Introduction ORIGINAL ARTICLE. Steve Y. Lee. Bret Coughlin. Jeannette M. Wolfe. Joseph Polino. Fidela S. Blank. Howard A.

Transcription

1 Emerg Radiol (2006) 12: DOI /s z ORIGINAL ARTICLE Steve Y. Lee. Bret Coughlin. Jeannette M. Wolfe. Joseph Polino. Fidela S. Blank. Howard A. Smithline Prospective comparison of helical CT of the abdomen and pelvis without and with oral contrast in assessing acute abdominal pain in adult Emergency Department patients Received: 11 July 2005 / Accepted: 16 December 2005 / Published online: 21 April 2006 # Am Soc Emergency Radiol 2006 Abstract Purpose: This prospective study compares the agreement of nonenhanced helical computed tomography (NECT) with oral contrast-enhanced computed tomography (CECT) in Emergency Department (ED) patients presenting with acute abdominal pain. Materials and methods: One hundred eighteen patients presenting to the ED with acute abdominal pain undergoing CT were enrolled over a 13-month period using convenience sampling. Exclusion criteria included acute trauma, pregnancy, unstable patients, and patients suspected of having urinary calculi. Patients were scanned helically using 5-mm collimation before and approximately 90 min after oral contrast administration. Both exams were prospectively interpreted by different attending radiologists in a blinded fashion using an explicit data sheet specifying the presence or absence of 28 parameters relating to various common diagnoses. Results: The 118 patients had a mean age of 49 years, a male: female ratio of 7:13, and a median height, weight, and BMI of 166 cm, 80 kg, and 29, respectively. The most common indications for the study included appendicitis This work was partially presented at the RSNA in Chicago 2004 S. Y. Lee. B. Coughlin (*). J. M. Wolfe. J. Polino. F. S. Blank. H. A. Smithline Department of Radiology, Baystate Medical Center, 759 Chestnut Street, Springfield, MA 01109, USA Bret.Coughlin@bhs.org Tel.: Fax: Steve.Lee@bhs.org S. Y. Lee. B. Coughlin. J. M. Wolfe. J. Polino. F. S. Blank. H. A. Smithline Department Emergency Medicine, Baystate Medical Center, 759 Chestnut Street, Springfield, MA 01109, USA (32%) and diverticular disease (12%). Pain maximally localized to the right lower quadrant in 37% and the left lower quadrant in 21%. There were 21 patients that had significant disagreement of interpretations between NECT and CECT resulting in a simple agreement of 79% (95% CI: 70 87%). For specific radiologic parameters, agreement ranged from 77 to 100%. A post hoc agreement analysis was subsequently performed by two radiologists and only five paired scans were identified as discordant between the NECT and CECT. For only one of these patients did both radiologists agree that there was a definite discordant result between the two studies. A final unblinded consensus review demonstrated that much of the disagreement between the interpretations was related to interobserver variation. Conclusion: There is 79% simple agreement between NECT and CECT in diagnosing various causes of acute abdominal pain in adult ED patients. Post hoc analysis indicates that a significant portion of the discordance was attributable to interobserver variability. This data suggests that NECT should be considered in adult ED patients presenting with acute abdominal pain. Keywords CT. Oral contrast. Abdominal pain. Technique. Emergency Introduction Helical CT has become the imaging modality of choice for diagnosing many causes of abdominal pain. Helical CT has a high sensitivity and specificity in diagnosis of abdominal pain and has become established in the evaluation of some gastrointestinal (appendicitis) and urologic (stone disease) conditions [1 6]. The increasing volume of Emergency Department (ED) patients [7] and potential seriousness of delayed or incorrect diagnosis in the ED make optimizing techniques for ED CT of the abdomen and pelvis important. Various protocols have been developed for studying ED

2 151 patients utilizing intravenous, oral, and rectal contrast agents [8 11]. Helical CT without any contrast (NECT) has been reported to be accurate for diagnosing or excluding acute appendicitis [12, 13]. Patients who present to the ED with acute abdominal pain spend disproportionately more time before disposition than other groups of ED patients [14]. The use of oral contrast for acute abdominal pain has not been thoroughly studied. Oral contrast has been continued since the early stages of development of CT (slower data acquisition). Oral contrast helped delineate soft tissue planes between intraperitoneal structures. Specifically, oral contrast was invaluable in defining the various segments of the duodenum from the pancreas, small bowel from mesenteric lesions as well as occasionally helping delineate intraperitoneal lesions from retroperitoneal one. Because dilute gastrograffin was commonly used, antispasmotic agents were also administered to minimize bowel