Neuroradiolog y / Head and Neck Imaging • Original Research

Saindane et al.
MRI of Empty Sella Turcica

Neuroradiology/Head and Neck Imaging

Original Research
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved

Factors Determining the Clinical

Significance of an “Empty”
Sella Turcica
Amit M. Saindane1 OBJECTIVE. Although often incidental, the “empty” sella turcica can reflect chronically
Paolo P. Lim1 elevated intracranial pressure (ICP). It is particularly common in the setting of idiopathic in-
Ashley Aiken1 tracranial hypertension (IIH). This study evaluated which clinical and MRI findings could be
Zhengjia Chen2 used to differentiate patients with chronically elevated ICP from those with incidental empty
Patricia A. Hudgins1 sella turcica.
MATERIALS AND METHODS. Forty-five patients with definite IIH and 92 patients
Saindane AM, Lim PP, Aiken A, Chen Z, Hudgins PA with “empty sella” reported on brain MRI were evaluated. Measurements of the sella turcica,
diaphragm sella, pituitary gland, infundibulum, and scalp and neck soft tissues were made
on MR images. These measurements, age, sex, clinical symptoms, and frequency of previ-
ously reported orbital findings of IIH were compared between the IIH and incidental empty
sella turcica groups. Measurements on MRI were correlated with patient age in each group.
RESULTS. The IIH and incidental empty sella turcica groups had statistically similar sel-
lar, pituitary, and infundibular measurements. The patients with IIH were significantly younger
than the patients with incidental empty sella turcica (mean age, 36.1 vs 54.3 years, respectively;
p < 0.05); were more likely to report headache (93.3% vs 32.6%; p < 0.05) and visual complaints
(66.2% vs 28.3%; p < 0.05); showed greater mean scalp thickness (9.0 vs 6.4 mm; p < 0.05) and
neck soft-tissue thickness (19.5 vs 13.8 mm; p < 0.05); and were more likely to have an orbital find-
ing suggestive of IIH (93% vs 14%). Age modestly correlated with the width of the diaphragm
sella (r = 0.53) in the IIH group only.
CONCLUSION. The significance of the MRI finding of an empty sella turcica can be
determined using a combination of clinical and imaging findings.

Keywords: empty sella turcica, idiopathic intracranial
hypertension, MRI
DOI:10.2214/AJR.12.9013
Received April 3, 2012; accepted after revision
May 13, 2012.
Presented at the 2011 annual meeting of the American
Society of Neuroradiology, Seattle, WA.
1Department of Radiology and Imaging Sciences, Emory
University Hospital, 1364 Clifton Rd NE, BG22, Atlanta,
GA 30322. Address correspondence to
A. M. Saindane (asainda@emory.edu).
2 Department of Biostatistics & Bioinformatics, Emory
University School of Medicine, Atlanta, GA.
AJR 2013; 200:1125–1131
0361–803X/13/2005–1125
© American Roentgen Ray Society
T
he “empty” sella turcica is char-
acterized by intrasellar herniation
of suprasellar arachnoid and sub-
arachnoid space CSF, resulting in
flattening of the pituitary gland. Chronically
transmitted CSF pulsations from the herniat-
ed subarachnoid space often lead to bony ex-
pansion and remodeling of the sella turcica.
The empty sella turcica has been associated
with elevated intracranial pressure (ICP); pos-
teriorly placed optic chiasm; and a reduction
in pituitary gland volume due to menopause,
multiparity, pituitary gland infarction, diabe-
tes, or bromocriptine treatment [1–5]. In most
cases, however, the empty sella turcica is con-
sidered an incidental finding and is consid-
ered a normal variant related to a deficiency in
the diaphragm sella [6–9]. In the absence of
surgery, radiation therapy, or medical therapy
for an intrasellar tumor, this entity has been
termed a “primary empty sella turcica” [6].
Idiopathic intracranial hypertension (IIH),
also known as pseudotumor cerebri, is a syn-
drome of unknown cause that results in ele-
vated ICP without an intracranial mass le-
sion or hydrocephalus [10, 11]. Most patients
with IIH have headaches, tinnitus, diplopia,
and transient visual obscurations associat-
ed with papilledema [12, 13]. Symptoms can
be improved with a reduction of CSF pres-
sure through pharmacologic therapy or CSF
diversion procedures. If untreated, chroni-
cally elevated ICP may lead to permanent vi-
sion loss [14, 15]. Although the diagnosis of
IIH is based on clinical findings and elevated
CSF pressure on lumbar puncture in the set-
ting of “normal” neuroimaging findings [10,
11], orbital findings on CT and MRI are com-
monly seen in patients with IIH. These CT and
MRI findings include flattening of the poste-
rior sclera, distention of the perioptic nerve
subarachnoid CSF space, vertical tortuosity

AJR:200, May 2013 1125

 Saindane et al.

fat thickness (as a correlate of obesity, which

is common in IIH) [25, 26].

