Detection of adrenocorticotropin-secreting pituitary adenomas by magnetic resonance imaging in children and adolescents with cushing disease.
ABSTRACT We recently showed that pre- and postcontrast spoiled gradient-recalled acquisition in the steady-state (SPGR) was superior to conventional pre- and postcontrast T-1 weighted spin echo (SE) acquisition magnetic resonance imaging (MRI) for the diagnostic evaluation of pituitary tumors in adult patients.
The present investigation assessed the use of SPGR vs. SE-MRI in the diagnostic evaluation of ACTH-secreting tumors in children and adolescents with Cushing disease.
Data were analyzed retrospectively from a series of patients seen over 7 yr (1997-2004).
The setting for this study was a tertiary care referral center.
Thirty children with Cushing disease (13 females and 17 males with a mean age of 12 +/- 3 yr) were studied.
Imaging results were compared with surgical and pathological findings and the clinical outcome.
Twenty-eight patients had microadenomas, and two had macroadenomas; the latter were identified by both MRI techniques. Precontrast SE and SPGR-MRI identified four and six of the microadenomas, respectively. Postcontrast SPGR-MRI identified the location of the tumor in 18 of 28 patients, whereas postcontrast SE-MRI identified the location and accurately estimated the size of the tumor in only five patients (P < 0.001).
We conclude that conventional MRI, even with contrast enhancement, mostly failed to identify ACTH-secreting microadenomas in children and adolescents with Cushing disease. Postcontrast SPGR-MRI was superior to SE-MRI and should be used in addition to conventional SE-MRI in the pituitary evaluation of children and adolescents with suspected Cushing disease.
- SourceAvailable from: George Edward Vates[Show abstract] [Hide abstract]
ABSTRACT: Recurrent pituitary disease presents unique challenges, including in some cases difficulty localizing a tumor radiographically. Here, we present the case of a patient with recurrent Nelson syndrome whose radiographic work-up was complicated by a significant parasellar metallic artifact. Positron emission tomography ultimately localized the lesion, and coregistration with computed tomography allowed for accurate intraoperative navigation. Additionally, we review a range of imaging techniques available in the evaluation of pituitary disease.Journal of neurological surgery reports. 06/2013; 74(1):33-6.
- Journal of Neuro-Oncology 12/2013; · 3.12 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Cushing's disease (CD) is the most common cause of endogenous Cushing's syndrome in children and adolescents and represents a rare cause of short stature. A 14-year-old boy came to our attention for progressive weight gain and short stature. At examination, height was 140 cm (3rd centile) and weight was 37.7 kg (10th centile). Tanner stage was G2, PH 3, testis 3 mL. Hypothyroidism and growth hormone deficiency were excluded. A marked increase of urinary free cortisol, a nonsuppressible serum cortisol after Liddle 1 test, and an elevated ACTH value confirmed the diagnosis of ACTH dependent Cushing's syndrome. Pituitary magnetic resonance imaging (MRI) showed a left microadenoma and a right focal area of lesser enhancement. Therefore, bilateral inferior petrosal sinus sampling (BIPSS) with CRH stimulation was performed to obtain an accurate preoperative localization of the adenoma: the interpetrosal sinus ACTH gradient indicated lateralization of ACTH secretion to the left side. The patient underwent transsphenoidal surgery with selective microadenomectomy, with an immediate ACTH decline in the postoperative phase. Histology confirmed the diagnosis of corticotrophic pituitary adenoma. Glucocorticoid replacement therapy was instituted. Clinical examination demonstrated a rapid catch-up growth (10th centile), with a normalization of body mass index and an adequate pubertal development.Case reports in endocrinology. 01/2014; 2014:851942.
