Pituitary Magnetic Resonance Imaging in Normal Human Volunteers: Occult Adenomas in the General Population

Methods

Seventy women and 30 men, 18 to 60 years old, were recruited from the general population and the normal-volunteers office of the National Institutes of Health. The volunteers had normal physical examinations and were selected by age and sex to correspond to the distribution of patients with symptomatic pituitary adenomas [6, 7]. Fifty-eight of the women (83%) were between 20 and 45 years old, and 27 of the men (90%) were between 30 and 60 years old. Random basal values of serum prolactin and α-subunit, plasma growth hormone, thyroid-stimulating hormone, thyroxine, triiodothyronine, free thyroxine, and thyroxine-binding globulin were measured. Young women selected for the study had normal menstrual cycles. Persons with previous or current endocrine disturbances were excluded. The protocol was approved by the Investigational Review Board of the National Institute of Neurological Disorders and Stroke, and all participants gave informed consent.

Magnetic Resonance Imaging

All scans were obtained with a 1.5-T scanner (Sigma, General Electric; Milwaukee, Wisconsin). T1-weighted coronal and sagittal images of the pituitary fossa were obtained with a repetition time (TR) of 600 ms and an echo time (TE) of 15 ms (TR/TE = 600/15). In the coronal plane, interleaved sections 3 mm in thickness without intersection gap were obtained with two repetitions and a 16-cm field of view. The acquisition matrix was 256 × 192. Gadolinium-DTPA (0.1 mmol/kg body weight) was administered intravenously over 2 minutes, and the T1-weighted coronal images were repeated immediately (Figure 1).

Evaluation of Pituitary Glands

Magnetic resonance scans were interpreted independently by three experienced reviewers. The scans of the normal volunteers were randomly mixed with scans of 57 patients with surgically confirmed Cushing disease. The reviewers were aware that the scans of the patients had been intermingled with those of the volunteers. Identifying information for patients and volunteers was masked. Reviewers evaluated the pituitary gland before and after the administration of Gd-DTPA by appraising gland and sellar size, stalk deviation, gland convexity, and position and size of focal pituitary abnormalities. Measurements were made with hand-held calipers. The readers also provided a summary interpretation of each MRI scan. The diagnosis of a pituitary abnormality, including the presence of an adenoma, was accepted only when the same area of the gland was independently interpreted as similarly abnormal by at least two of the three reviewers.

Results

Pituitary Gland Magnetic Resonance Scans in Normal Volunteers

Pituitary Size and Shape

The mean gland height in the 100 volunteers was 6.9 ±0.1 mm. It was greater in women (7.1 ±1.3 mm; mean ±SD) than in men (6.6 ±1.2 mm; P = 0.008). On coronal scans, the superior surface of the gland was convex upward in 33 persons (29 women). Upward convexity of the superior surface of the gland was limited to one side in 21 persons (12 right, 10 left) and occurred centrally in 11. The posterior pituitary gland was identified as a small focus of high-signal intensity in the posterior sella by at least two reviewers in 92 persons (by three reviewers in 77). One participant was considered to have an enlarged sella. Fifty-nine volunteers were interpreted by all three reviewers as having a normal pituitary gland. In 3, cerebrospinal fluid filled the superior portion of the sella (partially empty sella).

Focal Pituitary Abnormalities

Before Gd-DTPA was administered, 22 sites of focal abnormal signal intensity in the pituitary were detected by at least one reviewer in 21 volunteers (Table 1). All three reviewers considered the same site as an adenoma (by the presence of an area of low-signal intensity) in 1 person. Seven (6 women, 1 man) had focal areas of decreased signal, which were interpreted by at least two of the three reviewers as adenomas. Fourteen sites in 13 persons were considered abnormal by a single reviewer.

After Gd-DPTA was administered, 41 different sites of abnormal signal intensity in the pituitary gland were detected in 34 volunteers. Ten of them (7 women, 3 men) had focal areas of decreased signal intensity that were interpreted as pituitary adenomas by at least two reviewers. All three reviewers considered the same site as an adenoma in 2 persons. Six of the 7 women were 25 to 45 years old (1 was 48 years old) and the men were 22, 35, and 53 years old. The lesions were 3 × 3 mm to 6 × 6 mm in diameter (coronal plane). Thirty-two sites in 23 participants were considered abnormal by a single reviewer.

