Efficacy of diffusion-weighted magnetic resonance imaging in the middle ear cholesteatoma
Mikio Suzuki, MD; Akira Ganaha, MD; Asanori Kyuuna, MD; Hiroyuki Maeda, MD; Tetsuhiro Miyara, MD
University of the Ryukyus
ABSTRACT
INTRODUCTION
Objective: This study evaluated
the usefulness of diffusionweighted magnetic resonance
imaging (DWI) in the diagnosis
of middle ear cholesteatoma.
Methods: We performed DWI
(spin-echo-type EPI) on 73
patients suspected of having
middle ear cholesteatoma,
including 21 revision cases.
Results: The sensitivity,
specificity, and positive and
negative predictive values of
DWI for cholesteatoma were
69.4%, 92.8%, 97.5%, and
41.9%, respectively. Of the 21
subjects who received revision
surgery, the sensitivity,
specificity, and positive and
negative predictive values of
DWI for residual or recurrent
acquired cholesteatoma were
71.4%, 100%, 100%, and
63.6%, respectively.
Cholesteatoma mass
diameters were less than 5 mm
in 10 out of 18 subjects with
both cholesteatoma and
negative DWI findings.
Conclusions: Since DWI clearly
showed high specificity and
positive predictive value, it is
useful for diagnosing middle
ear cholesteatoma, including
postoperative recurrent
cholesteatoma of 5 mm
diameter or larger.
Radiological examinations, particularly computed
tomography (CT) and magnetic resonance imaging
(MRI), have become important in the evaluation of the
presence and extent of cholesteatoma.
Cholesteatoma generally appears as an isointense
lesion compared with brain on T1-weighted images
and a hyperintense lesion on T2-weighted images, and
without enhancement in delayed post-contrast T1
sequences. Recently, several studies have shown the
importance of diffusion-weighted magnetic resonance
imaging (DWI) in the diagnosis of cholesteatoma1.
Cholesteatoma exhibits a high signal intensity on a
diffusion-weighted image obtained with a b factor of
800 or 1000 sec/mm2. However, there are some major
limitations with standard clinical DWI sequences, such
as low resolution and the production of relatively thick
image sections that have distorted shapes due to
artifacts at the skull base.
The aim of the present prospective study was to
evaluate the usefulness and limitations of clinical DWI
in detecting middle ear cholesteatoma, including
residual cholesteatoma.
CONTACT
Mikio Suzuki
Dept. Otorhinolaryngology, Head
and Neck Surgery, University of
the Ryukyus
Email: [email protected]
Phone: +81-98-895-1183
Website:
http://www.ent-ryukyu.jp/
Poster Design & Printing by Genigraphics - 800.790.4001
®
RESULTS
1) Patient profiles and DWI findings
The clinical features of patients participating in
the present study are summarized in Table 1. The
correlation between DWI preoperative and
operative findings is summarized in Table 2. A
correct diagnosis using DWI was made in 54
(73.9%) of the 73 subjects examined, that is, 41
subjects were true-positive (TP) and 13 subjects
were true-negative (TN). DWI, however, failed to
detect the presence or absence of cholesteatoma
in 19 (26.0%) of the 73 subjects, that is, 18
subjects showed false-negative results and one
showed a false-positive result. Thus, the sensitivity,
specificity, and positive and negative predictive
values of DWI for cholesteatoma were 69.4%,
92.8%, 97.5% and 41.9%, respectively.
2) Correlation between type and size of
cholesteatoma and negative DWI findings
Patients were divided into two groups: a less than
5 mm group and an equal to or more than 5 mm
group, based on the size from entry to the bottom
of each cholesteatoma that was measured using
an excavator during otologic surgery. In the case of
a cholesteatoma mass equal to or larger than 5
mm, 37 (82.2%) of 45 subjects were found to be
DWI positive.
METHODS AND MATERIALS
DWI examination was performed on 73 patients
suspected of having middle ear cholesteatoma
according to medical history, otoscopic examination,
audiological examinations, and CT scan. Patients were
examined using spin-echo-type echo planar imaging
(SE-EPI) pulse sequences and imaging parameters of
1.5 T MR machine.
A radiologist evaluated all MR images without
knowledge of the otologic examinations or surgical
results. A positive finding of cholesteatoma was
defined as a markedly high DWI signal intensity of the
affected tissue compared with brain tissue. All cases
were classified as positive or negative according to the
above-mentioned signal characteristics. Each
operation was carried out within three months after the
DWI examination. We investigated the types and sizes
of cholesteatoma using intraoperative findings, and
clarified the sensitivity, specificity, and positive and
negative predictive values of preoperative DWI
examination regarding cholesteatoma.