J. Clin. Endocrinol. Metab. published online Jun 7, 2005; , doi: 10.1210/jc.2004-1778
Constantine A. Stratakis
Dalia Batista, Nickolas A. Courkoutsakis, Edward H. Oldfield, Kurt J. Griffin, Meg Keil, Nickolas J. Patronas and
(MRI) in children and adolescents with Cushing disease
Detection of ACTH-secreting pituitary adenomas by magnetic resonance imaging
Society please go to: http://jcem.endojournals.org//subscriptions/
or any of the other journals published by The Endocrine
Journal of Clinical Endocrinology & Metabolism
To subscribe to
Copyright © The Endocrine Society. All rights reserved. Print ISSN: 0021-972X. Online
- 1 -
Detection of ACTH-secreting pituitary adenomas by magnetic
resonance imaging (MRI) in children and adolescents
with Cushing disease
Dalia Batista MD 1#, Nickolas A. Courkoutsakis MD, PhD 1, 2,*#, Edward H.
Oldfield, MD 3, Kurt J. Griffin MD 1, Meg Keil, RN, PNP 1, Nickolas J.Patronas
MD2 & Constantine A. Stratakis MD, D(med)Sci1**
(1) Department of Diagnostic Radiology, Warren Grant Magnuson Clinical Center, National
Institutes of Health (NIH), Bethesda, MD 20892,
(2) Section on Endocrinology & Genetics (SEGEN), Developmental Endocrinology Branch,
(DEB), National Institute of Child Health and Human Development (NICHD), Bethesda, MD
(3) Surgical Neurology Branch, National Institute of Neurological Diseases and Stroke (NINDS),
Bethesda, MD 20892.
(4) Center for Statistical Sciences, Brown University, Providence, RI 02912
* Present address: Department of Radiology, Medical School, Democritos University of Thrace,
68100 Alexandroupolis, Greece.
# Drs. Courkoutsakis and Batista have contributed equally to this work and are thus sharing first
Key Words: Cushing disease, magnetic resonance imaging, pituitary, children, trans-sphenoidal
Short title: MRI for Cushing disease
** Adress all correspondence and reprint requests to: Dr. Constantine A. Stratakis, Section on
Endocrinology & Genetics, DEB, NICHD, NIH, Building 10, Room 10N262, 10 Center Dr., MSC 1862,
Bethesda, Maryland 20892-1862, Tel: 301-4964686/4021998, Fax: 301-4020574, E-mail:
Journal of Clinical Endocrinology & Metabolism. First published June 7, 2005 as doi:10.1210/jc.2004-1778
Copyright (C) 2005 by The Endocrine Society
- 2 -
Context: We recently showed that pre- and post-contrast spoiled gradient-recalled acquisition in
the steady state (SPGR) was superior to conventional pre- and post-contrast T-1 weighted spin
echo (SE) acquisition magnetic resonance imaging (MRI) for the diagnostic evaluation of
pituitary tumors in adult patients. Objective: The present investigation assessed the use of SPGR
versus SE-MRI in the diagnostic evaluation of ACTH-secreting tumors in children and
adolescents with Cushing disease. Design: Data were analyzed retrospectively from a series of
patients seen over 7 years (1997 to 2004). Setting: A tertiary care referral center. Patients: 30
children with Cushing disease (13 females and 17 males with a mean age of 12 + 3 years old)
were studied. Interventions and outcome measures: Imaging results were compared with surgical
and pathological findings and the clinical outcome. Results: 28 patients had micro- and 2 macro-
adenomas; the latter were identified by both MRI techniques. Pre-contrast SE and SPGR-MRI
identified 4 and 6 of the microadenomas, respectively. Post-contrast SPGR-MRI identified the
location of the tumor in 18 of 28 patients, whereas post-contrast SE-MRI identified the location
and accurately estimated the size of the tumor in only 5 patients (P<0.001). Conclusions: We
conclude that conventional MRI, even with contrast enhancement, mostly failed to identify
ACTH-secreting microadenomas in children and adolescents with Cushing disease. Post-contrast
SPGR-MRI was superior to SE-MRI and should be used in addition to conventional SE-MRI in
the pituitary evaluation of children and adolescents with suspected Cushing disease.
