| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Hôpital Erasme, Université Libre de Bruxelles, Route de Lennik 808, B-1070-Brussels, Belgium (C.K., M.Z., E.C., M.A.B., P.A.G., D.V.G.); and Statistical Unit, Institut de Recherche Interdisciplinaire en Biologie Humaine et Moléculaire, Université Libre de Bruxelles, Brussels, Belgium (V.D.M.). Received June 7, 2004; revision requested August 16; revision received September 2; accepted November 11. Address correspondence to C.K. (e-mail: caroline.keyzer{at}chu-charleroi.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: The institutional review board approved the research protocol. Written informed consent was obtained from all patients or, for those who were adolescents, from their parents. Ninety-four patients (59 female and 35 male patients) aged 16–81 years (mean, 38 years) who were suspected of having acute appendicitis underwent both US and unenhanced multi–detector row CT of the entire abdomen. The examinations were performed within 1–2 hours of each other. US and CT images were obtained and prospectively interpreted by a different radiologist from a group of abdominal radiologists or a group of residents and general radiologists. Radiologists proposed an overall diagnosis and an alternative diagnosis. Data from US and CT were compared, and the definite diagnosis was established with surgical findings (n = 40) or results of clinical follow-up (n = 54) as the reference standard. Comparisons were made for each group of radiologists and the patient's age, body mass index (BMI), and sex. Proportion comparisons were made by using the Pearson 
2 test or the Fisher exact test. Continuous variables were compared between groups with the Mann-Whitney U test.
RESULTS: Thirty patients had definite appendicitis. The sensitivity, specificity, positive and negative predictive values, and accuracy were not significantly different between US and CT or between groups of radiologists (P values ranged from .389 to >.99), regardless of the patient's BMI (P values ranged from .073 to >.99). Misclassifications were compared with the definite alternative diagnosis and were not significantly different between US and CT or between groups of radiologists (P = .061–.592), regardless of patient age (P = .875) or sex (P = .151 and >.99 for male and female patients, respectively). The frequency of inconclusive examinations, however, was significantly higher with US than with CT, regardless of radiologist experience (P = .020 and <.001, respectively).
CONCLUSION: Although the diagnostic performances of US and multi–detector row CT are comparable, more inconclusive images were obtained with US.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Although US is widely available and inexpensive, its accuracy is dependent on the skill of the operator. It has been reported that when patients suspected of having acute appendicitis are examined by experienced operators, the sensitivity of US is 76%–90%, specificity is 86%–100%, positive predictive value is 71%–95%, and negative predictive value is 76%–98% (4,5,11). CT, conversely, has corresponding values that are all greater than 95%, even without the use of enteric or intravenous contrast material (6,12). Furthermore, comparative studies have shown that when contrast material is used with CT (contrast-enhanced CT), the technique is also more reproducible than US (13,14). The use of contrast material, however, adds certain risks and costs to the procedure, and it can be more uncomfortable for the patient. Therefore, the aim of the present study was to prospectively compare the diagnostic performances of US and unenhanced multi–detector row CT in patients suspected of having acute appendicitis by using surgery or clinical follow-up as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The institutional review board approved our research protocol. Written informed consent was obtained from all patients or, for those younger than 18 years, from their parents.
US Examination
Gray-scale US of the entire abdomen, including the pelvis, was performed in all patients by using a 3.75-MHz convex-array transducer (Power Vision 8000; Toshiba Medical Systems, Tokyo, Japan). The right iliac fossa was examined with an 8-MHz linear-array transducer with use of the graded compression technique (16). Color Doppler US of the appendix was performed after completion of the gray-scale US examination to search for hyperemia within the appendix wall (17). While performing the US examination, radiologists were asked to record whether the appendix was visible and, if so, to measure its outer transverse diameter. In addition, they were asked to code the following findings as present or absent: (a) fluid-filled appendix, (b) lack of compressibility of the appendix, (c) hyperemia within the appendiceal wall at color Doppler US, (d) appendicolith, (e) pericecal fluid, (f) hyperechoic periappendiceal tissue, (g) abscess, and (h) maximal tenderness at the site of the appendix. These findings were considered as possible positive criteria for acute appendicitis. After separately coding each finding, radiologists were asked to propose an overall diagnosis of acute appendicitis or an alternative disease that could explain the patient's acute right lower quadrant pain. In case of doubt, it was not possible to code the level of confidence of radiologists in the proposed overall diagnosis.