Materials and Methods

Patient Selection
Institutional review board approval was ob-
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved

tained, and informed patient consent was not re-

quired for the retrospective review of the medi-
cal records and imaging studies for this study. The
electronic medical records were searched from Jan-
uary 2008 through August 2010 for MRI reports
containing the terms “idiopathic intracranial hy-
pertension,” “pseudotumor cerebri,” “IIH,” or “be-
nign intracranial hypertension.” This search yield-
ed 71 patients; of those cases, 26 did not meet the
A B inclusion criteria of having both a clinical diagnosis
of IIH based on Dandy criteria [11] documented in
the medical record and sagittal T1-weighted imag-
es available for review. Thus, the remaining 45 pa-
tients were included in the IIH group.
A search of MRI reports from January 2008
through August 2010 for the terms “empty sella,”
“empty sella turcica,” “partially empty sella,” “par-
tially-empty sella,” “CSF filled sella,” “CSF-filled
sella,” “flattened pituitary,” “flattening of pituitary,”
“flattening of the pituitary,” “intrasellar arachnoid
cyst,” “sellar arachnoid cyst,” or “arachnoid cyst
of the sella” yielded 657 reports. We excluded pa-
tients without an otherwise normal MRI examina-
tion (any intracranial abnormality present), patients
with a history of pituitary surgery or a known diag-
nosis of a pituitary tumor, patients whose records
C D
lacked sagittal T1-weighted images for review, and
Fig. 1—Techniques for measuring sella turcica, infundibulum, optic chiasm, scalp fat, and neck fat on MRI.
patients who had a documented diagnosis of IIH or
A, Magnified midsagittal T1-weighted image through sellar region shows anteroposterior distance of
diaphragm sella (1), anteroposterior distance of infundibulum along diaphragm sella (2), and maximum who were under evaluation for IIH; this yielded 92
anteroposterior dimension of sella (3). patients for inclusion in the incidental empty sella
B, Magnified midsagittal T1-weighted image through sellar region shows maximum craniocaudal dimension of turcica group.
sella (4) and craniocaudal distance of optic chiasm above diaphragm sella (5).
C, Midsagittal contrast-enhanced T1-weighted image shows scalp fat at level of coronal suture (6). The indication for brain MRI in the group with
D, Midsagittal contrast-enhanced T1-weighted image shows neck fat at C2–C3 interspace level (7). incidental empty sella turcica was headaches (n =
17), metastatic workup for known non-CNS ma-
of the optic nerve sheath complex, and protru- 22], and the overall incidence of an empty sel- lignancy (n = 9), dizziness (n = 7), mental status
sion or enhancement of the prelaminar optic la turcica on imaging has been estimated at change (n = 7), seizure disorder (n = 7), dementia
nerve [16]. However, these imaging findings 12% [23]. In contrast, the incidence of IIH is (n = 6), paresthesias or numbness (n = 6), hearing
lack sufficient sensitivity or specificity to be relatively rare, estimated at approximately 1 loss (n = 5), ataxia (n = 3), tinnitus (n = 3), weak-
diagnostic of the cause of elevated ICP, such case per 100,000 [24] individuals. Therefore, ness (n = 3), tremor (n = 1), disorder of smell or
as IIH [17]. most patients showing an empty sella turci- taste (n = 1), and nystagmus (n = 1). There were
The empty sella turcica is the most com- ca on imaging will not have IIH and should seven patients with endocrinologic indications for
monly described imaging sign in the setting not require further diagnostic evaluation for the MRI examination: hyperprolactinemia (n = 4),
of IIH and presumably is an imaging corre- the condition. The purpose of this study was hyperglycemia (n = 2), and infertility (n = 1); how-
late of chronically elevated ICP [17–20]. On to determine whether a patient with an imag- ever, none of these patients had evidence of pitu-
MRI, the empty sella turcica is shown by ing finding of an empty sella turcica can be itary tumor on MRI. Nine patients with incidental
varying degrees of flattening of the superior confidently classified as a case of incidental empty sella turcica had visual symptoms that were
surface of the pituitary gland and by CSF-in- empty sella turcica or as a case of empty sella the primary indications for MRI: diplopia (n = 2),
tensity signal in the sellar confines and is of- turcica associated with IIH using a combina- blurry vision (n = 3), optic atrophy (n = 2), and
ten associated with enlargement and remod- tion of patient demographics; presenting clini- pseudopapilledema (n = 2). All nine of these pa-
eling of bony sella turcica. cal symptoms; measurements of the pituitary tients were referred by a neuroophthalmologist
An anatomic defect in the diaphragm sella gland, infundibulum, and sella turcica; orbital who had documented the absence of papilledema
has been shown in up to 50% of adults [21, findings; and measurements of subcutaneous on funduscopic examination.