- 3 -
Detection of adenomas in the pituitary gland by imaging techniques has always been
problematic. ACTH-secreting adenomas, in particular, have been proven difficult to localize
because these tumors are typically very small (1-7). Magnetic resonance imaging (MRI) has been
considered for the last two decades as the imaging method of choice for detection of these
tumors (1-8). Various modifications of MRI techniques have been proposed to improve the rate
of tumor detection (7, 9-12). However, and even though MRI sensitivity for pituitary
microadenomas varies widely among different investigators, it is generally accepted that these
tumors are correctly identified by MRI only in approximately half of the cases (1, 2, 4-7, 10-12).
One of the main reasons for misdiagnosis is that pituitary tumors and the surrounding normal
pituitary gland often exhibit similar signal and enhancing characteristics. (2, 7, 11-15).
Successful treatment of ACTH-secreting adenomas requires accurate diagnosis and exact
localization. Because of limitations of imaging (12), many institutions use petrosal sinus
sampling (PSS) to distinguish a pituitary from an ectopic source of ACTH (11). While PSS has
high diagnostic accuracy, it is an invasive and expensive test that is not widely available and
carries a finite risk of serious complications. It is also difficult to perform in young children, who
have the added risks of sedation or anesthesia. Thus, an improved imaging technique for pituitary
tumors in children with Cushing disease would be a significant diagnostic advancement in the
work up of these patients.
- 4 -
In a recent pilot study of mostly adult patients, we demonstrated that spoiled gradient-recalled
acquisition in the steady state (SPGR) MRI was superior to the conventional T-1 weighted spin
echo (SE) technique in identifying pituitary tumors (11). As SPGR-MRI can be performed in
thin sections of one millimeter thickness, the spatial resolution of the acquired images is
substantially improved (16, 17). We hypothesized that this technique might increase the accuracy
of MRI for the detection of ACTH-secreting pituitary adenomas and assist in decreasing
morbidity associated with trans-sphenoidal surgery (TSS) for Cushing disease in childhood (18-
21). To evaluate this possibility, we compared the results of MRI using the SPGR technique and
the conventional T1-weighted SE technique in 30 children with Cushing disease and surgically-
proven ACTH-secreting microadenomas.
- 5 -
Subjects & Methods
Patients and protocol
A total of 30 patients (Table 1) were studied that were admitted consecutively to the National
Institutes of Health-Warren Magnuson Clinical Center from January 1997 to August 2003, under
investigational protocol 97-CH-0076, after an institutional review board approval and individual
patient consents. The mean age was 12 + 3 years (mean and standard deviation [SD]), and the
group consisted of 13 females and 17 males. Cushing syndrome was diagnosed in all patients by
standard testing as described elsewhere (22). After confirmation of hypercortisolism,
biochemical testing for Cushing disease included PSS, if the standard T1-weighted SE MRI
technique was negative. Twenty-four of our patients underwent PSS for confirmation of Cushing
disease, using the procedure that we have described elsewhere (23). All patients underwent TSS
for identification and excision of a pituitary adenoma. The size and the exact location of the
adenoma were identified at surgery using ring curettes of known diameter. Histo-pathological
examination of the surgical specimen and post-operative biochemical evidence of remission of
disease (hypocortisolism) confirmed the diagnosis of an ACTH-secreting adenoma.
All MRI scans were performed in a 1.5 T scanner (Signa, General Electric, Milwaukee, WI).
Two imaging techniques were used: a) coronal pre-contrast T1-weighted SE with repetition
time/echo time 400/9 miliseconds (msec); 192 x 256 matrix; two excitations; 12 cm field of
view; and 3mm in thickness interleaved sections without gap b) coronal pre-contrast SPGR with
repetition time/echo time 9.6/2.3 msec, a 20 degree flip angle, 160 x 256 matrix, six excitations,
- 6 -
and 12 cm field of view. Contiguous 1.5 mm thick sections were obtained in all patients. The
scan time was approximately 3.4 minutes. Both SE and SPGR studies were repeated after
intravenous administration of 0.01 mmol/kg gadopentetate dimeglumine (Magnevist, Berlex
Laboratories, Inc., Mintville, NJ).