CT Examination
For CT, patients were examined while in the supine position by using a commercially available scanner with four detector rows (Somatom Plus Volume Zoom; Siemens Medical Systems, Forchheim, Germany). A frontal 512-mm scout view was first obtained with 120 kVp and 50 mA. This was followed by helical scanning from the top of the liver to the symphysis pubis with 4 x 2.5-mm collimation (ie, four detector rows and 2.5-mm section thickness) at 120 kVp and 100 mAs (effective). As defined by Mahesh et al (18), an "effective milliampere-second" setting corresponds to the milliampere-second divided by the pitch, whereby "pitch" is defined by Silverman et al (19) as the ratio between the table feed per rotation and the x-ray beam width. In this case, the table feed was 15 mm per 0.5 second of scanner rotation (30 mm/sec), resulting in a pitch of 1.5:1. From the raw data of the acquisition, 3-mm-thick transverse sections were reconstructed with 1.5-mm increments. None of the patients received oral, rectal, or intravenous contrast material.
Immediately after multi–detector row CT, a radiologist read the unenhanced images at a clinical workstation with three-dimensional capabilities (Wizard; Siemens Medical Systems). The radiologist was asked to record whether the appendix was visible, to measure its outer transverse diameter (if seen), and to code the following findings as present or absent: (a) gas in the appendiceal lumen, (b) appendicolith, (c) periappendiceal fat stranding, (d) cecal wall thickening, and (e) abscess or phlegmon in the right iliac fossa. The presence of gas in the appendiceal lumen was considered as a possible negative criterion for acute appendicitis, whereas the other findings were considered as positive criteria. After separately coding each finding, the radiologist proposed an overall diagnosis of acute appendicitis or an alternative disease that could explain the patient's acute right lower quadrant pain. In case of doubt, there was no possibility to code the radiologist's level of confidence in the proposed overall diagnosis.
After the interpretation, the radiologist conducting the multi–detector row CT examination was free to acquire additional images after intravenous injection of the iodinated contrast material iopromide (Ultravist 370; Schering, Berlin, Germany). This occurred in 15 patients (16%). The interpretation with contrast material was not considered in the present study.
Radiologists and Blinding
For the purpose of this study, radiologists from our department were divided into two groups according to their expertise. Group 1 consisted of five board-certified radiologists with more than 10 years of experience in US and abdominal CT (M.Z., E.C., M.A.B., D.V.G.). Group 2 consisted of eight residents or general radiologists with 3–7 years of experience in radiology, including emergency imaging, but with no specific expertise in gastrointestinal imaging (C.K.).
Patients were examined by the radiologist responsible for the emergency examinations on that particular day and, consequently, were placed into that radiologist's corresponding group. In all patients, US and CT were performed within 1–2 hours of each other by separate radiologists from the same group, and each radiologist was blinded to the others' results.
Definite Diagnosis
In 40 patients, the definite diagnosis was made on the basis of the surgical findings. In the other 54 patients, the definite diagnosis was made on the basis of information from other diagnostic procedures (contrast material–enhanced multi–detector row CT, barium enema study, vaginal smear, colonoscopy with biopsy, and laboratory findings) obtained from medical charts during clinical follow-up and by telephone conversations conducted 1 month after the acute episode. The follow-up information was collected by one of the authors (C.K.). No patients were lost to follow-up.
Diagnostic Performance Measurements
Diagnostic performance was evaluated by considering the following measurements: sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of each finding and of the overall diagnosis suggested.
To evaluate the ability of US and CT to enable us to correctly rule out acute appendicitis or alternative diseases, we compared the frequency with which the normal appendix was not seen and no diagnosis of acute appendicitis or alternative disease was made with each technique. These examinations were considered inconclusive.
Statistical Analysis
Quantitative variables are expressed as the mean ± standard error of the mean.
For each group and for each imaging technique, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated for each finding and for the overall diagnosis of acute appendicitis. For each technique, the sensitivity and specificity of each finding were respectively compared between the two radiologist groups. The sensitivity, specificity, positive predictive value, and negative predictive value of the overall diagnosis of acute appendicitis were also compared between imaging techniques and between radiologist groups. To investigate a possible effect of BMI on the sensitivity and specificity of each imaging technique, patients were divided into three subgroups of BMI; each group contained the same number of patients.
The number of misclassifications of alternative diagnosis were determined and compared between US and CT for each group of radiologists and between groups of radiologists for each technique. To investigate a possible effect of patient sex, data from both groups of radiologists were pooled, and the proportions of misclassified alternative diagnoses were compared between sexes for each imaging technique and between techniques for each sex.
