unacceptable side effects caused by medication, patients with
dopamine-insensitive adenomas, and those patients who, after becoming
informed, for personal reasons select surgical over medical management.
In a significant proportion of patients referred to me for surgical
consultation, I have advised medical over surgical management based on
gender, age, and the probability of a surgical cure—"cure" in the
sense of long-term freedom from recurrence.
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 adenoma’s 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.
Unless a PRL-secreting adenoma is cystic, partly necrotic, or contains
a significant collection of blood, the levels of PRL in the blood
accurately reflect the tumor’s size. Furthermore, the preoperative
basal PRL value provides an accurate prediction of outcome, the
probability of surgical cure having a directly inverse relationship to
basal PRL value. An even more accurate prediction of outcome is
provided by the immediately postoperative PRL value, obtained 1–2 days
after surgery: an unmeasurable (<2 µg/L) value predicts a cure with
more than 90% probability, and higher values within the normal range
are inversely related to probability of cure (1). Postoperative values
even slightly above the upper limit of the normal range indicate
incomplete removal of the adenoma.
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
|
|---|
For almost three decades, the preferred primary treatment for the
patient with acromegaly has been surgery. In occasional cases of
acromegaly, uncontrolled diabetes, hypertension, or congestive heart
failure may counsel against anesthesia, and in such cases preoperative
medical preparation, including somatostatin, may be advisable. Still,
today (with the exception of such cases) the initial and usually
definitive treatment for acromegaly is transsphenoidal surgery. As a
general rule, younger patients have larger tumors and higher basal
growth hormone (GH) values, whereas older patients with acromegaly are
more likely to have smaller tumors and lower, or even normal, random GH
values.
Technically, surgery for acromegaly is often difficult. Because of the
greater distance from the patient’s 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 surgeon’s 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 anesthesia—the 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 (Cushing’s disease)
|
|---|
In the short term, Cushing’s disease is the most serious and
life-endangering condition caused by any pituitary adenoma; for the
surgeon, these tumors present the most difficult challenge of all
pituitary adenomas. As I begin each operation for Cushing’s disease, I
have to assume that the operation will be difficult from beginning to
end, and seldom am I pleasantly surprised by a truly "simple"
case.
The patient with Cushing’s 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 Cushing’s disease, and this is
especially true if the patient is a young child.
In adults with suspected Cushing’s 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 Cushing’s 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 Cushing’s 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 adenoma’s 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 Cushing’s 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, Cushing’s 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.
Cushing’s 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 surgeon’s 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 Cushing’s
disease—particularly pediatric patients—to 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
|
|---|
In this category are adenomas that produce no clinically
recognizable secretory product. At the time of diagnosis, the great
majority of cases present with impaired vision and some expression of
hypopituitarism, typically gonadotropic insufficiency, with or without
associated headache. Because of the increasingly prevalent use of
high-resolution MR imaging, a new category of incidental and
asymptomatic—presumed—pituitary adenomas requires the formulation of
a new management algorithm, which I will considered separately.
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 40–50% 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
|
|---|
Although I have kept no record of the number of patients with
presumed incidental pituitary adenomas whom I have evaluated since the
introduction of high-resolution MR images, I estimate that they
constitute one out of five patients referred to me with the diagnosis
of pituitary adenoma. Radiographically, some presumed adenomas are pure
cysts rather than cystic adenomas, but if a cyst has attained a
diameter of 1 cm, it should be managed in the same manner as an adenoma
because both the pure cyst of this size and an adenoma, whether smaller
or larger than 1 cm, have potential for further growth.
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.
Any incidental adenoma or cyst meeting the criteria just defined should
be left alone and the patient advised to have follow-up examinations by
a physician who accepts responsibility for periodic clinical,
laboratory, and radiographic assessments based on the patient’s age,
the size of the adenoma, and other concerns. As a rule, younger
patients require more frequent evaluation, and in that regard opinions
vary as to frequency in each particular case.
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
|
|---|
I have acquired a deep respect for pituitary adenomas. Although
rarely fatal because of their unrestrained growth, they can have a
profound effect on the quality of life, and in many cases lead to
secondary complications caused by endocrine dysfunction that affect not
only quality but also length of life. Quite apart from cosmetic
considerations, patients with active acromegaly are unhealthy, and
their life span is shortened significantly. Unless treated successfully
by either medical or surgical means, prolactinomas have consequences
that are profound for physical and psychological health as well as
fertility. The patient with Cushing’s disease becomes disabled rapidly
by a range of cortisol-produced complications, and a shortened life
expectancy is a certain consequence of uncontrolled hypercortisolism.
Endocrine-inactive adenomas threaten vision and pituitary function and,
even though they can be cured by irradiation (particularly when the
volume of tumor is small), irradiation of the normal anterior pituitary
gland (and hypothalamus) can and often does produce hypopituitarism.
Technologic advances during the past 2 decades have produced more
effective therapeutic tools than were available at any time in the
past, but even so, optimal management of any pituitary adenoma requires
thoroughness in the evaluation process, the thoughtful and expert use
of medical and surgical interventions, and in many cases a bit of good
luck!
Received January 8, 1997.
Revised May 6, 1997.
Accepted May 21, 1997.
|
References
|
|---|
-
Tyrrell JB, Lamborn KR, Hannegan LT, Applebury CB,
Wilson CB. 1996 Transsphenoidal microsurgical therapy of
prolactinomas: initial outcome and long-term results. Ann Intern Med.
In press.
-
Fraser KW, Halbach VV, Smith TP, et al. 1996 Bilateral simultaneous cavernous sinus sampling for adrenocorticotropin
hormone. AJNR. In press.
-
Mamelak AN, Dowd CF, Tyrrell JB, McDonald JF, Wilson
CB. 1996 Venous angiography is needed to interpret inferior
petrosal sinus and cavernous sinus sampling data for lateralizing
adrenocorticotrophin-secreting adenomas. J Clin Endocrinol Metab. 81:475–481.[Abstract]
-
Wilson CB, Mindermann T, Tyrrell JB. 1995 Extrasellar, intracavernous sinus adrenocorticotropin-releasing adenoma
causing Cushing’s disease. J Clin Endocrinol Metab. 80:1774–1777.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
R. J. Weil, A. O. Vortmeyer, L. K. Nieman, H. L. DeVroom, J. Wanebo, and E. H. Oldfield
Surgical Remission of Pituitary Adenomas Confined to the Neurohypophysis in Cushing's Disease
J. Clin. Endocrinol. Metab.,
July 1, 2006;
91(7):
2656 - 2664.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. Forti and C. Krausz
Evaluation and Treatment of the Infertile Couple
J. Clin. Endocrinol. Metab.,
December 1, 1998;
83(12):
4177 - 4188.
[Full Text]
|
|
|
Top
Introduction
The procedure