1126 AJR:200, May 2013


Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved
A B
D E
MRI Technique and Data Analysis
MRI was performed at either 3 T (Trio, Sie-
mens Healthcare) or 1.5 T (Avanto, Siemens
Healthcare; or Signa, GE Healthcare) using a
standard head coil. Although the imaging proto-
col varied, all patients in both groups had sagittal
T1-weighted images for review, as well as axial
T1- and T2-weighted images. Sixty-four patients
in the incidental empty sella turcica group and 37
patients in the IIH group had contrast-enhanced
sequences after a standard dose (0.1 mmol/kg) of
IV contrast material (gadobenate dimeglumine
[MultiHance, Bracco Diagnostics]).
Images were reviewed in consensus by an expe-
rienced neuroradiologist and a neuroradiology fel-
low. The following measurements were recorded
on the midsagittal T1-weighted images (Fig. 1): es-
timated anterior-posterior (anteroposterior) length
of the diaphragm sella (in millimeters), maximum
anteroposterior dimension of the sella (in millime-
ters), maximum craniocaudal dimension of the sel-
la (in millimeters), anteroposterior distance from
the anterior diaphragm sella to the pituitary stalk
(in millimeters), and craniocaudal distance of the
optic chiasm from the diaphragm sella. Pituitary
AJR:200, May 2013 1127
MRI of Empty Sella Turcica
tissue height on the sagittal T1-weighted images
was classified into one of five categories using the
system of Yuh et al. [20]: I, normal; II, mild su-
perior concavity (less than one third of the height
of the sella); III, moderate concavity (between
one third and two thirds of the height of the sel-
la); IV, severe concavity (more than two thirds of
the height of the sella); and V, no visible pituitary
tissue. Examples of each of these categories are
depicted in Figure 2. Subcutaneous fat thickness
was measured orthogonal to the coronal suture and
posteriorly at the level of C2–C3.
Reviewers were blinded to the clinical indica-
tion for imaging and the diagnosis of IIH. They
evaluated each patient’s full set of images for the
following imaging findings: increased periop-
tic nerve CSF, flattening of the posterior sclera,
protrusion of the optic disc, and vertical tortuos-
ity of the intraorbital optic nerve. The electronic
medical records were reviewed and the presence
or absence of the following was recorded: head-
ache, visual symptoms (including transient visu-
al obscurations and double vision), clinical evi-
dence of papilledema or secondary optic atrophy,
and known pituitary-related endocrine dysfunc-
C
Fig. 2—Midsagittal T1-weighted images through
sella turcica show categories of pituitary tissue
height based on system proposed by Yuh et al. [20].
A, Category I, normal.
B, Category II, mild superior concavity (less than one
third height of sella).
C, Category III, moderate concavity (between one
third and two thirds of height of sella).
D, Category IV, severe concavity (more than two
thirds of height of sella).
E, Category V, no visible pituitary tissue.
tion (laboratory results indicating hypofunction
or hyperfunction).
Statistical Analysis
The chi-square or Fisher exact test was used to
compare the following characteristics of the IIH and
incidental empty sella turcica groups: sex; frequency
of presenting clinical symptoms of headache, visu-
al complaints, known papilledema, and known pitu-
itary-related endocrine dysfunction; and frequency
of orbital findings. Age and MRI measurements of
the sella and infundibulum were compared between
the two groups using a Student t test, and pituitary
grade was compared between the two groups using
the Wilcoxon rank sum test. MR measurements of
the sella, infundibulum, and scalp and neck fat thick-
nesses were correlated with patient age for the two
groups using a Pearson coefficient, and pituitary
grade was correlated with patient age for each group
using a Spearman coefficient.
Results
Nearly all patients in the IIH group (96%)
and all patients in the incidental empty sel-
la turcica group (by definition) showed some
 Saindane et al.