Proportion comparisons were made by using the Pearson 2 test or Fisher exact test when appropriate. Continuous variables were compared between groups with the Mann-Whitney U test.
Statistical significance for all tests was set at a P value of less than .05. The statistical software used was SPSS for Windows (release 11.5; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Definite Diagnosis
Acute appendicitis was diagnosed in 30 patients, and this was confirmed with microscopic examination of the surgical specimen. Among these patients, the proportion of male patients was significantly higher than that of female patients (17 of 35 male patients vs 13 of 59 female patients, P = .008). When the two groups of radiologists are compared, 11 of 43 patients from group 1 and 19 of 51 patients from group 2 had a definite diagnosis of acute appendicitis. The difference between these proportions was not statistically significant (P = .226).
Thirty-six patients (27 female and nine male patients) had various alternative diagnoses, as summarized in Table 1. Among these patients, the proportion of female patients was not significantly higher than that of male patients (27 of 59 female patients vs nine of 35 male patients, P = .079). One male patient had two simultaneous alternative diagnoses: acute cholecystitis and cecal volvulus. Twenty-eight patients (19 female and nine male patients) were considered to have nonspecific abdominal pain because their symptoms could not be elucidated with any diagnostic modality and resolved without any specific treatment. Among these patients, the proportion of female patients was not different from that of male patients (19 of 59 female patients vs nine of 35 male patients, P = .515).
|
Visualization and Diameter of the Appendix
In patients in whom the appendix was detected by the radiologist, the number of patients with acute appendicitis and the number of patients without acute appendicitis are listed for each imaging technique in Tables 2 and 3. The number of patients in whom the appendix was not detected did not differ significantly between radiologist groups, regardless of the imaging technique used or whether the patient had acute appendicitis (P values ranged from .474 to >.99). For patients with definite acute appendicitis, the frequency of visualization of the appendix was slightly higher with CT than with US, but this difference did not reach statistical significance (P = .340 and >.99 for groups 1 and 2, respectively) (Fig 1). For patients without definite acute appendicitis, the frequency of visualization of the appendix was significantly higher with CT (26 and 29 patients from groups 1 and 2, respectively) than with US (nine patients from each group) (P < .001) (Fig 2).
|
|
|
|
|
|
|
|
|
|
With US, no statistically significant difference in sensitivity or specificity was detected between radiologist groups for any of the coded findings (P values ranged from .126 to >.99), except for hyperemia within the appendiceal wall, which had a higher sensitivity in group 1 than in group 2 (P < .001).
With CT, there was also no statistically significant difference in sensitivity and specificity between radiologist groups for any of the coded findings (P values ranged from .148 to >.99), with the exception of the presence of gas in the appendiceal lumen, which had a higher specificity in group 1 than in group 2 (P = .041).
Misclassifications of Acute Appendicitis
Among the 94 patients, 15 findings were misclassified with US (seven false-negative and eight false-positive findings), and nine findings were misclassified with CT (four false-negative and five false-positive findings). The difference in the frequency of these misclassifications between US and CT was not statistically significant (P = .190). The number of false-positive and/or false-negative findings with each technique, according to radiologist group, is listed in Table 7. If we consider false-positive and false-negative findings separately, there was no statistically significant difference between US and CT (P = .351 and .389, respectively).
|
|
With regard to the effect of age, we did not detect any statistically significant difference in the mean age of patients whose findings were misclassified with US (39 years ± 3) or CT (39 years ± 4) (P = .934). Between groups of radiologists, there was also no statistically significant difference in the age of patients whose findings were misclassified by radiologists from group 1 (38 years ± 3) and the age of patients whose findings were misclassified by radiologists from group 2 (39 years ± 4) (P = .974).
Inconclusive Examination Findings with Each Technique
With CT, five of the 43 examinations (12%) performed by radiologists from group 1 and one of the 51 examinations (2%) performed by radiologists from group 2 provided inconclusive findings. With US, 15 of 43 examinations (35%) performed by radiologists from group 1 and 17 of 51 examinations (33%) performed by radiologists from group 2 provided inconclusive findings.
Within either group of radiologists, the frequency of examinations with inconclusive findings was significantly higher with US than with CT (P = .020 and P < .001 for groups 1 and 2, respectively).
