HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Riddell, A. M. Articles by Khalili, K. Search for Related Content PubMed PubMed Citation Articles by Riddell, A. M. Articles by Khalili, K.

Assessment of Acute Abdominal Pain: Utility of a Second Cross-sectional Imaging Examination¹

¹ From the Department of Medical Imaging, University of Toronto, University Health Network, Princess Margaret Hospital, 610 University Ave, Room 3-964, Toronto, ON, Canada M5G 2M9. Received October 20, 2004; revision requested December 13; revision received February 3, 2005; accepted March 1; final version accepted May 4. Address correspondence to K.K. (e-mail: korosh.khalili{at}uhn.on.ca).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
Purpose: To retrospectively determine how often a second cross-sectional imaging examination provides useful additional information or alters management of acute abdominal pain.

Materials and Methods: The research ethics board approved this study; the informed consent requirement was waived. Authors assessed imaging reports and clinical charts of adult patients who presented to the emergency department and underwent both computed tomography (CT) and ultrasonography (US) of the abdomen within 72 hours. A total of 255 patients fulfilled study criteria. The second examination was categorized as providing additional useful information, providing no additional useful information, or providing contradictory information. It was also noted whether the second examination was recommended because of findings of the first and whether the results of the second altered clinical management. Follow-up was available in 149 patients, and a definitive diagnosis was established. For these patients, it was determined whether either examination favored the correct diagnosis. Fisher exact test, one- and two-sample tests for equality of proportions with continuity correction, and the ² test were used, where appropriate.

Results: In 85 patients (33.3%), findings of the second examination agreed with those of the first examination and provided additional information. In 153 patients (60.0%), findings of the second examination agreed with those of the first examination and provided no additional information. In 17 patients (6.7%), findings of the second examination were contradictory to findings of the first examination. The percentage of follow-up CT examinations that provided no additional useful information was significantly lower when recommended by the radiologist (38%) than when recommended by someone else (72%, P < .001). The percentage of follow-up US examinations that provided no additional useful information was significantly lower when recommended by the radiologist (42%) than when recommended by someone else (74%, P = .003). In the 149 patients in whom a final diagnosis was available, both sets of scans were correct in 87 patients (58.4%); only the second set of scans was correct in 43 (28.8%). Overall, findings of the second examination led to a change or could have led to a change in treatment of 23 patients (9.0%).

Conclusion: A second examination is significantly more likely to be useful when performed because of radiologist recommendation.

© RSNA, 2006

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
Cross-sectional imaging of the acute abdomen with computed tomography (CT) or ultrasonography (US) is widely performed in current clinical practice because it has been proved to be an accurate tool in the assessment of many diseases. Once the initial imaging examination has been performed, however, a second cross-sectional examination may be ordered for various reasons. The findings on the first scans may have been negative or equivocal in the face of concerning symptoms or laboratory findings. The patient's symptoms or signs may have changed since the physical examination, or the examining physician may change, thus giving way to a physician who interprets the clinical findings differently. The radiologist who interpreted the first set of scans may have recommended that a second set of scans be obtained. These scenarios occur frequently in our clinical practice, but how often do patients undergo both CT and US for acute abdominal pain? Does it matter which examination is performed first? How often does the second examination add useful information or disagree with findings of the first examination? How often do findings of the second examination change the clinical management? Does patient sex make a difference in the usefulness of the second set of scans? What if the radiologist recommended that the second set of scans be obtained? These are questions that may be asked by the physicians involved in the care of the patient while contemplating obtaining a second set of scans.

Thus, the purpose of our study was to retrospectively determine how often a second cross-sectional examination provides useful additional information or alters treatment in patients with acute abdominal pain.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
This retrospective study was approved by the research ethics board of our institution; the need to obtain informed consent was waived.

Patient Population
The clinical charts and imaging reports of all adult patients who presented to the emergency department with abdominal symptoms and who underwent both CT and US in the acute setting between January 1, 2002, and December 31, 2002, were reviewed by one author (A.M.R.).

A total of 3648 patients presented to the emergency department at a tertiary care hospital, and 4009 cross-sectional imaging examinations of the abdomen and/or pelvis were performed—1927 (48.1%) with US and 2082 (51.9%) with CT. From this initial cohort, patients were selected for the study if (a) they underwent both CT and US within 72 hours and (b) the initial CT or US examination was ordered by an emergency department physician.