RESULTS
DISCUSSION
In contrast, in cases of cholesteatoma masses less
than 5 mm, a significantly lower percentage of
subjects were DWI positive (Table 2; 4/14, 28.5%;
p<0.001, chi-square test).
3) Detection of residual or recurrent acquired
cholesteatoma in revision cases
In this study, revision operations were carried out
on 21 subjects, including nine with planned stage
surgeries. Of these 21 subjects, six had residual
cholesteatoma, six had recurrent acquired
cholesteatoma, two had both residual and recurrent
acquired cholesteatoma, and seven had no
cholesteatoma. The correlation between DWI
preoperative and operative findings in revision cases
is summarized in Table 3. Of the 10 subjects who
showed positive DWI findings, all (100%) had middle
ear cholesteatoma. Of the 11 subjects who showed
negative DWI findings, four (36.3%) had
cholesteatoma. Thus, the sensitivity, specificity, and
positive and negative predictive values of DWI for
residual or recurrent acquired cholesteatoma were
71.4%, 100%, 100%, and 63.6%, respectively.
Postoperative diagnosis
Otoscopic evaluation
(%)
CT evaluation
(%)
# of
subjects
presence of cholesteatoma
primary acquired
adhesion
59
41
suspicious 9 (75%)
suspicious 8
12
(67%)
attic
suspicious 25 (96%)
suspicious 22
(85%)
26
unclassified
suspicious 1 (33%)
suspicious 1
3
(33%)
residual/recurrent
congenital
suspicious 8 (61%)
suspicious 7
13
suspicious 2 (50%)
(53%)
suspicious 4
4
(100%)
iatrogenic
absence of cholesteatoma
suspicious 1 (100%)
suspicious 0 (0%)
1
suspicious 1 (7%)
suspicious 3
14
(21%)
Figure 1. Representative cases of middle ear cholesteatoma.
Positive on DWI
Negative on DWI
(n=42)
(n=31)
presence of cholesteatoma
true-positive
false-negative
(n=59)
(n=41)
(n=18)
size 5 (n=37)
size 5 (n=8)
size <5 (n=4)
size <5 (n=10)
absence of cholesteatoma
false-positive
true-negative
(n=14)
(n=1)
(n=13)
Table 1. Summary of patients.
Positive on DWI
Negative on DWI
(n=10)
(n=11)
presence of cholesteatoma
true-positive
false-negative
(n=14)
(n=10)
(n=4)
size 5 mm
(n=10)
size <5 mm
(n=4)
Absence of cholesteatoma
false-positive
true-negative
(n=7)
(n=0)
(n=7)
DWI, diffusion-weighted magnetic resonance imaging
Table 2. Correlation between DWI and oprative findings.
Several studies have reported that MRI without DWI
after primary surgery showed a specificity range of 63–
71% and a positive predictive value within 50–78%.2-4
In this study, DWI was shown to have a relatively high
specificity and positive predictive value in detecting
cholesteatoma, and it can also provide more
information on cholesteatoma to supplement findings
obtained by conventional MRI examination.
In this study, there were 18 false-negative subjects.
In 10 of these, the diameter of the cholesteatoma
mass was less than 5 mm. The remaining eight
subjects had an adhesion-type cholesteatoma or
partially evacuated attic cholesteatoma with a small
amount of keratin accumulation. In the case of a
cholesteatoma mass with a diameter of at least 5 mm,
the rate of correct diagnosis was 82.2% (n=45). The
rate of correct diagnosis was only 28.5% in subjects
with a cholesteatoma mass diameter less than this.
However, even with larger diameters, the volume of
accumulated keratin considerably affects the detection
rate of adhesion-type or evacuated cholesteatoma
masses.
Postoperative changes such as the presence of
granulation tissue, cholesterol granuloma, and other
nonspecific tissues occur in the tympanic cavity and
mastoid cavity, which can make it difficult to diagnose
postoperative recurrent cholesteatoma (especially
residual cholesteatoma) based on the results of
ordinary otologic examinations. In the present study,
all subjects who received revision surgery were also
SE-EPI DWI positive in the cases of cholesteatoma
masses of at least 5 mm diameter. Thus, DWI is useful
in the diagnosis of 5-mm-diameter or larger middle ear
residual or recurrent acquired cholesteatoma masses,
as well as in cholesteatoma without previous surgery.
Table 3. Summary of revision cases.
REFERENCES
1. Dubrulle F, et al. Radiology 2006; 238: 604-610.
2. Vanden Abeele D, et al. Acta Otolaryngol 1999; 119:
555-561.
3. Kimitsuki T et al. ORL J Otorhinolaryngol Relat Spec
2001; 63: 291-293.
4. Denoyelle F et al. Ann Otolaryngol Chir Cervicofac
1994; 111: 85-88.