239 / 280
abscess
[5]
. As
this
study did not
include
abscesses
from
all of
the
lateral
neck,
it
is
difficult
to
directly
compare
these findings with
the ultrasounds
studies. A
similar
study of 16 CT
scans of deep and
lateral
neck
abscesses
demonstrated
a
sensitivity
of
91%
but
a
specificity
of
60%
[6]
.
This
study was
limited
by
its
small
sample
size
and mixed
pathology.
A comparison of ultrasound and CT
in
the diagnosis of pediatric
lateral neck
abscesses
is necessary
in
order
to
establish
a practice
guideline
for
this
population.
CT
is
used
more
often
at
this
institution, most
likely because
there
is a CT
technician available at
all
times whereas ultrasound readings are only available during the
day.
If
it
can
be
shown
that
ultrasound
and
CT
are
comparable
in
accuracy
for
the diagnosis of
lateral neck abscesses,
then a practice
guideline can be developed based upon
the cost and safety profiles
of
the
two
procedures.
This
study
compares
the
accuracy
of
ultrasound
and
CT
to
the
gold
standard
outcome
of
attempted
drainage
in
order
to
promote
judicious
and
individualized
use
of
ultrasound
and
CT
in
the
diagnosis
of
children with
lateral
neck
abscesses.
2. Materials
and methods
IRB
approval was
granted
by
the University
of Oklahoma
for
a
retrospective
study
of
all
children
0–18
years
of
age with
lateral
neck
abscesses
who
underwent
preoperative
imaging
prior
to
attempted drainage at
an academic
tertiary
care
center
from 2005
to
2011.
This
allowed
evaluation
of
the
accuracy
of
CT
and
ultrasound relative
to
the surgical finding of presence or absence of
pus,
the
gold
standard
for
determination
of
an
abscess.
To
appropriately
power
the
study,
it
is
necessary
to
compare
approximately
40
ultrasound
and
40
CT
imaging
studies.
This
goal was
recommended
by
our
statistician
and
is
consistent with
the
power
of
previous
evaluations
of
CT
or
ultrasound.
Subjects
were
located by
a
search of
the medical
center billing database by
Current
Procedural
Terminology
codes.
The
patients were
evalu-
ated
in
two
groups
based
on whether
an
ultrasound
or
a
CT was
performed
prior
to
surgery.
There
are
no
specific
preferences
besides
availability
that
determined which
children
received
an
ultrasound or a CT
in
this
study. Currently
there
is no
institutional
protocol; rather
the decision
is
influenced by when a child presents
and
which
physician
initially
sees
the
patient. When
clinically
indicated,
some
children
are
taken
to
the operating
room without
receiving
either
imaging
study.
The majority
of
ultrasounds
and
CT
scans were
performed
at
this
institution
and
some
were
performed
at
outside
medical
facilities
prior
to
transfer.
All
imaging
studies
were
read
at
the
same academic
tertiary care center. The
initial final report
from
the
department
of
radiology
was
used,
and
only
studies
performed
within
3
days
of
surgery
are
included.
All
incision
and
drainages
were performed at this
facility.
In accordance with this
institution’s
protocol, stable children presenting with suspected neck abscesses
are
given
48 h
of
intravenous
clindamycin
before
drainage
is
attempted.
Lateral neck abscess
locations
in
this
study
include
the
anterior
and
posterior
triangles,
submandibular,
submental,
parotid,
and
parapharyngeal
spaces.
Clinical
and demographic
characteristics were
summarized
for
each
group
of
subjects
(CT
or
ultrasound).
The
mean
age
was
compared
between
groups
using
a
2-sample
t
-test.
The
distribu-
tion
of
gender was
compared
between
groups
using
a
Chi-square
test. Demographic
characteristics of
the CT and ultrasound groups
were
compared
after
excluding
patients who
underwent
both
CT
and ultrasound screening
tests. The sensitivity, specificity, positive
predictive value
and negative predictive value were
calculated
for
each method
separately.
