Diagnosis

Published on July 2016 | Categories: Documents | Downloads: 50 | Comments: 0 | Views: 386
of 9
Download PDF   Embed   Report

diagnosis

Comments

Content



AJ R:176, April 2001

933

Comparative Assessment of CT
and Sonographic Techniques for
Appendiceal Imaging

OBJECTIVE.

We performed a comparative assessment of CT and sonographic techniques
used to assess appendicitis.

MATERIALS AND METHODS.



One hundred patients with clinically suspected acute ap-
pendicitis were examined with sonography, unenhanced focused appendiceal CT, complete abdomi-
nopelvic CT using IV contrast material, focused appendiceal CT with colonic contrast material, and
repeated sonography with colonic contrast material. Each sonogram was videotaped for subsequent
interpretation by three radiologists and two sonographers. The mean sensitivity, specificity, positive
and negative predictive values, inter- and intraobserver variability, and diagnostic confidence scores
of all observers were used for comparative performance assessments. The three CT examinations
were filmed and interpreted separately by four radiologists. Patient discomfort was assessed on a 10-
point scale for each radiologic study. Diagnoses were confirmed by pathologic evaluation of resected
appendixes or clinical follow-up for a minimum of 3 months after presentation.

RESULTS.

Twenty-four of the 100 patients had positive findings for acute appendicitis. Both
sonographic techniques had high specificity (85–89%) and comparable accuracy (73–75%) but low
sensitivity (33–35%) and inter- and intraobserver variability (

κ

= 0.15–0.20 and 0.39–0.42, respec-
tively). Unenhanced focused appendiceal CT, abdominopelvic CT, and focused appendiceal CT with
colonic contrast material all significantly outperformed sonography (

p

< 0.0001), with sensitivities of
78%, 72%, and 80%; specificities of 86%, 91%, and 87%; and accuracies of 84%, 87%, and 85%, re-
spectively. Abdominopelvic CT gave the greatest confidence in cases with negative findings (

p

=
0.001), and focused appendiceal CT with colonic contrast material gave the greatest confidence for
cases with positive findings (

p

= 0.02). In terms of inter- and intraobserver variability, focused appen-
diceal CT with colonic contrast material yielded the highest, and unenhanced focused appendiceal
CT the lowest, agreement (interobserver

κ =

0.45 vs. 0.36 and intraobserver

κ =

0.85 vs. 0.76, respec-
tively) (

p

< 0.05). Colonic contrast material was unsuccessfully advanced into the cecum in 18% of
patients and leaked in another 24%. Patient discomfort was greatest with focused appendiceal CT us-
ing colonic contrast material and least with unenhanced focused appendiceal CT (

p

< 0.05).

CONCLUSION.

A standard abdominopelvic CT scan is recommended as the initial ex-
amination for appendicitis in adult patients. However, focused appendiceal CT with colonic
contrast material material should be used as a problem-solving technique in difficult cases.
any investigations have clearly
shown the usefulness of radiologic
imaging for examining patients who
are clinically suspected of having appendicitis.
CT and sonography are the modalities most fre-
quently recommended because of their ability to
lower the clinical negative appendectomy rate,
which ranges from 6.5% to 45% [1–4]. Further-
more, at least one study has shown that a prop-
erly performed and accurately interpreted
imaging study of relatively low cost can signifi-
cantly influence clinical management and be
cost-effective [5]. However, consensus in the lit-
erature is lacking for the specific imaging tech-
nique to be used in the evaluation of right lower
abdominal quadrant pain, and the choice of tech-
nique for CT has been particularly controversial.
At least three approaches have been advocated.
Some investigators have suggested an unen-
hanced technique in which thin-collimation (5
mm) CT is performed through the right lower
abdominal quadrant with or without a general
survey of the abdomen and pelvis [6, 7]. Other
investigators have advocated using oral and IV
contrast media and performing a similar CT as-
sessment [8, 9]. Finally, Rao et al. [10, 11] have
reported impressive results with focused appen-
diceal CT using rectally administered contrast

Scott W. Wise

1,2

Mark R. Labuski

1

Claudia J . Kasales

1

J udy S. Blebea

1

J on W. Meilstrup

1

Gregory P. Holley

1

Salvatore A. LaRusso

1

J ames Holliman

3

Francesca M. Ruggiero

4

David Mauger

5

Received August 18, 2000; accepted after revision
October 2, 2000.

1

Department of Radiology, Pennsylvania State University
College of Medicine, Milton S. Hershey Medical Center,
500 University Dr., Hershey, PA 17033.

2

116 Racehorse Dr., J onestown, PA 17038. Address
correspondence to S. W. Wise.

3

Department of Emergency Medicine, Pennsylvania State
University College of Medicine, Hershey, PA 17033.

4

Department of Pathology, Pennsylvania State University
College of Medicine, Hershey, PA 17033.

5

Department of Health Evaluation Sciences, Pennsylvania
State University College of Medicine, Hershey, PA 17033.

AJ R

2001;176:933–941
0361–803X/01/1764–933
© American Roentgen Ray Society

M

934

AJ R:176, April 2001

Wise et al.

material, with or without the use of additional
oral contrast material. However, to our knowl-
edge, a comparative assessment of the CT and
sonographic techniques has not been performed
on the same patient population with indepen-
dent interpretations rendered by a group of radi-
ologists. We undertook this investigation to
evaluate the following five appendiceal imaging
techniques: graded compression sonography,
unenhanced focused appendiceal CT, standard
abdominopelvic CT using IV contrast material,
focused appendiceal CT using colonic contrast
material, and sonography using colonic contrast
material. The intent is to provide a comprehen-
sive assessment of the diagnostic performance,
diagnostic confidence, inter- and intraobserver
variability, technical success, and patient com-
fort with each imaging technique.