motility [15, 16] The introduction of helical scanning and more recently multidetector devices has made the utility of oral contrast uncertain. The literature on the value of oral contrast for newer technology scanners is not conclusive. This prospective study was designed to assess the results of helical scanning NECT and CT with oral contrast (CECT) in 100 ED patients presenting with abdominal pain. Materials and methods The institutional review board for investigation of human subjects approved the study. This prospective observational study was conducted on a convenience sample of adult ED patients presenting with abdominal pain to an academic medical center. Convenience sampling was based on the availability of a research assistant and availability of a CT scanner. Due to the high demand for the medical center s CT scanners, enrollment was limited to two patients per weekday during daytime hours. Patients were approached for enrollment if they had an abdominal pelvic CT ordered by an ED provider for acute abdominal pain. Patients who had trauma, were suspected renal colic, were pregnant, need for IV contrast, or were clinically unstable were excluded. After informed consent, patients underwent an immediate helical CT of the abdomen and pelvis (Picker MX2000 or Picker MX 8000, Picker International Mansfield OH now Philips Electronics of NA, NY, NY) using 5 mm collimation, pitch 1.5, 130 kvp and 150 ma, and no contrast (NECT). The patients were then immediately given a standard oral contrast regimen of two 15 oz drinks from a mixture of 15 cc iothalamate meglumine 60% (Conray 60%, Mallinckrodt, St. Louis, MO, USA) and fruit juice or water. The patients were asked to drink 15 oz initially and then another 15 oz just before being placed on the scanner. The CT was repeated after 90 min with identical scanning parameters (CECT). The scan pairs were interpreted by different attending radiologists who had several years of experience. All department radiologists who routinely interpret ED abdomen and pelvis CT participated (n=19). Interpretations were performed using explicit data sheets that evaluated 28 different CT findings that are associated with six common abdominal/pelvic pathologic conditions. Each scan was interpreted immediately after performance and the radiologists did not have access to the matched scan or the interpretation of the other scan. The NECT was interpreted without any history. Agreement between the matched NECT and CECT was measured by the kappa statistic. All CT interpretations were sent to a panel consisting of two attending radiologist physicians and two attending ED physicians who individually determined if the scan interpretations were significantly discordant. A final determination was made by consensus agreement. A significant difference in interpretation was defined as a difference that would require different patient care/ management either in the ED or on follow-up. In addition, a post hoc analysis was done to better understand the discordant interpretations. All patients who had clinically significant discordant interpretations had their matched scans reviewed and separated. These scans were then randomly mixed with NECT and CECT pairs of 20 additional patients in the study population who had concordant interpretations. All of these paired studies were blindly reviewed by two radiologists using the same data collection sheet. Discordant interpretations were reviewed again in the same fashion as described above. Agreement was measured using the kappa statistic. Medical studies generally consider agreement as follows; kappa of <0.20 poor, fair, moderate, good, very good, and excellent [17, 18]. Finally, the images and data sheets were reviewed unblindly by two radiologists to determine whether the reason for discrepancy could be determined and, if so, whether it was related to scanning technique (specifically, presence of oral contrast), patient factors, or interobserver disagreement. The study design was to compare the techniques and not to determine the accuracy of the exams. Results One hundred nineteen patients were approached over a 13-month study period. Of these, 118 gave informed consent and 100 completed the protocol 18 patients did not complete the study, 11 patients could not be imaged due to unavailability of scanner due to high census, research patients inadvertent given oral contrast and patient refusal to continue the study protocol. Demographic data of the enrolled paients who completed the protocol are shown on Table 1. The most common locations for maximal tenderness were; right lower quadrant (37%), left lower quadrant (21%), and nonfocal tenderness (14%). The remaining patients had maximal tenderness distributed evenly throughout the remaining sections of the abdomen and pelvis. The indication for the study was documented for 62 patients. The most common clinically suspected diagnoses were appendicitis and diverticulitis.

