The pituitary is an important endocrine gland located at the base of the brain. It is involved in nearly all processes of homeostasis, as well as growth and development. This gland commonly develops pathology, which may result in a mass effect on adjacent intracranial structures or be hormonally active. Pituitary adenomas are present in 14% of autopsy studies and 23% of radiology studies. Given the high prevalence, either due to symptoms from an underlying disease process or through incidental discovery, it is crucial to understand the methods for imaging pituitary lesions. This article will discuss the various imaging modalities for the pituitary gland, their indications, and their limitations.
The pituitary gland is connected to the hypothalamus at the median eminence via the pituitary stalk. It sits within the bony cavity of the sella turcica, a superior depression within the sphenoid bone, named after its resemblance to a Turkish saddle. Laterally it is bordered by the cavernous sinus and internal carotid arteries. Superiorly, its border is the optic chiasm. This anatomic positioning is essential, as it has direct implications on different symptoms patients may develop due to mass lesions. For example, the superior extension of a pituitary mass may cause compression of the optic chiasm, resulting in visual field deficits, classically bitemporal hemianopsia. If the pituitary expands inferiorly, it may erode through the sella and into the sphenoid sinus, resulting in a cerebrospinal fluid leak, which typically presents as clear rhinorrhea. Aneurysmal dilatation of the cavernous carotids may cause a mass effect on the pituitary, which can uncommonly cause hypopituitarism. Pituitary masses that expand laterally can compress cranial nerves that run through the cavernous sinus and result in cranial nerve palsies in 1 to 14% of patients. Most commonly, patients will present with oculomotor nerve deficits, followed by abducens nerve palsy.
The pituitary is composed of two divisions: the anterior and the posterior divisions. The embryologic epithelial tissue from the primitive pharynx forms the anterior division, whereas the posterior tissues arise from neuroectodermal tissue. The differences in embryonic origins result in varied histologic compositions and functions between the anterior and posterior aspects of the pituitary gland. These different histologic compositions and cell type origins are notable since different tumor subtypes arise from these different anatomic regions, and this division is readily apparent on imaging.
Advanced imaging modalities using cross-sectional imaging have completely supplanted traditional methods for evaluating the pituitary gland. Before the 1970s, plain radiographs were used to assess for osseous changes in the sella turcica. Expansion of the sella turcica suggests the presence of a mass lesion within the pituitary fossa. However, the presence of abnormalities on skull radiographs is neither sensitive nor specific for pituitary lesions. Air contrast was also previously used for pneumoencephalograms to evaluate the sellar and parasellar regions. These techniques were quickly replaced by modern cross-sectional imaging modalities and are presented for historical reference.
Computed tomography (CT) uses axial-acquired helical x-rays to generate cross-sectional images. Post-processing reconstruction generates sagittal, coronal, and 3D-rendered images. Because CT uses X-rays, it depends on differences in the number of X-rays absorbed, also known as attenuation, of various tissues to generate the adequate contrast needed for image formation. The differences in attenuation of bone and calcified products to liquids and soft tissues make CT ideal for evaluating calcification or ossification within suprasellar lesions. This property is useful in distinguishing calcifications within a craniopharyngioma from hemorrhage within a pituitary adenoma since both may appear similar on magnetic resonance imaging (MRI). Similarly, CT is ideal for evaluating the sphenoid bone and geometric distortion or erosions of the sella turcica. An enlarged sella seen on skull radiograph can only suggest the presence of osseous remodeling around a mass lesion, whereas multiplanar CT images can fully characterize these lesions and help determine their underlying etiology. Note that sellar size may be abnormal due to conditions other than pituitary or vascular lesions, as genetic syndromes such as fragile X, trisomy 21, Turner syndrome, and Williams syndrome, to name a few.
In contrast to soft tissue mass lesions, non-contrast CT can quickly identify the presence of cystic lesions of the pituitary, which predominantly represent either Rathke’s cleft cysts or arachnoid cysts. The increased cerebrospinal fluid (CSF) pressure in patients with intracranial hypertension may cause a small amount of arachnoid and CSF to herniate through the aperture of the diaphragm sellae, the reflection of the dura mater which separates the pituitary fossa from the intracranial compartment. The arachnoid cyst results in partial effacement of the pituitary gland, with flattening of the gland on the posteroinferior aspect of the sella. Noncontrast CT can easily distinguish between the attenuation of CSF compared to the soft tissue of the pituitary gland in patients with a partially or completely empty sella. However, the presence of an arachnoid cyst is usually an incidental finding on non-contrast head CT, and it may be difficult to distinguish from other cystic lesions of the pituitary fossa.