TABLE 1:  Differences in Clinical Presentation and MRI Findings Between Patients With Idiopathic Intracranial
Hypertension (IIH) and Patients With Incidental Empty Sella Turcica
Incidental Empty Sella
Turcica (n = 92) IIH (n = 45) pa
Demographic and clinical parameters
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved

Age (y), mean (SD) 54.3 (14.3) 36.1 (11.7) < 0.0001
Sex, % of female patients 84.8 97.8 0.02b
Presence of headache, % of patients 32.6 93.3 < 0.0001c
Presence of pituitary dysfunction, % of patients 2.2 2.2 1.0c
Presence of visual complaints, % of patients 28.3 66.2 0.0001b
Presence of papilledema or secondary optic atrophy, % of patients 0 100 < 0.0001
MRI measurements (mm)
Anteroposterior length of diaphragm sella, mean (SD) 11.77 (2.3) 11.82 (2.6) 0.91
Maximum anteroposterior dimension of sella, mean (SD) 13.48 (3.1) 14.02 (2.3) 0.30
Maximum craniocaudal dimension of sella, mean (SD) 10.72 (3.4) 10.62 (2.5) 0.86
Anteroposterior distance of infundibulum along diaphragm sella, mean (SD) 9.46 (2.0) 9.22 (2.1) 0.53
Craniocaudal distance of optic chiasm from diaphragm, mean (SD) 1.70 (1.3) 1.93 (1.2) 0.74
Pituitary grade, mean (range) IV (II–V) IV (I–V) < 0.0001d
Fat measurements (mm), mean (SD)
Scalp fat 6.35 (2.5) 9.00 (2.5) < 0.0001
Neck fat 13.83 (4.9) 19.53 (5.3) < 0.0001
MRI orbital findings, % of patients
Increased perioptic nerve CSF 13 93.3 < 0.0001c
Flattened posterior sclera 5.4 42.2 < 0.0001b
Protrusion of optic disc 0 22.2 < 0.0001c
Vertical tortuosity of optic nerve 0 4.4 0.11c
aStatistical analysis was performed using a Student t test except when noted otherwise. Boldface indicates difference between groups was statistically significant.
bChi-square test.
cFisher exact test.
d Wilcoxon rank sum test.

degree of empty sella turcica on MRI, as de-
fined by a pituitary grade of II, III, IV, or V.
Comparisons of age, sex, sellar measure-
ments, and scalp and neck fat measurements
and frequencies of orbital findings and clini-
cal symptoms of the two groups are summa-
rized in Table 1. The mean age of the IIH
group (36.1 years [SD, 11.7]; range, 18–61
years) was significantly lower (p < 0.0001)
than that of the incidental empty sella group
(54.3 years [SD, 14.3]; range, 22–84 years).
Both groups were predominantly female
(IIH vs empty sella, 44/45 [97.8%] vs 78/92
[84.8%]), but there was a significantly great-
er percentage of females in the incidental
empty sella turcica group (p = 0.02). IIH pa-
tients were significantly more likely to have
headache and visual complaints than the pa-
tients with incidental empty sella turcica. All
the patients in the IIH group had document-
ed papilledema, whereas review of the medi-
cal records did not reveal papilledema in any
of the patients in the incidental empty sel-
la turcica group. Both groups had extremely
low reported rates of pituitary-related endo-
crine dysfunction in the medical record that
did not differ significantly.
The infundibulum was reliably seen on sag-
ittal images of all patients, allowing the mea-
surements described. All of the MRI measure-
ments related to the sella and infundibulum
(estimated anteroposterior width of the dia-
phragm sella, maximum anteroposterior di-
mension of the sella, maximum craniocaudal
dimension of the sella, anteroposterior posi-
tion of the infundibulum relative to the dia-
phragm sella, and craniocaudal distance of
the optic chiasm from the diaphragm sella)
did not significantly differ between the two
groups. The pituitary grade, however, was sig-
nificantly higher in the incidental empty sella
turcica group than the IIH group (p < 0.0001).
Patients with IIH showed significantly greater
scalp fat thickness at the level of the coronal
suture and greater neck fat thickness than the
incidental empty sella turcica group (both, p
< 0.0001). Figure 3 depicts scalp and neck fat
measurements in representative patients from
both groups.
Correlations of the MRI measurements
with age for both groups are listed in Table 2.
The estimated anteroposterior dimension of
the diaphragm sella, maximum anteroposte-
rior dimension of the sella, maximum cranio-
caudal dimension of the sella, anteroposterior
position of the infundibulum relative to the
diaphragm sella, and pituitary grade showed
significant age-related increases in the IIH
group but not in the incidental empty sella
turcica group. Figure 4 shows a plot of the
anteroposterior dimension of the diaphragm
sella versus age in both groups. The optic chi-
asm height did not show an age-related corre-