Effect of BMI
For the patients in group 1 in whom the appendix was visualized, there was no statistically significant difference in BMI between patients with acute appendicitis (23.9 kg/m2± 4.8) and those without acute appendicitis (24.0 kg/m2± 5.9) (P = .977). For the patients in group 2 in whom the appendix was visualized, the BMI also did not significantly differ between diagnosis (23.6 kg/m2± 3.3 for patients with acute appendicitis and 24.6 kg/m2± 4.2 for patients without acute appendicitis) (P = .369).
The sensitivity and specificity of the overall diagnosis with both imaging techniques were compared between the three subgroups of BMI. There was no statistically significant difference with either US (P = .073–.799) or CT (P > .99).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Our results are consistent with those reported by Poortman et al (20), who, when comparing the graded compression technique of US with focused unenhanced CT, used an investigation protocol similar to ours to reflect the daily practice in a general community teaching hospital. Poortman et al reported no significant difference in sensitivity, specificity, or accuracy between these imaging techniques. Similar predictive values were also reported for each technique, although statistical significance was not detailed. Although we also found similar concordance between US and CT, for each technique we observed higher specificity, accuracy, and negative predictive values and slightly lower positive predictive values than did Poortman et al. Differences in predictive values could be explained by a higher rate of definite appendicitis in their study group than in ours (66% vs 32%, respectively). By applying a prevalence of acute appendicitis similar to that of Poortman et al (66%), however, we recalculated our positive and negative predictive values and obtained results very close to theirs. In contrast to our study and that of Poortman et al, however, two other studies (13,14) reported better performance with contrast-enhanced CT (by using enteric and/or intravenous administration of contrast material) than with US. The better performance of US in our study and in that reported by Poortman et al could be explained, at least in part, by the fact that US examinations were performed by technologists instead of radiologists (14).
It is traditionally believed that US is very operator dependent, necessitating a high level of skill and expertise, and that it is less accurate and reproducible than CT. In almost all hospitals, radiologic studies are interpreted on an emergency basis by fellows, radiology residents, and/or general radiologists; therefore, it is important to determine whether clinical decisions made on the basis of these images result in differences in patient care compared with those made by body radiologists (21). Thus, we investigated the potential effect of radiologist expertise by dividing our patients into groups according to the gastrointestinal imaging experience of the radiologist on duty. We observed that the diagnostic performance for diagnosing acute appendicitis (as well as alternative diseases) is not dependent on the gastrointestinal imaging expertise of a specific radiologist, regardless of the imaging technique used.
As shown in a study involving standard-dose single–detector row CT with oral and/or colonic contrast material (22), the sensitivity and specificity of CT findings vary—with an enlarged appendix with periappendiceal fat stranding having the highest sensitivity and specificity. In a previous study in which low-dose and standard-dose unenhanced multi–detector row CT were compared (23), we also observed that these two CT findings are the two most predictive signs, that is, the signs with the highest probability of a correct diagnosis of acute appendicitis. In the present study based on unenhanced multi–detector row CT, we observed values comparable with those previously reported with contrast-enhanced standard-dose single–detector row CT, except for a lower sensitivity for periappendiceal fat stranding. More important, there was no difference in diagnostic performances between any findings when radiologist expertise was considered (except for the presence of gas in the appendiceal lumen, which had a higher specificity when assessed by experienced radiologists). This was true for US as well, where we also found no difference in diagnostic performances between any findings when we considered radiologist expertise (except for hyperemia in the appendiceal wall, which had a higher sensitivity when assessed by experienced radiologists, which is suggestive of an influence of radiologist expertise in color Doppler US).
Of the patients classified as definitely not having acute appendicitis, the appendix was visible at US in 28% and at multi–detector row CT in 81%–91%, regardless of radiologist expertise. This wide difference between US and CT could mean CT enables the radiologist to confidently exclude the diagnosis of acute appendicitis. This is supported by additional findings in our study, whereby the frequency of an inconclusive examination is significantly lower with CT than with US, regardless of radiologist expertise. The low frequency of visualization of the normal appendix with US is consistent with that in previous reports, even in pediatric patients (5,13,14,24), although others have reported frequencies as high as 63%–82% (11,25,26).
In addition to their performance in establishing true-positive and true-negative diagnoses of appendicitis, imaging techniques should also be able to reveal alternative diseases. CT is usually believed to be more useful than US in the detection of abdominal conditions unrelated to acute appendicitis (13,14). Nevertheless, this is not supported by our results, as the proportion of misclassified findings, when compared with each possible definite alternative diagnosis, is not higher with US than with multi–detector row CT. This can be explained by the fact that radiologists involved in the present study had performed US examinations of the entire abdomen. In other studies, the US abdominal examination was focused only on the right lower quadrant.