3 152 Table 1 Demographics Number Age (years) Height (cm) Weight (kg) Body mass index 29 7 Abdominal girth (cm) Gender Female 65% Male 35% Race Caucasian 64% Hispanic 26% Black 9% Asian 1% SD values represent mean standard deviation For 21 patients, there was clinically significant disagreement between the CECT and NECT interpretations yielding a simple agreement of 79% (95% CI: 70 to 87%) and kappa of 0.58 (95% CI: 0.06 to 0.72). Of these 21 patients, 11 had a normal NECT reading and an abnormal CECT reading (group 1), six had an abnormal NECT reading and a normal CECT interpretation (group 2), and four had an abnormal NECT and CECT but were felt to be significantly different from each other (group 3). The scan interpretations for these patients are listed in Table 2. Agreement, with 95% confidence intervals, between NECT and CECT on specific radiologic findings are listed Table 2 Diagnoses of discordant scans original readings Group 1 (N=11) Group 2 (N=6) Group 3 (N=4) CECT reading indeterminate Colitis (diverticulitis vs ischemia) Duodenitis Ovarian mass Sigmoid wall thickening Subcapsular hematoma CECT reading was normal Gastric tumor Ovarian mass Periumbilical hernia NECT reading NECT reading was normal SD Cervical Mass Duodenitis Iliopsoas mass unclear etiology Pancreatitis Cholelithiasis Iliopsoas hematoma in Table 3. Interpretations of the kappa values should take into account the percentage of abnormal readings. The different radiologists were not able to consistently quantify the level of confidence in their interpretations. Post hoc analysis is listed in Tables 4 and 5. A total of five patients were thought to have discordance between the NECT and CECT by either radiologist on post hoc review. Unblinded consensus review of all discrepant scan pairs by two radiologists determined that there were 20 discrepancies between the CECT and NECT (One pair was found to have no significant discrepancy when reviewed with the scans and data sheets simultaneously). Two sets of scans were felt to be discrepant primarily related to oral contrast protocol (Figs. 1 and 2), 18 were primarily related to interobserver variability (Figs. 3 and 4). Patient factors such as obesity were a factor in some of the difficulty in interpretation but did not differ between scans (Fig. 5). Most cases were similarly interpreted (Figs. 6 and 7). Discussion CT technology has rapidly evolved. The focus of early CT imaging of the abdomen and pelvis was often mitigating the degradation of image quality due to motion. The long image acquisition times combined with respiratory and peristaltic motion necessitated the use of all possible methods to improve bowel opacification and improve conspicuity of pathology. Most protocols used to evaluate the abdomen and pelvis included high volume oral contrast administration over a period long enough to opacify the small bowel. Some institutions advocated for rectal and IV contrast as well. Little study was given to nonenhanced imaging on these relatively slow scanners because the cost of oral was low and the direct adverse effects to the patients were essentially nil. CT technology has evolved and image acquisition is so rapid that respiratory and peristaltic motions have been dramatically decreased. Concurrently, the experience of imagers has advanced and many now feel comfortable to a greater or lesser degree with interpreting scans without contrast. Adding impetus to the development of new protocols is the current climate of rising health care costs and concern about optimal resource utilization [19]. The need to maximize resources focuses investigators on the development of protocols which might streamline the flow of patients through busy and expensive EDs and is expected to only intensify for the foreseeable future [20, 21]. The results show that there is moderate agreement between the interpretation of NECT (which would save at least 90 min of time) and CECT for adult patients who present for CT via the ED at a single institution. It also demonstrates that there are disagreements between CECT and NECT. The reason for the disagreement is not just the technique of bowel opacification. The post hoc analyses suggest that the discrepancy is largely due to interobserver variation. The interobserver variation demonstrated is consistent with several studies that confirm different observers will