1128 AJR:200, May 2013

 MRI of Empty Sella Turcica

Fig. 3—Scalp and neck fat measurements in patient

with incidental empty sella turcica and patient with
idiopathic intracranial hypertension (IIH).
A, Midsagittal T1-weighted image of 52-year-old
woman with incidental empty sella turcica (category
IV empty sella turcica) shows relatively little scalp
fat at level of coronal suture (short arrow) and at
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved

posterior neck at C2–C3 level (long arrow).

B, Midsagittal contrast-enhanced T1-weighted
image of 37-year-old woman with IIH (category
IV empty sella turcica) shows abundant scalp
fat at level of coronal suture (short arrow) and at
posterior neck at C2–C3 level (long arrow). Sella
turcica and pituitary gland appear similar in both
patients.

A B

TABLE 2:  Correlation of MRI Measurements With Age for Patients With Idiopathic Intracranial Hypertension (IIH)
and Patients With Incidental Empty Sella Turcica
Incidental Empty Sella Turcica (n = 92) IIH (n = 45)
MRI Measurements ra pb ra pb
Anteroposterior length of diaphragm sella 0.13 0.21 0.53 0.0002
Maximum anteroposterior dimension of sella −0.02 0.83 0.48 0.001
Maximum craniocaudal dimension of sella −0.08 0.43 0.36 0.02
Anteroposterior distance of infundibulum along diaphragm sella 0.12 0.27 0.42 0.004
Craniocaudal distance of optic chiasm from diaphragm 0.17 0.11 −0.16 0.31
Scalp fat −0.27 0.01 −0.08 0.59
Neck fat −0.15 0.16 0.02 0.88
Pituitary grade 0.19c 0.07 0.39c 0.008
aPearson correlation coefficients unless noted otherwise.
bBoldface indicates difference between groups was statistically significant.
cSpearman correlation coefficient.

lation in either group. Scalp fat exhibited an
age-related correlation (Pearson correlation
coefficient, –0.27; p = 0.01) in the incidental
empty sella turcica group but not in the IIH
group. There were no correlations between
neck fat thickness and age in either group.
Discussion
The “empty sella” is a term used to describe
a spectrum of findings related to the bony sella
turcica and pituitary gland, ranging from mild
superior concavity of the pituitary gland to ap-
parent absence of the gland and CSF expan-
sion of the bony confines of the sella turcica.
Milder appearances without bony expansion
and lesser degrees of pituitary compression are
often referred to as a “partially empty sella.”
Frequently during the course of interpreting
a brain MRI study, some degree of an empty
sella turcica is observed. Because the finding
may be seen incidentally (incidental empty
sella turcica) or pathologically as a manifesta-
tion of IIH, it would be helpful to advise clini-
cians about which patients should be evaluated
further for IIH, including funduscopic evalua-
tion and lumbar puncture with CSF pressure
measurement, and which patients might not
need further workup for elevated ICP.
In this study, we evaluated patients with
an MRI finding of an empty sella turcica
but no known diagnosis of IIH (i.e., inciden-
tal empty sella turcica) and patients with an
MRI finding of an empty sella turcica and a
clinical diagnosis of IIH. Not surprisingly,
the most important features that suggested
the diagnosis of IIH were clinical findings in-
cluding younger patient age; increased scalp
thickness and neck fat thickness; and pres-
ence of headache, visual symptoms, papill-
edema, and orbital findings suggestive of IIH.
Although none of these factors alone can dis-
tinguish between patients with IIH and those
with incidental empty sella turcica, taking
into account the presence or absence of each
of these features in the setting of an empty
sella turcica can improve confidence in pro-
posing the diagnosis and workup for IIH or
dismissing the finding as an incidental empty
sella turcica.
In this study, no difference was found in
the measurements of the sella between the
IIH patients and the incidental empty sella
turcica patients. This finding suggests that
the effect of elevated ICP on the sella tur-
cica and pituitary gland in IIH is not unique
and that the appearance itself is nonspecif-
ic. Varying appearances of empty sella tur-
cica have been described with IIH [20]. A
scatterplot of the size of the diaphragm sella
versus age (Fig. 3) shows that the diaphragm
sella may actually widen more over time in
the IIH group as a result of the elevated ICP,
whereas the lack of a correlation in the in-
cidental empty sella turcica group suggests
that size is a function of a preexisting defect
in the diaphragm sella that may not increase