A total of 361 patients (9.9%) met the previously mentioned criteria. Patients were assessed by the authors and excluded from the study if (a) they had undergone surgical or radiologic intervention in the interval between the two imaging examinations, thus altering the clinical scenario; (b) the second examination was deemed "routine" in our institution as part of clinical management (eg, CT is performed routinely to stage newly diagnosed malignancy); and (c) the scans could not be directly compared with each other because they depicted different areas of the abdomen (eg, a triphasic liver CT scan and a pelvic US scan).

A total of 255 patients (6.99%) met these criteria and formed the study group. There were 132 (51.8%) female and 123 (48.2%) male patients aged 17–95 years (mean age, 49.0 years). The mean time (± standard deviation) between the two imaging examinations was 20.5 hours ± 20.

Imaging Report Analysis
For each patient, the imaging reports of the two examinations were retrospectively reviewed in consensus by an abdominal imaging fellow (A.M.R.) and staff abdominal radiologist with 4 years experience (K.K.). The imaging findings were categorized as follows: (a) the findings of the second examination agreed with those of the first examination and additional useful information was obtained, (b) the findings of the second examination agreed with those of the first examination but no further useful information was obtained, and (c) the findings of the second examination disagreed with those of the first examination.

At the time of this categorization, the radiologists were blinded as to whether the second examination was recommended by the radiologist who interpreted the first scan and to the patient's ultimate outcome. Additional useful information was defined as new finding(s) noted on the second scans that may further aid diagnosis. Equivalent findings on the CT and US scans, such as fluid and/or inflammatory stranding of mesenteric fat at CT and echogenic fat at US, were not deemed additional useful information on the second scans when they had been detected on the first scans. A report was also categorized as providing additional useful information when the list of possible diagnoses given in the first imaging report was narrowed to a single diagnosis or a single diagnosis was more strongly favored in the second imaging report.

In addition, with use of the imaging reports and electronic clinical charts we determined whether (a) the recommendation for the second examination was noted in the initial radiology report and, if so, the reason given for follow-up imaging; (b) both scans were read by the same radiologist (to test for bias in reporting new findings when the same radiologist read both scans); and (c) the findings on the second scan resulted in or could have resulted in a change in clinical management if they were acted on. A change in clinical management was defined as a situation in which the findings on the second examination were suggestive of a different diagnosis, the treatment of which was different than that given or favored in the first report. In addition, it was noted which department (medicine, surgery, or emergency) ordered the second examination.

Follow-up Data
For all patients, we reviewed the clinical charts, including surgical and clinical notes, pathologic findings, and follow-up radiology reports. In 149 of 255 patients (58%), an established clinical diagnosis that explained the patient's symptoms was documented. On the basis of this diagnosis, the imaging reports were reclassified as follows: both scans enabled the correct diagnosis, only the first scans enabled the correct diagnosis, only the second scan enabled the correct diagnosis, or neither scans enabled the correct diagnosis. A report was deemed correct if the patient's established clinical diagnosis was favored within the differential diagnosis or given as the definite diagnosis.

Imaging Techniques
All CT scans were obtained with a multi–detector row unit (Lightspeed QXi or Plus; GE Medical Systems, Milwaukee, Wis). The CT protocols, including the administration of intravenous, oral, and rectal contrast material, varied depending on the clinical information given on the request. All scans, however, were obtained with 5-mm-thick sections reconstructed at 2.5-mm intervals. US examinations were performed initially by technologists using HDI 5000 series units (ATL, Bothell, Wash), with all scans reviewed by residents, abdominal imaging fellows, and/or abdominal staff radiologists at the time of imaging, which often resulted in further directed examination. For both CT and US, all scans obtained on an emergency basis were reviewed by staff abdominal radiologists before definitive reporting.