Materials and Methods

From March 9, 1998 to June 21, 1999, 100 consec-
utive adult (>18 years old) patients were examined us-
ing a series of CT and sonographic techniques for the
diagnosis of appendicitis (26 men, 74 women; age
range, 18–86 years; mean age, 38 years). The study
had the full approval of the institutional review board
for human subjects protection and our institutional ra-
dioisotope committee. Patient selection for the investi-
gation was determined by the referring clinicians in
that all enrolled patients presented to the emergency
department or outpatient clinic with appendicitis in
the first three of the differential diagnoses. Informed
written consent was obtained from all patients. An ad-
ditional 49 patients were not enrolled in the study for
a variety of reasons, most commonly refusal to con-
sent to all imaging studies (Table 1).
Oral contrast material (400–500 mL of 3% meglu-
mine diatrizoate solution [Gastrografin; Bristol-Mey-
ers Squibb, Wallingford, CT]) was given to each
patient, who was instructed to drink the solution as tol-
erated for 30–60 min. The patients were taken to the
emergency department CT scanner (PQ 5000; Picker
International, Cleveland Heights, OH), and a predeter-
mined array of examinations was performed. First,
graded compression sonography (128XPIO scanner;
Acuson, Mountain View, CA) of the right lower ab-
dominal quadrant was performed (by one of two dedi-
cated sonographers or one sonologist) using the graded
compression technique described in the literature [12].
Second, unenhanced focused appendiceal CT was per-
formed, obtaining an average of 21 consecutive 4-mm
axial images over approximately 8.4 cm of the right
lower abdominal quadrant (centered at the level of the
cecal apex) using a small field of view centered over
the right two thirds of the abdomen. Third, complete
helical CT of the abdomen and pelvis was performed,
obtaining 8-mm consecutive images using a field of
view large enough to encompass the entire peritoneal
cavity. Nonionic contrast material was administered
for this study (125 mL of Optiray 320 [ioversol];
Mallinckrodt, St. Louis, MO), injected through a pe-
ripheral IV catheter at 2.5 mL/sec (65-sec scan delay).
Fourth, focused appendiceal CT was performed after
the administration of 800–1200 mL of colonic contrast
material (3% meglumine diatrizoate solution) through
a rectal catheter placed by the radiologist monitoring
the study. Consecutive 4-mm small-field-of-view im-
ages were obtained focused on the cecal apex in a
manner similar to the unenhanced focused appendiceal
CT, with a mean anatomic coverage of 8.8 cm (mean,
22 images). A fifth study, repeated sonography of the
right lower abdominal quadrant after administration of
colonic contrast material, was performed. The tech-
nique used for each CT scan included 200–300 mA,
100–200 kVp, 1.5 pitch, and a 1.0 sec per revolution
acquisition time. Standard (180°) linear interpolators
and standard reconstruction kernels were used.
Patients were randomly assigned to one of two im-
aging sequences. Fifty percent of patients underwent
the studies in the order just outlined. For the remaining
50%, the order of the abdominopelvic CT and focused
appendiceal CT using colonic contrast material was re-
versed (the focused appendiceal CT with colonic con-
trast material in this sequence being performed before
the abdominopelvic CT instead of after it).
A digital scout image of the abdomen and pelvis
was obtained before each CT scan and was used to lo-
calize the cecum for the two focused appendiceal CT
scans. If the cecum could not be localized on the scout
image, the images were centered approximately 1 cm
inferior to the iliac crest.
The CT scans were evaluated on the monitor before
the second sonographic examination was performed. If
the diagnosis of appendicitis remained uncertain to the
radiologist monitoring the study, additional images (4-
mm axial slices) of the right lower abdominal quadrant
were obtained with the patient in the right or left lateral
decubitus position. The right lateral decubitus position
was chosen if the colonic contrast material did not ini-
tially opacify and distend the cecum. The left lateral
decubitus position was chosen if the cecum was well
filled with contrast material but the diagnosis was
equivocal. These decubitus scans were used for both
the prospective and retrospective interpretations.
The radiologist on duty for body imaging at the
time of the studies interpreted all CT and sono-
graphic studies as a unit, with the overall interpreta-
tion used for clinical treatment. The diagnosis of
appendicitis and the alternative diagnoses were then
documented using a confidence scale of 0–10. Find-
ings related to appendicitis and findings that were
nonappendiceal in nature were recorded as present,
equivocally present, or absent. Ancillary findings
were categorized according to the primary diagnosis
(e.g., fluid collection, bowel obstruction) or inciden-
tal findings. In addition, the radiologists interpreting
the studies prospectively recorded the specific exam-
ination that was considered most useful in the as-
sessment of that particular case.
Each CT scan was filmed separately, with the
only identification on the films being the patients’
six-digit medical record number. At a later date,
four observers independently interpreted each in-
dividual scan using the same scale described for
documenting the prospective diagnosis. Unen-
hanced focused appendiceal CT, abdominopelvic
CT, and focused appendiceal CT with colonic con-
trast material scans were interpreted separately
and in random order (for 16 months), whereas the
right and left lateral decubitus scans were inter-
preted only in conjunction with the focused appen-
diceal CT scans for which colonic contrast
material was used (after the focused scans were
first reviewed independently).
Three of the four CT observers were fellowship-
trained body imagers with various degrees of aca-
demic focus in gastrointestinal radiology. The fourth
observer was a third-year radiology resident.
Two experienced sonographers performed 94 of
the 100 sonographic studies (a sonologist performed
the remaining six). Graded compression sonography
of the right lower abdominal quadrant was performed
in the presence of the sonologist. The sonographers
were unaware of the CT results and performed the
second right lower abdominal quadrant sonography
after the final CT (while the colonic contrast material
was still in place).
The real-time sonography was videotaped, with
the only identification on the tapes being the six-
digit medical record numbers and whether the scan
was obtained before or after colonic contrast mate-
rial administration. Body markers and annotation
were added as appropriate. The videotaped scans
were then interpreted separately and in a blinded
a
Clinical CT scanner and sonography units were used,
and cases were prioritized according to clinical urgency.
b
Known cases; more might have existed.
TABLE 1
Reasons for Patient
Nonenrollment into
Investigation
Reason No.
Patient refusal (n =30)
Use of colonic contrast material 18
Mistrust of research 5
Desire to minimize time 1
Combination of time factor and
colonic contrast material
1
Fear of IV contrast material 1
Fear of an IV 1
Prior bad experience with oral and IV
contrast material
1
Desire only CT 1
No insurance 1
Technically not possible for patient (n=6)
History of allergy to IV contrast material 3
Pregnant 2
Spoke no English 1
Schedule conflict
a
4
Sonographer not available 2
Surgeon advised immediate surgery 7
b
Total 49