AJR:200, May 2013 1129


20.0
IIH
19.0
Incidental empty sella turcica
Anteroposterior Width of Diaphragm Sella (mm)
18.0
17.0
16.0
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved
15.0
14.0
13.0
12.0
11.0
10.0
9.0
8.0
7.0
6.0
5.0
15.0
Fig. 4—Scatterplot shows age-related correlation of anteroposterior width of diaphragm sella in patients
with idiopathic intracranial hypertension (IIH) and patients with incidental empty sella turcica. Solid black
line depicts linear trend line for IIH group (r = 0.53; p = 0.0002), whereas dotted line represents linear trend
line for incidental empty sella turcica group (r = 0.13; p = 0.21 [not significant]).
significantly or that may increase more slow-
ly over the age range listed under conditions
of normal ICP. The significant difference in
pituitary grade between the two groups may
be related to this age dependence in the IIH
group, because these patients were signifi-
cantly younger than patients in the inciden-
tal empty sella turcica group. The absence of
significant differences in any of the specific
measurements of the sella argues that the ex-
tent of the empty sella should not be used to
decide whether or not the patient is sympto-
matic from the finding.
The IIH group was more likely than the in-
cidental empty sella turcica group to have one
or more orbital finding of increased perioptic
nerve CSF, flattening of the posterior sclera,
protrusion of the optic disc, and vertical tortu-
osity of the intraorbital optic nerve. These or-
bital findings have been significantly associ-
ated with IIH [19, 27]; however, according to
Agid and Farb [17] and Agid et al. [28] et al.,
posterior globe flattening is the only sign that
strongly suggests the diagnosis of IIH (speci-
ficity, 100%; sensitivity, 43.5%; positive like-
lihood ratio, 49.7). Because patients with an
incidental empty sella turcica do not have el-
evated ICP, MRI correlates of papilledema
should not be present. The most common or-
bital finding in the incidental empty sella tur-
cica group was increased CSF surrounding the
optic nerve intraorbital segment. The results
of this study show, as previously described in
1130 AJR:200, May 2013
Saindane et al.
25.0 35.0 45.0 55.0 65.0 75.0 85.0
Age (y)
the literature [17, 28], that this finding is a rela-
tively nonspecific sign for chronically elevat-
ed ICP. Rohr et al. [29] found that a combi-
nation of at least two imaging signs including
optic nerve sheath hydrops, reduction in pitu-
itary height, and venous outflow obstruction
discriminated between patients with elevated
ICP from a variety of causes and age-matched
control subjects.
Scalp and neck subcutaneous fat thick-
nesses were significantly greater in the IIH
group than in the incidental empty sella tur-
cica group. These findings have not been pre-
viously reported, but as a potential crude im-
aging marker of body mass index (BMI),
subcutaneous fat thickness would be expected
to be associated with IIH because IIH is a dis-
ease predominantly of obese patients [30–33].
Normal BMI and age greater than 50 years
are rare or atypical for patients with IIH, and
this atypical subset of patients is more likely
to have visual complaints [33]. Both groups
were predominantly female, as has been pre-
viously described [3]; however, the incidental
empty sella turcica group did have a signifi-
cantly higher percentage of males.
Patients in the IIH group were more like-
ly to present with headache than those in the
incidental empty sella turcica group. Previ-
ous studies have shown that the most com-
mon presenting symptom for IIH is head-
ache, occurring in more than 90% of cases
in most series [11, 26]. Although headache is
common in the general population and is a
frequent indication for brain imaging, most
patients with incidental empty sella turcica
did not have the symptom of headache de-
scribed in their clinical records or as an in-
dication for imaging. The patients with IIH
were more likely to present with visual symp-
toms than the patients with incidental empty
sella. Previous studies have shown that visual
symptoms including transient visual obscura-
tions, blurred vision, photophobia, and dou-
ble vision occur in 40–70% of patients with
IIH [26]. These symptoms should not occur
in patients with incidental empty sella turci-
ca. Finally, pituitary symptoms and laborato-
ry evidence of pituitary hypofunction or hy-
perfunction have been described in patients
with incidental empty sella turcica [34]. It is
possible that there were subclinical or labora-
tory abnormalities in these patients that were
not mentioned in the medical records.
There are several limitations to this study. A
major limitation is that the clinical symptoms
and diagnosis were based on retrospective re-
view of electronic records. It is possible that
some of the cases of incidental empty sella tur-
cica were not appropriately diagnosed as IIH.
Although funduscopic examination was not
performed in all of the incidental empty sella
turcica patients, it was performed in the nine
patients presenting with visual complaints as
the indication for MRI and did not show papill-
edema. Most of the remaining patients did not
have headache or other typical symptoms sug-
gestive of IIH, again making it less likely that
these patients had subclinically elevated ICP.
Second, there is likely some inaccuracy in
the measurements of the diaphragm sella and
sella because the diaphragm sella is not al-
ways clearly visible but must be inferred and
because differences in slice position for sagittal
images could affect the size of the sella turci-
ca. These errors would, however, be systematic
errors that would not be expected to differen-
tially affect one group over the other. Finally,
there was variability in the imaging protocols
used, and it is possible that patients presenting
with visual complaints were more likely to have
dedicated orbital images on which orbital find-
ings were easier to detect and that the orbital
findings were underestimated in the incidental
empty sella turcica group on that basis.
The imaging findings related to pituitary
compression and bony sellar expansion in an
empty sella turcica alone are nonspecific. Us-
ing a combination of the available clinical in-
formation of patient age and sex, headache,
visual symptoms, and papilledema and imag-