The scan of only 1 volunteer interpreted as having an adenoma had stalk deviation. An upward convex shape of the superior surface of the pituitary gland occurred in 8 of the 10 volunteers considered to have an adenoma. However, in only five of these studies was the elevation on the side of the lesion.

Endocrine Screening Tests

The endocrine studies done were normal in the 10 persons with an abnormal MRI scan. In the 90 volunteers whose MRI scans were interpreted as normal, 3 (1 woman, 41 years old; 2 men, 37 and 48 years old) had evidence of mild primary hypothyroidism. Three women (33, 34, and 44 years old) had elevated growth hormone levels (>10 µg/L; single random sample).

Pituitary Gland Magnetic Resonance Scans in Patients with Cushing Disease

Cushing disease was confirmed by adrenocorticotropin staining of an adenoma or by remission of hypercortisolism after selective adenomectomy or hemihypophysectomy. Five macroadenomas and 45 microadenomas were identified at surgery in the 57 patients with Cushing disease. In the 50 patients with adenomas identified at surgery, 56% (all patients with macroadenomas and 51% of 45 patients with microadenomas) had areas of focal low-signal intensity after administration of Gd-DTPA that were diagnosed by at least two reviewers as adenomas. However, in 4 of the 23 patients with microadenomas and a focal pituitary MRI abnormality, the position of the adenoma in the gland was incorrect as read on the MRI scan. In two patients, the MRI scan indicated a lateral tumor, but the tumor was found in the midline at surgery. In the third patient the reverse occurred, and in the fourth the tumor was found on the opposite side of the gland. Thus, 26 of 45 (58%) of the 45 microadenomas that were large enough to be found and selectively excised at surgery were not detected by MRI. The size of the tumors that were not detected (6 ±2 mm) did not differ significantly from the size of those that were (5 ±2 mm).

Discussion

The configuration of the pituitary gland is influenced by age and sex, the location of the carotid arteries, the shape of the pituitary fossa, transmission of cerebrospinal fluid pulsation into the sella turcica permitted by an incompetent diaphragma sella, and the presence of an intrasellar mass [8-12]. The results of our study of the size and shape of the normal pituitary gland confirm the findings of other studies using computed tomography (CT) or MRI scanning without Gd-DPTA enhancement [11, 12]. A convex contour of the superior surface of the pituitary gland was once considered to be unusual or abnormal [3, 11, 12]. (^However), recent studies using CT or MRI scanning found that convex superior contours occur in as many as 35% to 44% of women of childbearing age [9, 12, 13]. An upward convex shape of the superior surface of the pituitary gland occurred in 34% of our 70 women. Deviation of the pituitary stalk also has been suggested as an indirect sign of the presence and site of an adenoma. However, rates of stalk deviation as high as 46% have been reported in normal persons [14]. Stalk deviation occurred in 13% of our volunteers. The 4% incidence of an empty sella in our volunteers agrees with the 3% to 4% prevalence reported by others using CT [3, 12] and MRI [13].

Autopsy series estimate the prevalence of asymptomatic pituitary adenomas to be 1.5% to 27% [1-6]. Microscopic examination of the pituitary gland at autopsy is much more sensitive than contrast-enhanced MRI scanning for detecting adenomas, and immunohistochemical techniques allow direct hormonal assessment of the tissue. However, the prevalence of asymptomatic adenomas in a young adult cohort is unknown because autopsy series do not accurately reflect this population. Young women, the group in which symptomatic pituitary tumors occur most frequently, are under-represented in autopsy studies.

Previous imaging techniques, such as plain roentgenography, polytomography, and pneumoencephalography, were unreliable for detecting pituitary microadenomas [1]. Shortly after its introduction, CT scanning became the diagnostic procedure of choice for examining the pituitary gland because it was noninvasive, offered improved resolution, and, for the first time, permitted direct visualization of the pituitary gland. By high-resolution, contrast-enhanced CT scanning, Wolpert and colleagues [12] identified pituitary lesions suggestive of adenoma in 12 of 107 normal women. However, the sensitivity of CT scanning for detecting microadenomas subsequently identified at surgery is as low as 17% to 22% [15-18].