CT and Sonography of the Appendix

AJ R:176, April 2001

935

fashion by three radiologists, including two staff so-
nologists and a third-year radiology resident.
Before the interpretative stage of the investiga-
tion, each observer reviewed the appendicitis-related
radiology literature published in the preceding 4
years. Furthermore, prior clinical cases were avail-
able to review specific imaging findings and varia-
tions of appendicitis.
One observer interpreted each CT scan and
sonogram twice to assess intraobserver variability.
Given the multiplicity of the number of individual
scans for 100 patients and the number of observers
involved in the study, a total of 3110 interpreta-
tions were performed (Table 2).
After each sonographic and CT technique was
performed, the patient was asked to rate the level
of discomfort for that procedure on a scale of 0–10
(0 being no discomfort and 10 being maximal dis-
comfort), enabling an assessment of patient prefer-
ence for techniques relative to one another.
An experienced emergency medicine physician es-
tablished the clinical diagnosis for each nonsurgical
case. Findings at pathologic assessment of resected ap-
pendixes were used as the gold standard for patients
who underwent appendectomy. An experienced gas-
trointestinal pathologist reviewed each resected speci-
men to confirm the presence or absence of
appendicitis. Patients who did not have surgery were
contacted by telephone for clinical follow-up 1 week,
1 month, and 3 months after presentation. The final di-
agnosis for each patient incorporated information from
the surgical and pathologic findings, clinical data at
discharge, and the clinical follow-up.
In addition to parameters of diagnostic perfor-
mance, we assessed diagnostic confidence with each
imaging technique. A 10-point scale was used for the
interpretations, and the mean scores for the cases with
negative and positive findings were assessed as sepa-
rate groups. In this analysis, a technique that is perfect
would have a mean value of 0 for a case with negative
findings and 10 for a case with positive findings. This
assessment incorporated both relative accuracy and the
diagnostic confidence level of the interpretations for
each technique.
Several technical aspects of CT were assessed.
First, the anatomic coverage of the appendix was as-
sessed for the focused techniques using a 10-point
scale. Second, the amount of colonic contrast material
reaching the cecum was assessed for focused appen-
diceal CT using colonic contrast material. Cecal opaci-
fication and distention were assessed on a 10-point
scale from no contrast in the cecum (0) to excellent
opacification (10). If the degree of cecal opacification
was considered poor (<5), the suspected reason for the
failure (colonic stool, sigmoid redundancy, poor effort
by radiologist, or other) was recorded. Third, for the
second half of the study, the frequency of colonic con-
trast material leakage and problems related to the IV
contrast material were assessed. The amount of co-
lonic contrast leakage was assessed qualitatively
(none, mild, moderately, and severe), whereas prob-
lems related to IV contrast material (e.g., extravasa-
tion) were recorded as they occurred. Last, the
radiologist interpreting the studies prospectively was
asked his or her preference among the three CT tech-
niques and sonography.
Descriptive statistical analysis included means and
standard deviations for continuous outcomes and fre-
quency tables for categoric outcomes. Diagnostic pa-
rameters for each technique—sensitivity, specificity,
positive predictive value, negative predictive value, and
accuracy—were calculated separately for each ob-
server. The means of all observers were used in the
analyses unless otherwise indicated. Statistical com-
parisons among techniques with respect to categoric
outcomes, such as diagnostic parameters, were made
using the Cochran and Mantel-Haenszel tests, which
accounts for the fact that all techniques were used on
the same patients. Statistical comparisons among tech-
niques with respect to continuous outcomes, such as
diagnostic confidence score, were made using re-
peated measures analysis of variance with technique
and observer as within-subject factors. Intra- and inter-
observer agreement with respect to diagnostic confi-
dence were quantified using the weighted kappa
statistic. Values of the kappa statistic greater than 0.7
typically indicate good agreement, whereas values be-
tween 0.4 and 0.7 reflect moderate agreement. All
analyses were carried out using the SAS statistical
software system version 8.0 (SAS, Cary, NC).

Results

Diagnostic Performance

Twenty-four of the 100 patients had positive
findings for acute appendicitis. The prospective
CT interpretations had the best overall diagnos-
tic performance, with a sensitivity of 96%, spec-
ificity of 92%, positive predictive value of 79%,
negative predictive value of 99%, and accuracy
of 93% (Table 3). The retrospective interpreta-
tions produced comparable results among the
three primary CT imaging techniques; however,
considerable variation was seen among the ob-
servers. The range in accuracy was 75–94% for
unenhanced focused appendiceal CT, 82–91%
for abdominopelvic CT, and 75–92% for fo-
cused appendiceal CT with colonic contrast ma-
terial. The two focused appendiceal CT
techniques had slightly higher sensitivity for de-
tecting appendicitis than abdominopelvic CT
(79% vs. 71%), but this was not statistically sig-
nificant (

p

= 0.21; range among observers, 5–
76%). The addition of either right or left lateral
decubitus scanning to focused appendiceal CT
with colonic contrast material did not improve
diagnostic performance.
The prospective and retrospective sono-
graphic interpretations had comparable accu-
racy, but the prospective assessment provided
significantly higher sensitivity (

p

< 0.01). No
differences were noted in the diagnostic perfor-
mance of sonography with or without the use of
colonic contrast material (

p

> 0.05). Both sono-
graphic techniques were inferior to CT (

p

<
0.0001), with the greatest difference being a low
sensitivity for sonography (range, 13–65%
among observers).