For the purposes of the present study, all imaging studies were reviewed retrospectively by two
radiologists (NJP, NAC) in a blinded fashion, e.g. without knowledge of the surgical outcome
and final histology. The radiological interpretation of all images was made first independently by
each radiologist; a meeting with the protocol team was then arranged if there was disagreement.
A scan was considered “positive” only if both radiologists agreed either after their first
interpretation or their re-evaluation after their meeting. Imaging findings were then recorded
(presence, size and position -right, left, central- of any lesion) by the protocol team in a blinded
manner: all analyses and comparisons to surgical and histopathologic findings were done by
members of the study team that had not participated in the radiological interpretation meetings.
Sensitivity and specificity of both techniques were calculated by comparing the imaging data to
the surgical findings, considering the surgical results to represent the gold standard (Table 2).
The difference between the sensitivities of both imaging techniques were calculated using a
sample test of correlated proportion (Table 3); 95 percent confidence intervals (CI) were then
calculated (24, 25). Differences in tumor size were compared by paired t-tests. A P value < 0.05
was considered significant.
- 7 -
All tumors detected on the pre-contrast scans were hypo-dense; likewise, all lesions
identified on the post-contrast scans demonstrated decreased enhancement with respect to the
normal pituitary parenchyma regardless of the technique (SE or SPGR) (Figure 1). When tumors
were visible by both techniques, the SPGR technique generally provided a sharper, clearer image
(compare figure 1b with figure 1c). Two patients had a macroadenoma and were identified by
both techniques. While the largest tumors were identified with both methods, the smaller tumors
were only seen with the SPGR technique (Table 1). No tumors found on the SE-MRI were
missed by the SPGR scans. No patients with a micro-adenoma missed on the post-contrast scans
had an indirect sign of a pituitary tumor (such as deviation of the stalk or unilateral superior
convexity) that could have assisted in the correct diagnosis (2, 9). There was no good correlation
between PSS and imaging findings. PSS was performed in 24 of the 30 children; localization of
the tumor by PSS was correct in approximately only 50% of the cases, but a larger investigation
of PSS data in children is currently ongoing (Batista, Stratakis, unpublished data)
Surgical findings and outcome
No children with ectopic ACTH production were identified. All 30 patients (Table 1)
were in remission immediately after TSS, confirmed by undetectable levels of urinary free
cortisol postoperatively; however, recently we identified 2 patients that recurred and in whom
- 8 -
histology had identified ACTH-producing cells. These 2 patients, who also had a negative MRI,
were considered as having negative surgical exploration (Table 2), even though the surgical
report identified a lesion in each case (Table 1). For one of these patients, pre-contrast SE- and
SPGR-MRI studies were not available; in an additional 2 patients with negative MRIs the tumor
was not found in the first exploration (listed as a “negative” surgical result in Table 2), but
eventually a tumor was found in a second procedure and its size and location are recorded in
In 2 adolescent patients, a macro-adenoma was evident by both SE and SPGR techniques.
TSS identified adenomas with vertical diameters measuring 22mm and 18mm, respectively. The
remaining patients had microadenomas (size < 1 cm) with 9 found on the right, 8 on the left side,
and 11 near the midline of the gland. Histologic examination did not reveal tissue degeneration,
cyst formation, hemorrhage or necrosis in any of the excised adenomas. The size of the micro-
adenomas varied from 2.5 to 10mm; there were no significant differences between the mean size
of the tumor at the time of surgery versus that in post-contrast SPGR-MRI (7.6+0.4 vs. 5.6+0.98,
Data analysis: SE- versus SPGR-MRI.
Two patients had a macroadenoma and were identified by both techniques; these patients
were excluded from this analysis. When we compared imaging results versus surgical findings in
the remaining patients (Table 2), in the pre-contrast SE-MRI, 14.2% (4/28) patients had a
microadenoma. In the post-contrast SE-MRI, 18% (5/28) of the patients had a microadenoma
- 9 -
that was identified correctly after taking into account the surgical report. In contrast, in SPGR-
MRI, in the pre-contrast study, 21% (6/28) of the patients had a microadenoma; following
contrast-enhancement, SPGR-MRI identified an adenoma in 64% (18/28) of the patients.