The results of our study also show that patient sex and age do not have a significant influence on the performance of US and CT in the diagnosis of alternative diseases. Within each sex, there was no difference between US and CT in the misclassification of alternative diseases. Between sexes, there was no difference with US; however, there was a higher rate of misclassified alternative diseases with CT in female patients than in male patients, a finding that may reflect a worse performance of CT in the detection of gynecologic disorders. Although this finding conflicts with results of a previous study by Raman et al (27), who reported no difference between sexes in the detection of alternative diseases at CT, the difference could be explained by a lower frequency of gynecologic disorders in their series than in ours (10% vs 34%).
It is also traditionally believed that US is highly dependent on patient body habitus. In our study, however, the BMI did not have an effect on the visualization of the appendix or the diagnostic performance of either CT or US, regardless of radiologist expertise.
Our study has several limitations. First, if the patient did not undergo surgery, we had no absolute confirmation that he or she did not have acute appendicitis. Because this limitation applied equally to US and CT, however, there is no risk for bias. Second, we did not investigate radiologist confidence in the diagnostic proposal with each technique. The much higher frequency of visualization of a normal appendix with CT than with US (in patients without acute appendicitis), however, could convey a higher level of confidence with CT than with US in excluding acute appendicitis. Third, different patients were examined by each group of radiologists, but because the proportions of patients with definite appendicitis interpreted by each group of radiologists were not different, there is also no risk of bias. Fourth, only nine underweight patients and 11 obese or extremely obese patients, as classified according to the BMI categories recommended by the World Health Organization (15), were included in our study sample. The absence of BMI effect on the diagnostic performance of either technique could be explained by any limitation of US in obese patients being offset by the limitation of CT in underweight patients.
In conclusion, the results of this study suggest that the diagnostic performance of US does not differ from that of unenhanced multi–detector row CT in the detection of both acute appendicitis and alternative diseases, regardless of radiologist expertise in gastrointestinal imaging or patient sex, age, and body size. Nevertheless, examinations with inconclusive findings occur more frequently with US.
|
FOOTNOTES |
|---|
Abbreviations: BMI = body mass index
Authors stated no financial interest to disclose.
Author contributions: Guarantor of integrity of entire study, C.K.; study concepts and design, C.K., M.Z., D.V.G.; literature research, C.K., D.V.G.; clinical studies, C.K., M.Z., E.C., M.A.B., D.V.G.; data acquisition, C.K., M.Z., E.C., M.A.B., D.V.G.; data analysis/interpretation, C.K., V.D.M., P.A.G., D.V.G.; statistical analysis, V.D.M.; manuscript preparation, C.K., V.D.M., P.A.G.; manuscript definition of intellectual content, C.K., P.A.G., D.V.G.; manuscript editing, C.K., P.A.G.; manuscript revision/review, C.K., V.D.M., P.A.G., D.V.G.; manuscript final version approval, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. van Randen, S. Bipat, A. H. Zwinderman, D. T. Ubbink, J. Stoker, and M. A. Boermeester Acute Appendicitis: Meta-Analysis of Diagnostic Performance of CT and Graded Compression US Related to Prevalence of Disease Radiology, October 1, 2008; 249(1): 97 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Keyzer, S. Pargov, D. Tack, V. Creteur, P. Bohy, V. De Maertelaer, and P. A. Gevenois Normal Appendix in Adults: Reproducibility of Detection with Unenhanced and Contrast-Enhanced MDCT Am. J. Roentgenol., August 1, 2008; 191(2): 507 - 514. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Gaitini, N. Beck-Razi, D. Mor-Yosef, D. Fischer, O. B. Itzhak, M. M. Krausz, and A. Engel Diagnosing Acute Appendicitis in Adults: Accuracy of Color Doppler Sonography and MDCT Compared with Surgery and Clinical Follow-Up Am. J. Roentgenol., May 1, 2008; 190(5): 1300 - 1306. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. D. Bixby, B. C. Lucey, J. A. Soto, J. M. Theysohn, A. Ozonoff, and J. C. Varghese Perforated versus Nonperforated Acute Appendicitis: Accuracy of Multidetector CT Detection Radiology, December 1, 2006; 241(3): 780 - 786. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