The
accuracy
of
the
imaging method
is
summarized
using
a
two-sided
95%
exact
confidence
interval.
Positive and negative predictive values were calculated assuming a
positive
abscess
prevalence
of
0.90.
3. Results
One
hundred
thirty-two
patients
are
included
in
the
analysis
with 31 who underwent ultrasound, 93 who underwent CT,
and 8
who underwent both ultrasound
and CT
scans. The median
age
of
the
sample was
1.5
years
(range
one month
to
18
years) with
a
mean
age
of
2.9
years
(standard
deviation
3.5
years).
Although
those who
underwent
a
CT
scan were
on
average
one
year
older
than
the
ultrasound
group,
this
difference
was
not
statistically
significant. The gender distribution was well balanced between
the
groups
(
Table
1
).
A
total
of
140
imaging
studies
were
available
for
review,
including
39
ultrasound
studies with
gold
standard
results
of
34
positive
and
5
negative
and
101
CT
studies
with
gold
standard
results
of 90 positive
and 11 negative.
The
overall prevalence
of
a
pus-positive
abscess
in
children
undergoing
the
gold
standard,
incision
and
drainage, was
89%.
Table
2
presents
the
estimated
sensitivity,
specificity,
positive
predictive
value
and
negative
predictive
value
for
each
method
along
with
a
95%
confidence
interval
for
the
estimate.
The CT scan test method has very
low specificity
(2/11, 18%) and
a
very
low
negative
predictive
value
(6%)
assuming
a
positive
abscess
prevalence
of
0.9.
The
sensitivity
is
reasonable
(61/90,
68%). The positive predictive value
(88%)
is
slightly
lower
than
the
assumed
prevalence
of
90%.
Based
on
the
assumed
prevalence
value,
the
probability
of
a
pus-positive
abscess
is
90%
(without
knowledge of
the CT
test
result) while
the positive predictive value
suggests
that
the
probability
of
a
pus-positive
abscess
is
88%
among
those
with
a
positive
CT
scan.
Similarly,
the
estimated
negative predictive value
(6%)
is
less
than
the assumed prevalence
of
a pus-negative abscess
(10%). The ultrasound
test method has
a
high estimated
specificity
(5/5, 100%) but a
low
sensitivity
(18/34,
53%). The positive predictive value
(96%)
is high while
the negative
predictive
value
is
low
(16%)
assuming
a
positive
abscess
prevalence
of
0.9.
Table
3
demonstrates
the
sensitivity
and
specificity
of
ultra-
sound
and
CT
by
location
of
the
abscess.
Twelve
of
140
imaging
studies were
excluded
from
this
analysis
because
they
included
Table
1
Demographic
characteristics
of
patients who
underwent
ultrasound
and
computed
tomography.
Baseline
values
a
All
patients
b
(
n
= 132)
Diagnostic
screening
approach
p
-Value
c
(comparing
CT
to
ultrasound)
CT
(
n
= 101)
Ultrasound
(
n
= 39)
Age
in
years
2.9
(3.5)
3.2
(3.5)
2.2
(2.9)
p
= 0.16
1.5
[0.04–18]
2.0
[0.08–18]
1.3
[0.04–16]
Male
sex
69
(52%)
55
(54%)
19
(49%)
p
= 0.53
Legend:
CT
–
computed
tomography;
n
–
total
number
in
category.
a
Distributions
summarized
using
the mean
(standard
deviation)
and median
[range]
for
continuous measures
and
count
(column
%)
for
categorical measures.
b
Data
are
available
for
132
patients,
eight
of whom
underwent
both
CT
and
ultrasound
screening.
c
Statistical
comparisons
of
the mean
or
proportions were made
after
excluding
eight
patients
undergoing
both
CT
and
ultrasound
screening.
B.
Collins
et al.
/
International
Journal
of
Pediatric Otorhinolaryngology
78
(2014)
423–426
217