When the sensitivity was calculated comparing imaging testing vs. the location of the
adenoma at surgery (Table 2) the sensitivity for SE-MRI (pre-contrast T1) was 16% ( 4/25, CI95:
5 - 40%) and the specificity was 66.7% (2/3, CI95: 13 – 98%)). For the post-contrast SE-MRI the
estimated sensitivity was better at 21% (5/24, CI95:8 – 43%), with a specificity of 50% (2/4, CI95:
9 90%). However, the corresponding values for SPGR-MRI were significantly higher. The
sensitivity for pre-contrast SPGR-MRI was 24% (6/25 CI95: 8 – 41%) and specificity was 67%
(2/3 CI95: 13 – 98%). For the post-contrast SPGR MRI, sensitivity was 75% (18/24, CI95: 53 –
89%) and the specificity was 50% (CI95: 9 – 91%). The difference between the sensitivities of
the post-contrast SE-MRI vs. that of the SPGR-MRI was quite significant at 55% (CI95: 36 –
74%; p=0.001) (Table 3).
Both techniques had high false negative rates, but the numbers for SPGR-MRI were
significantly better. For SE-MRI (post-contrast) this was 90% (CI95: 68 - 98%) with a false
positive rate of 28% (CI95: 5 - 70%). For SPGR-MRI (post-contrast), the values were 75% (CI95:
36 - 96%) and 10% (CI95: 2 - 33%), respectively. Finally, the overall probability that post-
contrast SE-MRI would be positive in a child with surgically-proven Cushing disease was 25%
(CI95: 11% - 45%), whereas the probability that it would be negative was 75% (CI95: 55 - 86%).
On the other hand, for post-contrast SPGR-MRI the probability that the test would be positive
was 71% (C95:51 – 86), whereas the probability that the test would be negative was 29% (CI95:
- 10 -
14 - 50%). Two representative examples of patients with negative SE-MRIs but positive SPGR-
MRIs are shown in figures 2 and 3. In figure 3, the use of contrast enhancement identified a
lesion that was not visible by plain SPGR-MRI, too.
- 11 -
Our overall success rate for TSS for Cushing disease at the National Institutes of Health
for the period 1997 – 2004 is 97.6% (without long-term follow up data, unpublished
information). The SPGR technique improved by approximately two-fold the detection rate by
MRI of pituitary adenomas in a pilot study of mostly adult patients (11). In adults, SE-MRI
(post-contrast) had a detection rate of 49% (CI95: 34-63%), whereas that of the SPGR-MRI (post-
contrast) was 80% (CI95: 68 – 91%) (11).
Cushing disease is extraordinarily rare in children and adolescents (19, 20, 21). The
majority of pediatric patients with Cushing disease have small microadenomas as the cause of
their disease; in addition, pituitary “incidentalomas” are rare in children (20-22, 26). The present
study is the largest that specifically evaluated the use of any type of MRI for detection of
pituitary corticotropinomas in children; but we also evaluated a new modality, SPGR-MRI . The
detection rate of 64% with a sensitivity of 75% represents a significant improvement over
conventional imaging techniques that are currently in use.
The low detection rate of these tumors in children by conventional imaging, which in the
present study was at approximately 20%, was less than half of that reported in adults (1, 7, 11).
Possible explanations for this difference are the relatively smaller size of ACTH-secreting
adenomas in children (22) and the absence of tissue degeneration, necrosis or cyst formation in
pituitary adenomas of younger patients(24). These tissue changes provide additional elements of
- 12 -
separation from adjacent normal tissue, which make the presence of an adenoma more obvious
upon radiological studies (2, 5, 6).
A number of factors may account for the apparent superior performance of the SPGR
compared with the SE technique (7-12). The increased soft tissue contrast with SPGR is well-
acknowledged; with it, images can be obtained with 1.5-mm thin sections (16, 17) (Figure 1).