4 153 Table 3 Agreement prospective readings Parameter Percentage abnormal Simple agreement (%) 95% CI Kappa 95% CI Appendix Inflammation Wall thickening Free fluid Appendicolith Distention Abscess Large bowel Diverticulum Inflammation Free fluid Wall thickening Gall bladder Gallstones Wall thickening * Free fluid Inflammation * * Distention * * Small bowel Wall thickening Free fluid Inflammation Distention Abscess Stricture/transition point Pancreas Inflammation Abscess * Free fluid Ovaries/uterus Abscess Free fluid Mass Inflammation Percentage abnormal the percentage of scans where either or both are abnormal *Insufficient data to perform calculation Table 4 Diagnoses of discordant scans post hoc analysis CECT reading NECT reading Group 1 Colitis (diverticulitis vs ischemia) NECT reading was normal Duodenitis Group 2 CECT reading was normal Ureteral calculi Appendicolith Table 5 Overall agreement between radiologists and between scans Between radiologist #1 and #2 for NECT Between radiologist #1 and #2 for CECT Between NECT and CECT for radiologist #1 Between NECT and CECT for radiologist #2 Simple agreement (%) Interquartile (%) Kappa 95% CI

5 154 Fig. 1 Discordance due to technique. Axial images through the region of the pancreatic head and descending duodenum in a 60-year-old male presenting with epigastric pain. The NECT (a) interpreted as no significant findings and the CECT (b) interpreted as showing duodenal wall thickening (arrow) Fig. 2 Discordance due to technique. Magnified axial CT images through region of appendix and cecum. NECT (a) interpreted as probable appendicitis due to thickening and possible stranding (arrow) around appendix and cecum. CECT (b) interpreted as interpret studies differently in an amount which varies by how the study is designed [22, 23]. Factors that may influence the interobserver variation include level of experience, level of training and amount of clinical information available, and definition of discrepancy. Studies have identified variation of interpretation between 1 and 38% for studies including head CT, trauma, urologic disease, CT of the abdomen and pelvis, chest CT, and computed tomography angiography [24 31]. The interobserver variation confounds attempts to discern the actual influence of oral contrast on emergency CT of the abdomen probably normal but rectal contrast was recommended (c) which demonstrated normal appendix (shortarrow) filled with contrast and contrast distended cecum without deformity (long arrow) and pelvis. The retrospective review suggests the impact of oral contrast is minor. There are several reasons for expecting nearly equivalent level of diagnostic accuracy of NECT and CECT for many causes of abdominal pain in the ED population. Solid visceral abnormalities should not have significant differences related to the presence or absence of oral contrast. IV contrast is known to improve the detection of focal lesions in solid viscera. This is not a frequent cause for acute abdominal pain and, if symptomatic, would likely be advanced enough to be detected without IV contrast. Inflammatory conditions often manifest with abnormalities Fig. 3 Discordance primarily related to interobserver variability of a 48-year-old female with RLQ pain. NECT (a) and CECT (b) through similar levels of ileo cecal valve and appendix region demonstrate similar findings of appendicolith (short arrows) and appendiceal thickening (short arrows) with mild surrounding stranding. The CECT was interpreted as appendicitis and the NECT interpretation did not include appendicitis as highly likely

6 155 Fig. 4 Discordance primarily related to interobserver variability of the CT of the upper abdomen in a 41-year-old female with abdominal pain. NECT (a) interpreted as some fluid and unable to exclude appendicitis. Gastric wall thickening (arrow), mesenteric studding, and omental disease not noted. CECT (b) demonstrates gastric wall thickening (arrow). Omental studding and mesenteric disease was present on lower images on both NECT and CECT. Gastric cancer with carcinomatosis diagnosed at exploration 7 days later Fig. 5 Patient-related factors contributing to difficulty in interpretation and interobserver variability of a 54-year-old obese female with abdominal pain. NECT (a) demonstrates artifact-limiting evaluation of pelvis. Stranding around sigmoid (arrows) felt >95% likely due to diverticulitis. CECT (b) through pelvis also demonstrates significant artifact felt to be most likely normal (5 24% likelihood of diverticulitis) Fig. 6 Concordance between NECT (a) and CECT (b) in 80- year-old female with acute RLQ pain. Images through the lower abdomen interpreted as demonstrating thickening and stranding (arrows) on NECT and CECT Fig. 7 Concordance between NECT (a) and CECT (b) through the mid-abdomen demonstrate dilated fluid-filled loops of small bowel (short arrows) and decompressed large bowel (long arrow) interpreted similarly