Inter- and Intraobserver Variability

Table 4 shows the kappa values for inter- and
intraobserver variability with each imaging
technique. Regardless of the specific imaging
technique, intraobserver variability was signifi-
cantly less than interobserver variability (

p

<
0.0001). The sonographic techniques produced
significantly worse inter- and intraobserver
agreement than the three CT techniques. Unen-
hanced focused appendiceal CT was the CT
technique with the greatest inter- and intraob-
server variability, whereas focused appendiceal
CT with colonic contrast material had the least
variability (

p

= 0.04).
Note.—N/A =not applicable.
a
Includes five observers for sonography and four for CT, plus one observer in each group interpreting each study twice.
TABLE 2 Number of Interpretations
Type of Study
Studies
Performed
Interpretations
Total
Prospective Retrospective
a
Unenhanced sonography 100 100 600 700
Contrast-enhanced sonography 100 100 600 700
Focused appendiceal CT
Unenhanced 100 N/A 500 500
Enhanced with colonic contrast material 100 N/A 500 500
Right lateral decubitus 8 N/A 40 40
Left lateral decubitus 14 N/A 70 70
Abdominopelvic CT 100 N/A 500 500
All three CT studies together 100 100 N/A 100
Total 3110

936

AJ R:176, April 2001

Wise et al.

Diagnostic Confidence

Figure 1 shows the mean confidence scores
for each imaging technique. Again, the prospec-
tive interpretations of CT produced the highest
confidence scores for both the negative and posi-
tive findings. For the retrospective CT interpreta-
tions, abdominopelvic CT provided the greatest
level of confidence for negative findings (mean,
2.21) relative to the two focused appendiceal CT
techniques (mean, 2.75;

p

= 0.001). However,
abdominopelvic CT produced significantly
lower confidence for the positive cases than did
both focused CT methods (

p

= 0.0001), with the
greatest confidence seen in positive cases with
focused appendiceal CT using colonic contrast
material (

p

= 0.02).
The retrospective interpretations of the
sonograms were comparable to the CT scans
for negative findings but were significantly
inferior to all CT techniques for cases with
positive findings (

p

< 0.0001). Prospective
sonography produced comparable confi-
dence to the retrospective interpretations of
the negative findings but significantly
greater confidence for the positive findings
(

p

< 0.006).
During imaging, the appendix was identi-
fied with greater frequency on CT than on
sonography, regardless of technique (

p

<
0.0001) (Table 5). Significant interobserver
differences were noted for both CT and
sonography (

p

< 0.01) but without signifi-
cant trends among the techniques.

Use of Colonic or IV Contrast Material for CT

The use of colonic contrast material did
not improve the diagnostic performance of
CT (

p

> 0.33) (Table 6). The addition of co-
lonic contrast material for focused appen-
diceal CT and abdominopelvic CT did not
alter diagnostic confidence for the cases with
negative findings (

p

= 0.30) but significantly
improved confidence in cases with positive
findings (

p

= 0.02).
Focused appendiceal CT using colonic con-
trast material was assessed between the two
imaging protocols. (Focused appendiceal CT
with colonic contrast material either preceded
or followed abdominopelvic CT with IV con-
trast material.) This comparison enabled as-
sessment of the use of IV contrast material, at
least with delayed scanning. As seen in Table
6, no significant difference was noted between
focused appendiceal CT scans with colonic
contrast material that also used IV contrast
material and those that preceded IV contrast
administration (

p





0.18).

Other Analyses

When we used the discomfort scale of 0–10
(0 = least discomfort, 10 = most discomfort), the
mean patient discomfort scores for each tech-
nique were as follows: sonography without co-
lonic contrast material, 6.1; sonography with
colonic contrast material, 6.7; unenhanced fo-
cused appendiceal CT, 4.2; abdominopelvic CT,
5.3; and focused appendiceal CT with colonic
contrast material, 6.7. Patients reported signifi-
cantly less discomfort for unenhanced focused
appendiceal CT than for any other technique (

p

<
0.0001). Focused appendiceal CT with colonic
contrast material produced significantly higher
patient discomfort than the other CT techniques
but did not significantly differ from sonography
(

p

< 0.0001 and

p

= 0.051–0.075 for CT and
sonography comparisons, respectively).
The following technical problems were en-
countered during CT. Colonic contrast mate-
rial leaked onto the table in 12 cases (six
mild, one moderate, and five severe leakage)
(24%), nausea or vomiting occurred in three
patients (6%) during scanning, and one pa-
tient (2%) experienced extravasation of IV
contrast material. In addition, cecal contrast
opacification or distention was graded as
suboptimal (<5 on the 10-point scale) in
18% of the 100 focused appendiceal CT
scans with colonic contrast material, with 16
of the 18 failures caused by abundant stool in
the right colon. (Of the two remaining fail-
ures, one case was related to excessive sig-
moid redundancy and one was suspected to
be inadequate effort by the radiologist.)
The accuracy of localizing the cecum and the
appendiceal region using the scout topogram re-
vealed that 8% of the unenhanced focused ap-
pendiceal CT scans and 4% of the focused
appendiceal CT scans with colonic contrast ma-
terial were graded as relatively poor for accu-
racy of scanning the appendiceal region.
The radiologists monitoring the cases
were prospectively asked their preference
among the techniques for establishing the di-
agnosis with greatest confidence. Abdomi-
nopelvic CT or focused appendiceal CT with
colonic contrast material was preferred to
sonography or unenhanced focused appen-
diceal CT in 87 of 100 cases (

p

< 0.0001).
Focused appendiceal CT with colonic con-
trast material was preferred in 45 cases and
abdominopelvic CT in 42 cases.
Note.—Dash (—) indicates insufficient cases to assess. PPV =positive predictive value, NPV =negative predictive value.
TABLE 3 Diagnostic Performance per Imaging Technique
Technique Sensitivity Specificity PPV NPV Accuracy
Unenhanced sonography 34 86 42 81 74
Enhanced sonography, colonic contrast
material
32 89 47 81 76
Abdominopelvic CT 71 91 71 91 87
Focused appendiceal CT
Unenhanced 83 85 58 95 85
Enhanced, colonic contrast material 74 87 70 90 84
With right lateral decubitus scan 62 73 28 92 71
With left lateral decubitus scan — 91 — 100 91
Prospective sonography
Unenhanced 62 71 37 87 69
Enhanced, colonic contrast material 57 82 46 87 76
Prospective CT 96 92 79 99 93
Note.—Kappa scores can range from0 to 1.0, with 0 indicative of no agreement and 1.0 signifying complete agreement.
Numbers in parentheses are kappa scores for radiologists (sonographers excluded).
a
With colonic contrast material.
TABLE 4
Inter- and Intraobserver Variability (Kappa Scores) for Each Imaging
Technique
Variability
Sonography Focused Appendiceal CT
Abdominopelvic
CT
Unenhanced Enhanced
a
Unenhanced Enhanced
a
Interobserver 0.20 (0.18) 0.15 (0.11) 0.36 0.45 0.41
Intraobserver 0.43 0.39 0.76 0.85 0.79