Statistical Analysis
The Fisher exact test was used to compare the proportion of second scans that agreed with first scans but added no new information to the sum of the proportion of second scans that agreed with first scans and added additional information and the proportion of second scans that disagreed with first scans, as stratified by report recommendation. The Fisher exact test was also used to compare the proportion of second scans that alone enabled the correct diagnosis in patients with definitive follow-up according to whether the recommendation for follow-up imaging was noted in the radiology report. The one-sample test for equality of proportions with continuity correction was used in patients with follow-up data to test the statistical significance between the difference of the findings on the first and second scans. The two-sample test for equality of proportions with continuity correction was used to compare rates of second scans obtained and rates of second scans deemed correct with follow-up, stratified according to modality (US vs CT). The ² test was used to compare proportions of (a) second examinations with findings that agreed with those of the first examination and provided additional information, (b) second examinations with findings that agreed with those of the first examination but provided no additional information, and (c) second examinations with findings that disagreed with those of the first examination when results were stratified according to the examination that was performed first and to whether the same radiologist interpreted the scans. Differences were considered statistically significant if the P value was less than .05. We used statistical software for analysis (S-Plus, version 6.1; Insightful, Seattle, Wash).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 
During the 1-year study period, 159 of 1927 patients (8.2%) who underwent US subsequently underwent CT, and 96 of 2082 patients (4.6%) who underwent CT subsequently underwent US. The likelihood of CT being performed in patients who initially underwent US was higher than the reverse (P < .001, two-sample test for equality of proportions with continuity correction; 95% confidence interval: 0.02, 0.05). The second examinations were recommended by the radiologist in 97 of 255 patients (38.0%), members of the emergency department in 92 (36.1%), members of the department of medicine in 49 (19.2%), and members of the department of surgery in 17 (6.7%).

The imaging reports (Table 1) revealed that the second examination provided additional information in only 85 of the 255 patients (33.3%) and that the findings of the second examination disagreed with findings of the first examination in 17 (6.7%). In 153 of the 255 patients (60.0%), the findings of the second examination agreed with those of the first examination but provided no additional useful information. The results did not significantly differ when stratified according to which scans were obtained first (P = .60, ² test).

View larger version (24K): [in this window] [in a new window] [Download PPT slide]   Figure 1: Chart shows the percentage of patients who underwent US followed by CT in whom findings of the second examination agreed with findings of the first examination and added additional useful information (category A), findings of the second examination agreed with those of the first examination but provided no additional information (category B), and findings of the second examination disagreed with those of the first examination (category C). Data are stratified according to whether CT was recommended in the US report.

View larger version (24K): [in this window] [in a new window] [Download PPT slide]   Figure 2: Chart shows the percentage of patients who underwent CT followed by US in whom findings of the second examination agreed with those of the first examination and added additional useful information (category A), findings of the second examination agreed with those of the first examination but provided no additional information (category B), and findings of the second examination disagreed with those of the first examination (category C). Data are stratified according to whether US was recommended in the CT report.

Follow-up Data
For the 149 patients in whom clinical follow-up helped establish a final diagnosis, findings of both examinations were deemed correct in 87 patients (58.4%) and only findings of the second examination were deemed correct in 43 (28.8%) (Table 5). Findings of the initial examination alone were correct in six of the 149 patients (4.0%). In 13 of the 149 patients (8.7%), neither examination enabled the correct diagnosis. In patients who underwent US followed by CT, there was disagreement between findings on the scans in 32 of 83 patients (38%). In 29 of these 32 patients (91%), the CT scans, being the second set of scans obtained, were correct (P < .001, one-sample test for equality of proportions with continuity correction). In patients who underwent CT followed by US, there was disagreement between findings on the scans in 17 of 66 patients (26%). In 14 of these 17 patients (82%), the US scans, being the second set of scans obtained, were correct (P = .015, one-sample test for equality of proportions with continuity correction). Therefore, in patients in whom the ultimate diagnosis was determined and in whom there was disagreement between findings on the two sets of scans, the findings of the second examination (be it US or CT) were correct in most patients (ie, 82.3% for US, 90.6% for CT). The difference between the two rates was not statistically significant (P = .70, two-sample test for equality of proportions with continuity correction).

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   Figure 3: Chart shows the percentage of patients with an established clinical diagnosis in whom the diagnosis was made on the basis of both examinations, the second examination only, the first examination only, or neither examination. Data are stratified according to whether the recommendation to perform a second examination was noted on the first examination report.

Stratification of Results according to Radiologist
The results were categorized on the basis of whether the two sets of scans were read by the same radiologist. The differences between the categories (Table 6) were not statistically significant (P = .50, ² test). The largest difference was in cases in which findings of the second examination agreed with findings of the first examination but no new information was obtained: There was an 8% (95% confidence interval: –6%, 22%) higher frequency of findings of second examinations agreeing with findings of first examinations but not yielding further information when the two sets of scans were read by different radiologists.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

In an ideal clinical practice, the referring clinician and radiologist would consult with each other regarding the ordering of costly imaging examinations and their findings. Seltzer et al (1) showed that consultation before CT of the abdomen is ordered leads to an increase in the clinical information provided and a decrease in additional diagnostic examinations. Other investigators (2,3) showed that the interpretation of imaging studies with the correct clinical information can lead to improved sensitivity and accuracy. In a community hospital setting, Bova and Villalobos (4) analyzed 93 patients who were referred for multiple imaging tests before the review of any test results. They found that 60% of the tests were determined to be inappropriate and 79% of the tests were negative, with no significant difference between referrals made by primary care physicians and specialists. From a medicolegal point of view, previous court rulings have emphasized the importance of consultation with a radiologist, including the provision of correct clinical history and the selection of imaging modalities (5).