SE-imaging, on the other hand, is usually obtained with 3-mm thickness and averaging of
different tissues can obscure smaller pituitary microadenomas (7-11). The main drawback of the
SPGR technique is its inferior signal to noise ratio compared with SE-MRI (11, 16, 17). In this
study we doubled the number of excitations from 3 to 6 to alleviate this problem. In doing so, the
specificity of the findings was improved while the scan time was still shorter than that of the SE
technique, thus, minimizing motion artifacts that can be encountered with prolonged scanning.
As was the case in adult patients (11), we were surprised to detect pituitary tumors in
pediatric patients who were referred after a previously negative MRI had been obtained
elsewhere. Is it possible that these tumors grew sufficiently in the interval between the two
studies to allow detection? Although this is a possibility, it is unlikely, as our imaging usually
took place within six months from the previous study. First, pituitary tumors typically do not
increase in size over an interval of only a few months. Second, the tumors that were detected by
SPGR-MRI in the course of our study, after a recent negative MRI elsewhere, covered the same
range of sizes as other corticotropinomas that have been reported in the literature (1, 2, 11); a
recent size increase would suggest that the newly detectable tumors would be, on average, of
smaller size. Third, in the previous study (11), we reviewed films obtained at other institutions
- 13 -
and found a great variability in the technical aspects of the scanning procedure and,
consequently, the quality of imaging (11). Optimal detection of pituitary adenomas requires the
use of imaging techniques developed for pituitary studies, not brain; it is generally better, if these
studies, especially in children, are obtained at specialized, tertiary-referral centers for Cushing
The difference in the rate of detection using the two techniques was not statistically
different in the pediatric population, whereas in adult patients, there were too few patients with
false positive findings to make a meaningful comparison (11). The false positive rate in pediatric
patients was 28% (CI95: 5 - 70%) and 10% (CI95: 2 - 33%) for the SE- and SPGR-MRI (post-
contrast), respectively. This is comparable to the 18%-20% rates reported by other centers for
conventional pituitary MRI, but contrasts with the 4% rate found in the study of adult patients
(11), and the 0% rates in earlier studies from the NIH (1). It is possible, therefore, that SPGR-
MRI, at least in children, increases the chances of falsely detecting a pituitary tumor. From the
surgical and pathological findings, we know that MRI-detected abnormalities did not represent
additional pituitary lesions, which in some cases of Cushing disease may co-exist (27).
The surgical cure rate for Cushing disease is 80 to 90% when MRI localizes a tumor, but
drops to 50 to 70% when the lesion cannot be demonstrated upon pre-operative imaging (28 -
31). Since in a patient with endocrine testing consistent with Cushing disease a positive MRI
almost always indicates the site of the ACTH-secreting adenoma, as it did in the current series,
and since pituitary exploration with multiple incisions in the gland can produce hypopituitarism
in some patients, our surgeon (EHO) begins exploration for the adenoma based on the site of the
- 14 -
positive MRI. If a tumor is found there, the tumor is removed and no further exploration of the
gland is performed. Thus, the correct localization of the tumor on MRI not only provides higher
remission rates, but also, by limiting the exploration of the pituitary gland that is necessary, it
may reduce the incidence of postoperative complications, such as cerebrospinal fluid leak and
hypopituitarism. Suggestive localization on MRI guides the surgeon during TSS, even if the
imaging study was not definitive (32).
We conclude that coronal post-contrast SPGR images should be added to conventional
SE-imaging protocols of the pituitary gland: SE images are complementary to SPGR images for
the diagnosis of corticotropinomas and both techniques should be used for the investigation of
the pituitary gland in all patients with Cushing syndrome. While imaging alone can not establish
the diagnosis of Cushing disease in patients with Cushing syndrome, it plays a crucial supporting
role to diagnostic endocrine testing.
We would like to thank, first, our patients who participated in NICHD 97-CH0076
investigational protocol (PI: Dr. Stratakis) and, second, Dr. Alicia Y. Toledano, from the Center
for Statistical Sciences at Brown University, (Providence, RI 02912) for her review of our
statistical analysis and presentation.