CT and Sonography of the Appendix

AJ R:176, April 2001

937

Alternative diagnoses were established us-
ing imaging in 30 cases. Specific diagnoses
included ovarian cyst (

n

= 13), cecal diverticu-
litis (

n

= 3), sigmoid diverticulitis (

n

= 2), in-
flammatory bowel disease (

n

= 2), biliary
colic (

n

= 1), mesenteric adenitis (

n

= 1), pel-
vic inflammatory disease (

n

= 2), pelvic he-
matoma (

n

= 1), endometriosis (

n

= 1),
fibroid (

n

= 1), segmental omental infarction
(

n

= 1), choledocholithiasis (

n

= 1), and ure-
teral calculus (

n

= 1). The latter three diag-
noses required complete abdominopelvic
scanning for identification.
The number of ancillary (incidental) find-
ings (e.g., gallstones) identified was aver-
aged for all observers to provide a relative
(but fractional) assessment on a per-patient
basis. For ancillary findings related to the di-
agnosis, sonography without colonic contrast
material revealed 0.1 findings per patient;
sonography with colonic contrast material,
0.1 findings; unenhanced focused appen-
diceal CT, 0.1 findings; abdominopelvic CT,
0.3 findings; and focused appendiceal CT
with colonic contrast material, 0.1 findings.
For incidental findings, sonography without
colonic contrast material revealed 0.1 find-
ings; sonography with colonic contrast mate-
rial, zero findings; unenhanced focused
appendiceal CT, 0.1 findings; abdominopel-
vic CT, 0.3 findings; and focused appen-
diceal CT with colonic contrast material, 0.1
B A
Fig. 1.—Graphs show diagnostic
confidence per observer and imag-
ing technique.
Aand B, Each numbered line repre-
sents mean confidence score of single
observer using each technique for
cases with negative (A) and positive (B)
findings for appendicitis. Optimal score
would be 0for cases with negative find-
ings and 10for cases with positive find-
ings. FACT =focused appendiceal CT,
A/P CT = abdominopelvic CT. “En-
hanced” indicates use of colonic con-
trast material.
Note.—Percentages given in parentheses are based on a maximumof 76 normal appendixes that could be identified (24
appendicitis cases in study).
a
With colonic contrast material.
TABLE 5
Number of Normal Appendixes Identified by Observer and Imaging
Technique
Sonography
Observer
Sonography
CT Observer
CT
Unenhanced Enhanced
a
Focused Appendiceal
Abdominopelvic
Unenhanced Enhanced
a
1 21 (28) 18 (24) 1 18 (24) 20 (26) 19 (25)
2 21 (3) 1 (1) 2 28 (38) 24 (32) 35 (46)
3 10 (13) 11 (14) 3 35 (47) 40 (53) 36 (47)
4 13 (17) 15 (20) 4 47 (64) 56 (74) 51 (67)
5 4 (5) 6 (8)
Note.—PPV =positive predictive value, NPV =negative predictive value.
a
On a scale of 0–10. Optimal score would be 0 for cases with negative findings and 10 for cases with positive findings.
b
Focused appendiceal CT with colonic contrast material plus abdominopelvic CT (half of abdominopelvic CT before and half after administration of colonic contrast material).
c
Focused appendiceal CT with colonic contrast material either before or after abdominopelvic CT with IV contrast material.
TABLE 6 Effect of Colonic and IV Contrast Material
Type of Imaging and Contrast
Material
Sensitivity
(%)
Specificity
(%)
PPV
(%)
NPV
(%)
Accuracy
(%)
Confidence Level
a
κ
Negative
Findings
Positive
Findings
Intraobserver Interobserver
CT
b
With colonic contrast material 78 89 68 93 86 2.5 7.6 0.82 0.45
Without colonic contrast material 76 87 65 92 84 2.6 7.1 0.78 0.37
Focused appendiceal CT
c
With IV contrast material 77 90 77 92 87 2.6 7.9 0.85 0.43
Without IV contrast material 83 83 61 94 83 2.8 8.2 0.86 0.46

938

AJ R:176, April 2001

Wise et al.

findings. Abdominopelvic CT was superior
to the other techniques in the identification
of ancillary findings regardless of whether
they were related to the primary diagnosis
(

p

< 0.01). Renal cysts, gallstones, diverticu-
losis, and one large hemangioma were inci-
dental findings that were frequently identified
on abdominopelvic CT.