ing correlates of body fat thickness and orbital
findings suggestive of IIH, the imaging find-
ing of an empty sella turcica can in most cas-
es be confidently triaged to further workup as
suspicious for IIH or can be recognized as an
incidental finding.
Downloaded from www.ajronline.org by 181.175.138.180 on 08/21/18 from IP address 181.175.138.180. Copyright ARRS. For personal use only; all rights reserved
References
1. Jaffer KA, Obbens EA, El Gammal TA. “Empty”
sella: review of 76 cases. South Med J 1979;
72:294–296
2. Buonaguidi R, Faggionato F, Puglioli M, Lutzem-
berger L, Ferdeghini M. Primary empty sella syn-
drome: clinical and radiological study, and evalu-
ation of pituitary function (in Italian). Minerva
Med 1984; 75:1589–1599
3. Jordan RM, Kendall JW, Kerber CW. The prima-
ry empty sella syndrome: analysis of the clinical
characteristics, radiographic features, pituitary
function and cerebrospinal fluid adenohypophysi-
al hormone concentrations. Am J Med 1977;
62:569–580
4. Zayour DH, Azar ST. Silent pituitary infarction
after coronary artery bypass grafting procedure:
case report and review of literature. Endocr Pract
2006; 12:59–62
5. Demura R, Kubo O, Demura H, Shizume K, Kita-
mura K. Changes in computed tomographic find-
ings in microprolactinomas before and after bro-
mocriptine. Acta Endocrinol (Copenh) 1985;
110:308–312
6. Sage MR, Blumbergs PC. Primary empty sella
turcica: a radiological-anatomical correlation.
Australas Radiol 2000; 44:341–348
7. Estopiñán V, Fortea L, Gracia P, Baiges JJ. Pri-
mary empty sella turcica: clinical aspects and
hormonal study of 15 cases (in Spanish). An Med
Interna 1991; 8:537–541
8. Akiyama Y, Yamasaki T, Kagawa T, Moritake K.
Empty sella syndrome (in Japanese). Nippon
Rinsho 1993; 51:2731–2736
9. Borgoni F, Giusti F, Venturini M, Furbatto A, Bi-
agi P. Primary empty sella syndrome: report of 6
cases (in Italian). Minerva Med 1993; 84:233–238
10. Dandy WE. Intracranial pressure without brain
AJR:200, May 2013 1131
MRI of Empty Sella Turcica
tumor: diagnosis and treatment. Ann Surg 1937;
106:492–513
11. Friedman DI, Jacobson DM. Diagnostic criteria
for idiopathic intracranial hypertension. Neurolo-
gy 2002; 59:1492–1495
12. Giuseffi V, Wall M, Siegel PZ, Rojas PB. Symp-
toms and disease associations in idiopathic intra-
cranial hypertension (pseudotumor cerebri): a
case-control study. Neurology 1991; 41:239–244
13. Wall M. Idiopathic intracranial hypertension.
Neurol Clin 2010; 28:593–617
14. Corbett JJ, Savino PJ, Thompson HS, et al. Visual
loss in pseudotumor cerebri: follow-up of 57 pa-
tients from five to 41 years and a profile of 14 pa-
tients with permanent severe visual loss. Arch
Neurol 1982; 39:461–474
15. Celebisoy N, Gokcay F, Sirin H, Akyurekli O.
Treatment of idiopathic intracranial hypertension:
topiramate vs acetazolamide, an open-label study.
Acta Neurol Scand 2007; 116:322–327
16. Brodsky MC, Vaphiades M. Magnetic resonance
imaging in pseudotumor cerebri. Ophthalmology