Noncontrast CT offers limited delineation of soft tissue structures both within and adjacent to the brain, given their similar attenuation characteristics. CT is inferior to MRI in distinguishing soft tissue from small structures, such as the optic pathways. Intravenous (IV) iodinated contrast can increase the utility of CT by increasing the differences in attenuation, particularly for vascular structures, and CT angiography of the head is useful when an aneurysm is suspected.
Presurgical evaluation for patients undergoing transsphenoidal surgery requires CT imaging for image co-registration with bony landmarks. Additionally, evaluation for anatomic variants such as under-pneumatization of the sphenoid sinus or dehiscence of the bony roof overlying the internal carotid arteries is better using CT than MRI. CT angiography can be useful in evaluating for cavernous sinus invasion, as this impacts preoperative planning.
Despite the utility of CT in evaluating pituitary lesions, it is not the first-line initial imaging modality for evaluating the pituitary fossa. The acquisition speed of CT makes it an ideal study for ruling out many pathologies on an emergent basis, such as intracranial hemorrhage from pituitary apoplexy. These non-contrast CT studies often reveal incidental pituitary lesions, requiring a follow-up MRI for further characterization. In rare circumstances, patients with severe claustrophobia may not tolerate a full MRI of the brain. In these cases, performing a CT with IV contrast optimized for imaging the pituitary (sella protocol) is the next best option. It is important to note that CT has the additional disadvantage of ionizing radiation, where MRI does not. Thus, patients screened for pituitary abnormalities should receive an MRI rather than CT whenever possible. Furthermore, precontrast CT imaging offers little additional diagnostic value; thus, avoiding precontrast pituitary imaging is recommended. It is noteworthy to mention that patients with intracranial lesions often require multiple follow-up imaging studies, for which clinicians should consider the patient’s anticipated cumulative radiation dose.
Magnetic resonance imaging (MRI) is the ideal imaging modality for assessing nearly all pituitary lesions. MRIs do not require ionizing radiation like CT or traditional radiography use and instead can generate tissue contrast based on differences in the magnetic properties of hydrogen atomic nuclei. MRI can easily differentiate between solid and cystic lesions and also between different soft tissues, such as gray and white matter. Specific MRI imaging protocols are available for evaluating the pituitary. These include both high-resolution coronal and sagittal T1 weighted images with a small field-of-view centered about the pituitary. Sagittal plane images allow assessment of the pituitary, the stalk, the infundibulum, and the optic chiasm. The intrinsic T1 signal hyperintensity from the secretory granules of the posterior pituitary gland allows differentiation between its anterior and posterior divisions. Scrutinizing the relationship between the pituitary gland and the cavernous sinus structures is performed on coronal images. T2 weighted images, in combination with Fluid Attenuation Inversion Recovery (FLAIR) sequences, can reliably characterize an empty sella without the need for intravenous contrast.
An optimal assessment of sellar and suprasellar mass lesions requires both non-contrast and gadolinium-enhanced MRI imaging. Regions exhibiting gadolinium enhancement demonstrate T1 shortening, which increases the imaging signal intensity, further accentuating image contrast differences between anatomic structures. Regions of the brain without a blood-brain barrier avidly enhance postcontrast imaging. Both the pituitary gland and stalk demonstrate enhancement given their rich vascularity from the pituitary capillary plexus. Differential enhancement between a seller mass and the pituitary is particularly relevant during the presurgical evaluation, as inadvertent damage of the normal pituitary gland may result in pituitary insufficiency. Contrast also increases the sensitivity of detection of small adenomas, which typically are hypo-enhancing lesions. Intravenous contrast administration allows for the assessment of vascular encasement within the cavernous sinus. Contrast-enhanced magnetic resonance angiography (MRA) is not routinely performed during the initial evaluation of sellar lesions unless there is a concern for vascular pathologies such as an aneurysm or cavernous-carotid fistula. Follow-up MRA may be required for surgical planning if there is evidence of vascular encasement or narrowing. Gadolinium-based contrast agents are well-tolerated, but their use is associated with some risks, including rare allergic reactions, gadolinium deposition, and nephrogenic systemic sclerosis in patients with impaired renal function. A complete discussion of these risks is beyond the scope of this article.