Discussion

The usefulness of sonography and CT in the
diagnosis of appendicitis has been reported ex-
tensively in the literature. Because the surgery
literature reports an average rate of negative
findings at appendectomy of 20% [13, 14], im-
aging techniques are useful to prevent unneces-
sary appendectomies and to help avoid costly
hospital admissions. Furthermore, imaging can
expedite the diagnosis of appendicitis, mini-
mizing surgical delays and the subsequent risk
of appendiceal perforation. Rao et al. [5]
showed the cost-effectiveness of a focused ap-
pendiceal CT technique in the clinical treat-
ment of patients with suspected appendicitis. In
that study, a 98% interpretive accuracy of CT
resulted in a net savings of $44,731 for the
treatment of 100 consecutive patients. The au-
thors concluded that an immediately available
and accurate imaging examination could diag-
nose or exclude appendicitis in a timely fash-
ion, decreasing patient morbidity and saving
significant money when widely implemented.
Several appendiceal imaging techniques have
been advocated in the literature. Puylaert et al.
[12] described a graded compression sono-
graphic technique for appendiceal imaging re-
ported to be 83–96% accurate for the diagnosis
of appendicitis [8, 15, 16]. Complete CT of the
abdomen and pelvis using oral and IV contrast
material has also been advocated and shown to
be 94% accurate in the examination of patients
with possible appendicitis [8]. This technique,
initially described by Balthazar et al. [8], entails
the use of thin-section (5 mm) imaging through
the cecal region, with thicker images in the up-
per and lower portions of the abdomen and pel-
vis. Unenhanced thin-section CT through the
right lower abdominal quadrant with or without
imaging the remainder of the peritoneal cavity
has also been described [6, 7], yielding an accu-
racy of 93–97% without using oral, IV, or co-
lonic contrast material.
Rao et al. [10, 11] developed the focused ap-
pendiceal CT technique in which a helical scan
with 5-mm collimation is obtained through the
appendiceal region after the administration of
approximately 1 L of colonic contrast material.
This technique, with or without the addition of
orally administered contrast material, has a re-
ported accuracy of 98% in the diagnosis of ap-
pendicitis [10, 11]. Furthermore, this technique
enables a relatively quick diagnostic examina-
tion (there is no delay related to oral contrast ma-
terial) while maintaining the benefit of enteric
contrast material and cecal distention.
All these techniques offer promising results.
However, few comparative studies have been
performed [8]. We chose to undertake a com-
prehensive comparative study of appendiceal
imaging techniques evaluating not only the di-
agnostic performance of each technique, but
also the inter- and intraobserver variability, di-
agnostic confidence for negative and positive
findings, number of normal appendixes identi-
fied, and relative patient discomfort with the
various techniques.
The prospective interpretations of the CT
scans in our study had the highest overall di-
agnostic performance, despite the fact that the
initial observers were frequently residents at
various stages of training with or without ad-
ditional faculty input. This observation has
two explanations. First, the combination of
the various CT techniques with sonography is
better than any individual study. The useful-
ness of performing tandem techniques has not
been assessed to date and may warrant further
investigation. Second, when examining a pa-
tient with abdominal pain, it is difficult to
have prospective interpretations in which the
diagnostician is completely unaware of clini-
cal data. The influence of clinical correlation
may have significant diagnostic impact in
cases that are equivocal for appendicitis at im-
aging. Bias can be introduced from the im-
pressions of referring clinicians, from
laboratory data, or simply through patient ob-
servation during provocative imaging maneu-
vers such as compression with a sonographic
transducer or the introduction of colonic con-
trast material. This effect may account for the
overall poorer performance of the retrospec-
tive interpretations when compared with the
performance data in the literature that consis-
tently reports accuracies greater than 90%.
This effect may also account for the overall
higher sensitivity for prospective rather than
for retrospective sonographic interpretations.
CT had significantly better diagnostic perfor-
mance than sonography in our study, but no sig-
nificant differences were noted among the
individual CT techniques tested. Furthermore,
the use of colonic or IV contrast material (at
least in delayed scanning) had no effect on diag-
nostic performance. These findings support the
conclusion postulated in the literature that com-
parable performances can be obtained regardless
of the specific CT technique used [17] (Fig. 2).
The assessment of diagnostic performance in
appendiceal imaging is undoubtedly important,
particularly with a panel of observers who have a
range of subspecialty clinical expertise and aca-
demic interest in the subject. Radiology investi-
gations tend to emphasize diagnostic
performance; however, diagnostic confidence
and observer variability are also important in the
diagnosis or exclusion of surgical conditions
such as acute appendicitis. If surgeons recognize
excessive interpretive variability or suboptimal
diagnostic confidence of radiologists, they will
be more likely to discount the efficacy of the im-
aging examination regardless of the accuracy re-
ported in the literature.
The optimal imaging approach should have
several key characteristics. Imaging tests
should be sensitive, specific, and accurate; en-
able observers to be confident of both negative
and positive findings with little inter- and in-
traobserver variability; and be easily per-
formed with little risk and discomfort for
patients. Rao et al. [11] addressed several is-
sues that support the use of limited CT for ap-
pendiceal imaging rather than CT of the entire
abdomen and pelvis. First, considerable cost
savings can be realized by limiting the CT and
reducing the cost and charge for the study to
that for a pelvic CT. Second, the amount of ra-
diation exposure for patients is significantly
less with limited CT than with CT of the entire
abdomen and pelvis. Most patients referred for
appendiceal examination are relatively young,
and gonadal radiation should be minimized,
particularly for young women. Third, the use
of IV contrast material has some inherent risk
and intrinsic cost that is best avoided if a clear-
cut clinical advantage is not seen. Fourth, the
study can be performed rapidly (without de-
lay), resulting in quick diagnoses.
IV contrast material was used in each ab-
dominopelvic CT examination. When assessing
the focused appendiceal CT studies, no statisti-
cal influence was seen for the use of IV contrast
material on the diagnostic performance of the
focused appendiceal CT with colonic contrast
material that followed the abdominopelvic CT
(even though appendiceal mural enhancement
can be indicative of appendicitis). This use of
IV contrast material, however, combined with
full coverage of the abdomen and pelvis during
CT, does raise the cost of the study and may
negatively influence the reported cost-effective-
ness of CT for the examination of patients with
possible appendicitis. The technique of helical