With advances in communication technology (most requests are now made electronically) and shrinking healthcare budgets, the interaction and consultation between the emergency and imaging departments is of utmost importance and should be simplified. The results of our study, which emphasize the value of the radiologist's recommendation with respect to obtaining a second set of scans, argue for a remodeling of current ordering and reporting practices. This is especially important when there is a discrepancy between the clinical and laboratory findings and the imaging report. Standardized reports that document the recommendation of additional imaging and mandatory consultation before additional examinations are ordered for the same indication are possible solutions.

Although the percentage of second scans that provided useful clinical information was significantly higher when they were obtained on the basis of a radiologist's recommendation, it is important to note that a sizable minority of second scans were still deemed to provide useful additional information when there had not been a specific recommendation for follow-up imaging in the initial report. When performed as the second examination but not recommended in the report of the first examination, 22% of CT findings and 25% of US findings were still deemed useful. Therefore, we emphasize that the imaging report cannot be interpreted independently of the clinical and laboratory findings. The consultation process between the referring clinicians and radiologists therefore should be a two-way exchange of information that leads to informed and shared decision making. A "gatekeeper" approach, in which the responsibility of the decision to perform the second examination is shifted to the radiologist, may also lead to insufficient imaging of patients.

On the basis of the criteria described in Materials and Methods, a second examination was considered useful if the findings enabled the differential diagnosis given in the first imaging report to be narrowed to strongly favor a single diagnosis or to provide a definitive diagnosis on the second imaging report. As imagers, we acknowledge that sometimes a second examination, although it may not reveal new findings, is useful nonetheless. This may occur when the finding discovered on or diagnosis entertained with the first scans is subtle enough or a proposed therapy serious enough to warrant a confirmatory second examination. We refer to this as "improved diagnostic confidence" provided by the second examination, and we tried to indirectly account for it with the given definition of a useful second examination. Despite this wider definition of a useful second examination, most second examinations were not deemed to provide additional useful information. We do not believe that alteration of clinical management alone is a valid method for evaluating the usefulness of the second examination. We would further emphasize that although only a small minority (9.0%) of second scans led to or could have led to an alteration of clinical management, many did result in the initiation or maintenance of correct therapy by virtue of improved diagnostic confidence.

The reasons why the radiologist who interpreted the first scans requested a follow-up examination were dependent on the modality. Follow-up CT was requested most commonly when an appendix was not visualized in the face of concerning US findings, when the image quality of the scan was poor, and when hydronephrosis was noted but the cause of obstruction was not clear. A follow-up US examination was requested most commonly when adnexal masses needed to be characterized, when acute cholecystitis was in question, to search for gallstones, and to further assess hepatic lesions. Because of the retrospective nature of the study, we cannot accurately determine the reason why clinicians requested a second examination.

In a similarly designed study, Harvey and Miller (6) retrospectively reviewed radiologic reports in patients with right upper quadrant pain who underwent both CT and US of the abdomen. They found that US was more accurate than CT as the first imaging investigation and was much more likely to be contributory to the diagnosis than CT as the second examination. In our study, there was no significant difference in the overall ability of the second examination—be it CT or US—to provide useful additional information. This was even so in the female patients, in whom US is recommended over CT in the evaluation of pelvic organs in the acute setting. We believe, however, that there is no contradiction between our results and those of the previous study. Each modality has its strengths with respect to depiction of various abnormalities within the abdomen; however, by combining all indications—as in our study—the differences diminish. Furthermore, in both studies, the population analyzed was a selective one that was likely to be a diagnostic challenge for the referring clinicians because the patients underwent two imaging examinations.

The rate of second scans providing additional information was higher when both scans were read by the same radiologist than when scans were read by different readers. This trend, although not statistically significant, can have several possible explanations. The radiologists who analyzed both scans could have been biased toward justifying the recommendation of the second scan. Alternatively, the radiologists may have obtained further knowledge of the patient's other clinical findings by means of increased interaction with the patient and referring clinicians.