Sellar and parasellar lesions comprise a heterogeneous group of pathologies that may be developmental, neoplastic, vascular, infectious, inflammatory, or degenerative. Most pituitary lesions are discovered incidentally and are comprised mostly of pituitary adenomas and Rathke’s cleft cysts, both easily diagnosed using contrast-enhanced MRI. Other rare diagnoses such as pituitary metastasis, abscess, infarction/hemorrhage, or epidermoid cyst demonstrate unique imaging characteristics and rarely pose a diagnostic dilemma. Seldomly, large meningiomas and craniopharyngiomas may occupy the sellar and suprasellar regions, mimicking a pituitary mass. However, associated features such as a dural tail, calcifications, or hyperostosis of the adjacent calvarium often cinch the diagnosis. Unspecified sellar and suprasellar lesions may require a biopsy for definitive diagnosis, particularly if the imaging features are concerning for metastatic disease from an unknown primary site.
Since ultrasound cannot penetrate the calvarium, there is no utility in using ultrasound for evaluating the pituitary gland.
Additional advanced imaging studies have not shown proven benefits for evaluating pituitary lesions. For example, pituitary lesions are often discovered incidentally on fluorodeoxyglucose positron emission tomography (FDG-PET) studies, but there is poor characterization since gamma camera resolution is more than 1 cm. FDG-PET is also limited since the brain uses glucose as its primary metabolite, which offers no distinction between tumor hypermetabolism and normal brain metabolism.
Radionuclide cisternography is a method that may be useful in the detection of occult cerebrospinal fluid leaks. In this technique, a radionuclide is injected into the CSF via lumbar puncture and then later imaged via a gamma camera. CSF leak is confirmed if scintigraphy demonstrates the accumulation of radioisotope in the region of the presumed leak. For patients with sellar lesions or chronic intracranial hypertension, expansion of the sella and thinning of the posterior wall of the sphenoid sinuses may result in eventual breakdown and herniation. Similar processes can occur through the ethmoid sinuses and cribriform plate. Typically, patients receive high-resolution skull base CT imaging for evaluation of bony abnormalities in patients with CSF leaks or presumed skull base lesions. However, if CT imaging is unrevealing, radionuclide cisternography may prove to be a useful ancillary study.
Traditional catheter angiography is the gold-standard for the assessment of the intracranial vasculature. However, it is an invasive procedure and presents inherent procedural risks associated with vascular access as well as thromboembolic events within the cerebral vasculature. The workup of sellar and suprasellar lesions does not routinely include traditional angiography, but before the advent of CT and MRI, traditional angiography was the imaging modality to evaluate for possible mass effect on the cavernous carotids. The single advantage traditional angiography presents over CTA and MRA is that catheter-based interventions can take place at the time of angiography. If a patient requires endovascular treatment of an intracranial aneurysm, the vascular surgeon can perform them at the time of the angiogram. Outside of this indication, traditional angiography is typically not indicated.
Before the development of multi-detector computed tomography and isotropic acquisition, patient positioning during CT studies was incredibly valuable. Current technology allows for acquisition in nearly any position, and post-processing adjustments will occur after image acquisition. Standard views of the pituitary gland can be obtained by image reconstruction and formatting to standard coronal, sagittal, and axial anatomic landmarks. Since MRI can acquire imaging data in multiple planes, patient positioning is also less critical.
Pituitary lesions are common and incidentally discovered on many CTs performed for unrelated reasons. The use of traditional skull radiography and pneumoencephalography to assess pituitary lesions is only of historical significance. Contrast-enhanced CT imaging of the brain using a pituitary protocol offers limited diagnostic value compared to MRI when evaluating sellar and suprasellar lesions. However, calcifications and bony structures are optimally evaluated using CT and may be necessary for preoperative planning for surgical resection of pituitary masses. The initial recommended diagnostic test for pituitary lesions is a specialized MRI pituitary protocol, both with and without intravenous contrast. Few pituitary lesions demonstrate overlapping imaging findings on MRI and may require tissue biopsy for definitive diagnosis. Other advanced imaging modalities such as ultrasound play no role in evaluating the pituitary, but in select circumstances, radionuclide cisternography may prove to be a useful adjunct study in evaluating for occult CSF leaks.