CT and Sonography of the Appendix

AJ R:176, April 2001

939

abdominopelvic CT using prospectively
planned thin sections through the appendiceal
region (during dynamic enhancement) was not
specifically tested in this study. However, this
approach in theory takes advantage of dynamic
enhancement IV contrast material as well as
thin sections.
The limited scope of focused CT may result
in pathologic disorders not being fully imaged.
Rao et al. [10] reported in their initial investi-
gation that 8% of patients underwent extended
scanning (beyond the focused appendiceal CT
anatomic coverage) “to follow the cranial or
caudal extent of pathologic abnormalities
noted on the initial limited scan.” In a subse-
quent study, the same group reported the need
to extend scanning into the upper abdomen or
pelvis in 5% of cases [11].
Our study evaluated this issue by perform-
ing focused appendiceal CT with less overall
coverage than was used in prior studies. The
technique used in these prior investigations en-
tailed 5-mm images with approximately 15 cm
of coverage using a field of view encompass-
ing the entire peritoneal cavity. Conversely, we
used 4-mm images with 8.4–8.8 cm of cover-
age and a field of view encompassing the right
two thirds of the peritoneal cavity. (We chose
this technique to limit the amount of patient ra-
diation and to magnify the performance differ-
ences of focused appendiceal CT vs. standard
abdominopelvic scanning.)
Despite this focused approach to the fo-
cused appendiceal CT studies, we had good
anatomic coverage of the appendix in most pa-
tients. However, three alternative diagnoses
(choledocholithiasis, ureteral calculus, and
segmental omental infarction) were encoun-
tered that would have required additional im-
aging of the upper abdomen or lower pelvis for
detection. In addition, ancillary findings were
more frequently identified on abdominopelvic
CT than on either focused CT or sonography.
Although most of these findings would proba-
bly not have altered clinical treatment of the
patients, abdominopelvic CT provided a more
complete assessment of the peritoneal cavity.
These observations, along with the superior
level of observer confidence in cases with
negative findings and relatively good inter-
and intraobserver variability, lead us to con-
clude that abdominopelvic CT may be the
best initial examination for suspected appen-
dicitis in adult patients. Other observations
also support this conclusion. The patients ex-
perienced a relatively low degree of discom-
fort with the technique. Radiologists preferred
interpreting the abdominopelvic CT much
more than the unenhanced focused appen-
diceal CT scans or sonograms. The identifica-
tion of normal appendixes using 8-mm
abdominopelvic images was comparable to
identification on thin-section scans. (The latter
observation differs slightly from a previous
investigation that reported significantly higher
frequency of normal appendix identification
with 5-mm as opposed to 10-mm image colli-
mation [18].)
D
A
Fig. 2.—Acute appendicitis readily diagnosed in
24-year-old man using all CT techniques but not
identified on sonography. Minimal periappen-
diceal inflammatory stranding is present on all
CT images.
A, Unenhanced focused appendiceal CT scan
shows appendicolith (black arrow) and slightly
enlarged appendix (white arrow).
Band C, Abdominopelvic CT scans show equivo-
cal cecal apical thickening (arrowhead, B) and ap-
pendicolith (arrow, C) that is less conspicuous.
Dand E, Focused appendiceal CT scans with
colonic contrast material show cecal apical
thickening (arrowhead, D) and appendicoliths of
intermediate conspicuity (arrow, E).
C
E
B

940

AJ R:176, April 2001

Wise et al.

The patients themselves were proponents
for excluding the use of colonic contrast mate-
rial, which became apparent through several
observations. First, most patients refusing to
enroll in the study did so to avoid receiving
colonic contrast material. Second, the tech-
nique with the greatest relative patient dis-
comfort was the one using colonic contrast
material. Finally, the colonic contrast material
could not be advanced successfully into the
cecum in 18% of patients and was associated
with leakage onto the CT table in 24%.
Clearly, if an examination produces greater
patient discomfort and offers greater technical
challenges than rival techniques, its use
should be justified through a significant im-
provement in diagnostic performance, greater
degrees of diagnostic confidence, or lower in-
ter- and intraobserver variability.
The diagnostic benefits of colonic contrast
material for appendiceal CT relate to the
ability to identify specific cecal apical find-
ings (e.g., “arrowhead” and cecal bar signs)
and to exclude appendicitis when colonic
contrast material fills the appendiceal lumen.
Focused appendiceal CT with colonic con-
trast material offered the greatest diagnostic
confidence for cases with positive findings
and the least intra- and interobserver vari-
ability. The significant advantage of diagnos-
tic confidence in positive cases was also seen
when the interpretations of the abdominopel-
vic CT scans were divided as to the presence
or absence of colonic contrast material and
were included in the analysis. However, fo-
cused appendiceal CT with colonic contrast
material did not yield a higher diagnostic
performance than the other CT techniques
and resulted in greater patient discomfort
and technical challenges. Therefore, we con-
clude that focused appendiceal CT with co-
lonic contrast material is probably best used
as a problem-solving technique for equivocal
appendiceal cases after an initial assessment
with standard abdominopelvic CT.
Sonography had a lower diagnostic per-
formance and observer confidence than CT.
Furthermore, the tenderness evoked by the
transducer pressure produced the second
highest rating of patient discomfort among
the imaging techniques.
Before the investigation, we theorized that
distending the cecum with contrast material
could provide a useful window for the sono-
graphic examination of appendixes, particularly
those in a retrocecal location. Initial in vitro
sonography with diluted iodinated colonic con-
trast material revealed it to be a primarily
anechoic solution with scattered specular reflec-
tors. However, the technique was unsuccessful
in improving the accuracy of appendiceal
sonography (Fig. 3). If the cecum is not com-
pletely fluid-filled and retains a small amount of
gas, the retrocecal region can still be obscured.
The primary limitation of this study is the
possibility of spontaneously resolving appen-
dicitis in a few patients. In three patients in our
series, one of the CT techniques raised the pos-
sibility of mild appendicitis, but the patient did
not undergo surgery and subsequently the
symptoms resolved (Fig. 4). Resolution of
mild appendicitis has been widely recognized
during the past several years [19, 20], and ap-
pendicitis may recur in approximately 38% of
cases [21]. One case of apparent mild appendi-
citis was missed prospectively in the early
stage of our study, but the patient was reen-
rolled 7 weeks later when symptoms recurred.
Repeated imaging of this patient showed clas-
sic findings of appendicitis.
Our study has unique limitations to the
evaluation of appendiceal sonography.
Sonographic interpretations were limited by
the study design in that there was no com-
munication between the sonographer and
the reviewing physician (for the retrospec-
tive interpretations) and no opportunity for
the reviewing physician to personally per-
form a real-time examination. These factors,
plus the inability to correlate sonographic
findings with patient tenderness during the
examination, may explain the poor perfor-
mance of sonography. (Patient body habitus,
which may have been a significant limiting
factor, is being evaluated separately.) Our
study specifically excluded pediatric pa-
tients—a subgroup of patients considered
ideal for appendiceal sonography—which is
another factor that likely contributes to the
improved performance of CT over sonogra-
phy. The predominance of female patients
(with relatively greater differential diag-
noses) in our series likely added to the diag-
nostic challenge of the imaging techniques.
In conclusion, our results suggest that ab-
dominopelvic CT with a standard protocol
should be the initial approach to imaging
patients being examined for appendicitis.
Focused appendiceal CT using colonic con-
trast material should be used for further ex-
amination of equivocal cases. This
approach enables maximum diagnostic con-
fidence and reduction in intra- and inter-ob-
server variability.