Our study has limitations. Its retrospective nature introduces several factors that can lead to under- or overestimation of the value of the radiologist's recommendation. We have presumed that the study report accurately reflected the recommendation passed on to the emergency department. It is possible that, on occasion, a radiologist may have verbally recommended a second examination but did not include this recommendation in the report. In addition, a radiology trainee may have recommended the second examination in a preliminary and/or verbal report, whereas the final reporting staff radiologist may have considered it unnecessary and therefore did not document the recommendation in the final report. Because there was no consistent documentation of verbal reports given to the emergency department, it is not possible to determine whether the decision for the patient to undergo repeat imaging was made in light of a final report of the first imaging examination in all cases. In patients whose scans were interpreted by two different radiologists, it is also not known whether the findings from the first examination were reviewed by the radiologist who interpreted findings of the second examination. These uncertainties, however, do not alter the key finding of our study: that the rate of diagnostic information being obtained from the second examination was significantly higher when it was recommended by the radiologist who interpreted findings of the first examination.

Our study could have been performed in a prospective fashion, with the referring clinicians documenting the reason for obtaining the second set of scans and the radiologists documenting their exact communications and recommendations. This would not necessarily have been a more accurate study, however, because the knowledge of future scrutiny of practice could alter the practice itself, which is known as the Hawthorne effect (7). There was no change in the practice pattern of referring clinicians as a result of our retrospective study design.

Another potential weakness of our study is the subjectivity in the interpretation and analysis of reports. Because we do not use structured reporting at our institution, reported findings, their severity, and entertained diagnoses were worded in many different ways, which inevitably leads to variability in their interpretation. We tried to minimize this factor by analyzing each report in consensus and being consistent in the analysis while remaining blinded as to whether the second examination was recommended by the radiologist who interpreted the first scans.

In conclusion, we have shown that, for patients who present with acute abdominal pain, a second cross-sectional examination was significantly more likely to be useful when it was recommended by the radiologist. Our results emphasize the importance of consultation between clinicians and radiologists before additional imaging investigations are performed in the acute abdomen.

	ACKNOWLEDGMENTS

 
The authors thank and acknowledge the aid of Mostafa Atri, MD, FRCPC, for valuable input regarding study design, and George Tomlinson, PhD, for statistical analysis.

	FOOTNOTES

 
Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, A.M.R., K.K.; study concepts/study design or data acquisition or data analysis/interpretation, A.M.R., K.K.; manuscript drafting or manuscript revision for important intellectual content, A.M.R., K.K.; approval of final version of submitted manuscript, A.M.R., K.K.; literature research, A.M.R., K.K.; clinical studies, A.M.R., K.K.; and manuscript editing, A.M.R., K.K.

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION References

 

1. Seltzer SE, Beard JO, Adams DF. Radiologist as consultant: direct contact between referring clinician and radiologist before CT examination. AJR Am J Roentgenol 1985;144(4):661–664.[Abstract/Free Full Text]
2. Mullins ME, Lev MH, Schellingerhout D, Koroshetz WJ, Gonzalez RG. Influence of availability of clinical history on detection of early stroke using unenhanced CT and diffusion-weighted MR imaging. AJR Am J Roentgenol 2002;179:223–228.[Abstract/Free Full Text]
3. Leslie A, Jones AJ, Goddard PR. The influence of clinical information on the reporting of CT by radiologists. Br J Radiol 2000; 73(874):1052–1055.
4. Bova JG, Villalobos LB. Utilization review of simultaneously ordered multiple radiologic tests for the same symptom. Am J Med Qual 1998;13(2):81–84.[Abstract/Free Full Text]
5. Mozumdar BC, Jones G. Medico-legal issues in radiological consultation. Radiol Manage 2003;25(5):40–43.[Medline]
6. Harvey RT, Miller WT Jr. Acute biliary disease: initial CT and follow-up US versus initial US and follow-up CT. Radiology 1999;213:831–836.[Abstract/Free Full Text]
7. Fletcher RH, Fletcher SW, Wagner EH. Clinical epidemiology: the essentials. 2nd ed. Baltimore, Md: Williams & Wilkins, 1988; 137.

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Riddell, A. M. Articles by Khalili, K. Search for Related Content PubMed PubMed Citation Articles by Riddell, A. M. Articles by Khalili, K.