1998; 105:1686–1693
17. Agid R, Farb RI. Neuroimaging in the diagnosis
of idiopathic intracranial hypertension. Minerva
Med 2006; 97:365–370
18. Davis S, Tress B, King J. Primary empty sella
syndrome and benign intracranial hypertension.
Clin Exp Neurol 1978; 15:248–257
19. Gibby WA, Cohen MS, Goldberg HI, Sergott RC.
Pseudotumor cerebri: CT findings and correlation
with vision loss. AJR 1993; 160:143–146
20. Yuh WT, Zhu M, Taoka T, et al. MR imaging of pi-
tuitary morphology in idiopathic intracranial hyper-
tension. J Magn Reson Imaging 2000; 12:808–813
21. Sage MR, Blumbergs PC, Fowler GW. The dia-
phragma sellae: its relationship to normal sellar
variations in frontal radiographic projections. Ra-
diology 1982; 145:699–701
22. Sage MR, Blumbergs PC, Mulligan BP, Fowler
GW. The diaphragma sellae: its relationship to the
configuration of the pituitary gland. Radiology
1982; 145:703–708
23. Foresti M, Guidali A, Susanna P. Primary empty
sella: incidence in 500 asymptomatic subjects ex-
amined with magnetic resonance (in Italian). Ra-
diol Med 1991; 81:803–807
24. Durcan FJ, Corbett JJ, Wall M. The incidence of
pseudotumor cerebri: population studies in Iowa
and Louisiana. Arch Neurol 1988; 45:875–877
25. Radhakrishnan K, Thacker AK, Bohlaga NH,
Maloo JC, Gerryo SE. Epidemiology of idiopathic
intracranial hypertension: a prospective and case-
control study. J Neurol Sci 1993; 116:18–28
26. Galvin JA, Van Stavern GP. Clinical characteriza-
tion of idiopathic intracranial hypertension at the
Detroit Medical Center. J Neurol Sci 2004;
223:157–160
27. Ball AK, Howman A, Wheatley K, et al. A ran-
domised controlled trial of treatment for idiopath-
ic intracranial hypertension. J Neurol 2011;
258:874–881
28. Agid R, Farb RI, Willinsky RA, Mikulis DJ,
Tomlinson G. Idiopathic intracranial hyperten-
sion: the validity of cross-sectional neuroimaging
signs. Neuroradiology 2006; 48:521–527
29. Rohr AC, Riedel C, Fruehauf MC, et al. MR im-
aging findings in patients with secondary intra-
cranial hypertension. AJNR 2011; 32:1021–1029
30. Friedman DI, Rausch EA. Headache diagnoses in
patients with treated idiopathic intracranial hy-
pertension. Neurology 2002; 58:1551–1553
31. Verdin E, Smitz S, Thibaut A, Born J, Legros JJ,
Luyckx A. Adipsic hypernatremia in a patient
with pseudotumor cerebri and the primary empty
sella syndrome. J Endocrinol Invest 1985;
8:369–372
32. Silver RI, Moonis G, Schlosser RJ, Bolger WE,
Loevner LA. Radiographic signs of elevated in-
tracranial pressure in idiopathic cerebrospinal
fluid leaks: a possible presentation of idiopathic
intracranial hypertension. Am J Rhinol 2007;
21:257–261
33. Bruce BB, Kedar S, Van Stavern GP, Corbett JJ,
Newman NJ, Biousse V. Atypical idiopathic intra-
cranial hypertension: normal BMI and older pa-
tients. Neurology 2010; 74:1827–1832
34. De Marinis L, Bonadonna S, Bianchi A, Maira G,
Giustina A. Primary empty sella. J Clin Endocri-
nol Metab 2005; 90:5471–5477