Because of its superior contrast resolution and the facility of imaging in the coronal and sagittal planes, MRI is the preferred imaging modality to investigate pituitary dysfunction. Before contrast enhancement, the normal pituitary gland is isointense with brain on T1-weighted images. A hypointense focus in the gland is considered suggestive of a pituitary adenoma [19]. The introduction of Gd-DTPA and MRI scanning with 1.5-tesla magnets has increased the sensitivity of MRI for detecting small microadenomas [20-23]. The absence of the blood-brain barrier in the pituitary stalk and gland is responsible for their marked contrast enhancement on T1-weighted MRI scans. A focal hypointense area within the gland immediately after Gd-DTPA is abnormal and is the most common appearance of an adenoma [19-23]. However, asymptomatic focal pituitary lesions, such as pars intermedia cysts, metastases, infarctions, epidermoid cysts, and abscesses [3], also may appear on MRI scans as an area of low-signal intensity after administration of Gd-DTPA. In 6 (11%) of our 57 patients with surgically confirmed Cushing disease, a hypointense lesion on preoperative MRI did not correspond to the site of a microadenoma at surgery. False-positive pituitary MRI scans also have been reported in patients with ectopic adrenocorticotropin secretion and the Cushing syndrome [24, 25].^However, because of the high prevalence of an adenoma at surgery at the site of a focal hypointense area after administration of Gd-DTPA, this finding is accepted as diagnostic of a pituitary microadenoma in a patient with pituitary endocrinopathy [19-23, 26].^Further, our study showed a positive predictive value of 86% for a positive MRI scan for the presence of an adenoma at that site at surgery in the patients with Cushing disease whose scans were intermixed with those of normal persons. Thus, most focal areas of low-signal intensity in the pituitary gland after Gd-DTPA administration are adenomas.

About 70% of pituitary tumors occur in women between the ages of 20 and 45 years [6, 7]. Symptomatic pituitary tumors occur much less often in men and affect an older age group, with a peak incidence between 35 to 60 years [6, 7]. Because previous reports emphasize the sensitivity of MRI in detecting functioning adenomas but do not address the frequency of positive studies in patients who do not have an endocrinopathy, the prevalence of “incidentalomas” detected by MRI is unknown. To simulate the group known to have the highest prevalence of symptomatic pituitary adenomas, we evaluated 70 women, 83% of whom were between 20 and 45 years old, and 30 men, 90% of whom were 30 to 60 years old. Because the scans were intermixed with the scans of 57 patients with surgically proven adrenocorticotropin-secreting adenomas, the review closely simulated the clinical situation in which the radiologist or clinician is asked to interpret an MRI examination of the pituitary gland in a patient known, or suspected to have, an endocrinopathy. Ten percent of the 70 asymptomatic women and 10% of the 30 asymptomatic men had focal pituitary lesions detected by high-resolution MRI with Gd-DPTA enhancement that were interpreted as adenomas by at least two of three blinded reviewers. The true prevalence of asymptomatic adenomas is probably higher because the sensitivity of MRI for detecting tumors smaller than 10 mm is limited (it was only 42% in our 45 patients with Cushing disease and microadenomas at surgery) and is certainly less than that of postmortem microscopic studies, in which as many as 95% of asymptomatic adenomas are less than 3 mm in diameter [1, 4].

These findings concur with the assumption that most pituitary adenomas remain asymptomatic [2]. Women and men in the age range with the highest prevalence of symptomatic adenomas of the pituitary gland have a high prevalence of focal pituitary lesions that are similar to adenomas on MRI scanning, which are not symptomatic and which only rarely become symptomatic. In patients with endocrinopathy potentially attributable to a pituitary disorder, MRI may reveal a focal pituitary abnormality in about 10%, regardless of the cause of the endocrinopathy. This prevalence will increase with technical advancements in imaging that enhance resolution and increase the sensitivity for detecting the smaller microadenomas.