Acknowledgments

We appreciate Doreen Katzaman’s assis-
tance in the preparation of this manuscript.
Additionally, we thank the Penn State Col-
lege of Medicine residency classes of 1998,
1999, and 2000 for their contributions to our
investigation.
B A
Fig. 3.—Sonography of 36-year-old woman shows that right lower abdominal quadrant anatomy is better seen
without colonic contrast material.
A, Sonogramobtained before administration of colonic contrast material depicts right lower abdominal quadrant
anatomy with reasonably good quality. Note right iliac vessel adjacent to right iliopsoas muscle (arrow).
B, Sonogramafter administration of colonic contrast material shows distended cecum(arrow) with floating
specular reflectors but diminished anatomic visualization.
CT and Sonography of the Appendix
AJ R:176, April 2001 941
References
1. Lewis FR, Holcroft JW, Boey J, Dunphy JE. Ap-
pendicitis: a critical review of diagnosis and treat-
ment in 1000 cases. Arch Surg 1975;110:677–681
2. Berry J Jr, Malt RA. Appendicitis near its cente-
nary. Ann Surg 1984;200:567–575
3. Izbicki JR, Knoefel WT, Wilker DK, et al. Accu-
rate diagnosis of acute appendicitis: a retrospec-
tive and prospective analysis of 686 patients. Eur
J Surg 1992;158:227–231
4. Ramirez JM, Deus J. Practical scores to aid deci-
sion making in doubtful case of appendicitis. Br J
Surg 1994;81:680–683
5. Rao PM, Rhea JT, Novelline RA, Mostafavi AA,
McCabe CJ. Effect of computed tomography of the
appendix on treatment of patients and use of hospi-
tal resources. N Engl J Med 1998;338:141–146
6. Lane MJ, Liu DM, Huynh MD, Jeffrey RB, Min-
delzun RE, Katz DS. Suspected acute appendici-
tis: nonenhanced helical CT in 300 consecutive
patients. Radiology 1999;213:341–346
7. Malone AJ Jr, Wolf CR, Malmed AS, Melliere
BF. Diagnosis of acute appendicitis: value of un-
enhanced CT. AJR 1993;160:763–766
8. Balthazar EJ, Birnbaum BA, Yee J, Megibow AJ,
Roshkow J, Gray C. Acute appendicitis: CT and US
correlation in 100 patients. Radiology 1994;190:31–35
9. Rowling SE, Jacobs JE, Birnbaum BA. Thin-section
CT imaging of patients suspected of having appendi-
citis or diverticulitis. Acad Radiol 2000;7:48–60
10. Rao PM, Rhea JT, Novelline RA, et al. Helical
CT technique for the diagnosis of appendicitis:
prospective evaluation of a focused appendix CT
examination. Radiology 1997;202:139–144
11. Rao PM, Rhea JT, Novelline RA, Mostafavi AA,
Lawrason JN, McCabe CJ. Helical CT combined
with contrast material administered only through
the colon for imaging of suspected appendicitis.
AJR 1997;169:1275–1280
12. Puylaert JBCM, van der Zant FM, Rijke AM.
Sonography and the acute abdomen: practical
considerations. AJR 1997;168:179–186
13. Rothrock SG, Green SM, Dobson M, Coluciello
SA, Simmons CM. Misdiagnosis of appendicitis
in nonpregnant women of childbearing age. J
Emerg Med 1995;13:1–8
14. Reynolds SL. Missed appendicitis in a pediatric
emergency room. Pediatr Emerg Care 1993;9:1–3
15. Jeffrey RB, Laing FC, Lewis FR. Acute appendi-
citis: high-resolution real-time US findings. Radi-
ology 1987;163:11–14
16. Jeffrey RB, Laing FC, Townsend RR. Acute ap-
pendicitis: sonographic criteria based on 250
cases. Radiology 1988;167:327–329
17. Birnbaum BA, Jeffrey RB Jr. CT and sonographic
evaluation of acute right lower quadrant abdomi-
nal pain. AJR 1998;170:361–371
18. Weltman DI, Yu J, Krumenaker J, Huang SM,
Moh PP. Comparison of 5 mm and 10 mm CT
section in the same patient in the diagnosis of
acute appendicitis. Radiology 2000;216:172–177
19. Migraine S, Atri M, Bret PM, Lough JO, Hinchey
JE. Spontaneously resolving acute appendicitis:
clinical and sonographic documentation. Radiol-
ogy 1997;205:55–58
20. Jeffrey RB, Jain KA, Nghiem HV. Sonographic
diagnosis of acute appendicitis: interpretive pit-
falls. AJR 1994;162:55–59
21. Cobben LPJ, Otterloo ADM, Puylaert JBCM.
Spontaneously resolving appendicitis: frequency
and natural history in 60 patients. Radiology
2000;215:349–352
C
B A
Fig. 4.—CT in 21-year-old man indicates suspicion of appendicitis
but follow-up produces negative results.
A, Unenhanced focused appendiceal CT scan shows 8-mmap-
pendix (arrow) without definite inflammatory stranding.
B, Abdominopelvic CT scan shows appendix (arrow) that was
seen on only one or two images.
C, Focused appendiceal CT scan with colonic contrast material
shows 8-mm appendix (arrow) with hazy periappendiceal in-
crease in attenuation, suggesting inflammation. Sonography
failed to show appendix.

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close