7 156 in the fat of the peritoneal cavity and omentum. The stranding and hazy fat changes aid localization of pathology and should be detectable with or without oral contrast. The presence of oral contrast might lead to a higher level of confidence in the diagnosis. Experience would be expected to allow higher accuracy over time using a NECT protocol. Abscesses should frequently be detectable without oral contrast. Bowel wall pathology should be better detected when manifesting as only thickening if effort is made to distend the bowel; however, the detection of pneumatosis should not be impacted by oral contrast. Note that most oral contrast regimens do not distend the colon. Direct visualization of wall edema would likely be more affected by the presence of IV contrast and an ideal protocol might include the use of IV contrast for the age group most likely to be affected by bowel ischemia. Gynecologic pathology can be challenging to diagnose and CT using most protocols will detect large abnormalities but the tissue contrast capability of U/S makes it a useful test for gynecologic diseases. A NECT can be as accurate as CECT in the diagnosis of many etiologies of abdominal pain. A retrospective review suggests that discrepant readings were related to interobserver variability and technique with interpretations affected by patient factors such as habitus. The results of this study support the further investigation of use of NECT for evaluation of adult patients presenting to the ED for abdominal pain. Interobserver variability should be addressed by training and ongoing quality assurance programs to detect systematic differences. One possible protocol, which is currently being compared to CECT, is the performance of CT with rectal contrast immediately administered. The administration of IV contrast would likely improve detection of bowel wall pathology and with the introduction of multidetector devices improve detection of vascular diseases. Such protocols could improve the care of ED patients by decreasing the time to diagnosis and disposition and by expediting the initiation of the appropriate treatment pathway. Another possible improvement is the potential cost savings from a more streamlined diagnostic evaluation of this often-vexing patient population. The current study is limited by several factors. There is no true final diagnosis possible in many cases that did not go to surgical exploration or definitive follow-up. However, the purpose of the study was to determine the relative agreement and compare two protocols. As described, there are possible confounding factors for why there were disagreements between CECT and NECT other than oral contrast. These include type of pathology, observer differences, patient factors such as amount, and distribution of fat. Current research is being conducted to assess these possibilities. Due to patient care factors, the amount of clinical information known to some observers, prospectively, may have differed and could influence interpretations (e.g., house staff may have provided information when attempting to learn results of test). Logistical factors limited sampling to two cases per daytime weekday shift. Explicit data sheet collection may not completely reflect actual interpretations. The usefulness of a CT protocol which does not require a long oral contrast protocol is not established. Previous research has shown that there is a significant difference in the time a patient spends in the ED when presenting with abdominal pain requiring CT of abdomen and pelvis. There are possible cost savings and patient satisfaction issues, which were not specifically studied and which are currently being prospectively evaluated. In conclusion, there is a 79% simple agreement between NECT and CECT for adult ED patient evaluation of abdominal pelvic pain. The contribution of the oral contrast protocol to the disagreement appears to be much less than the degree of interobserver variability and is the topic of ongoing study. References 1. Bendeck SE, Nino-Murcia M, Berry GJ, Jeffrey RB (2002) Imaging for suspected appendicitis: negative appendectomy and perforation rates. Radiology 225: Raman SS, Lu DSK, Kadell BM, Vodopich DJ, Sayre J, Cryer H (2002) Accuracy of nonfocused helical CT for the diagnosis of acute appendicitis: a 5 year review. Am J Roentgenol 178: Birnbaum BA, Wilson SR (2000) at the Millenium. Radiology 215: Smith RC, Rosenfield AT, Choe KA et al (1995) Acute flank pain: comparison of non contrast enhanced CT and intravenous urography. Radiology 194: Portis AJ, Sundaram CP (2001) Diagnosis and initial management of kidney stones. Am Fam Physician 63: Akbar SA, Mortele KJ, Baeyens K, Kekelidze M, Silverman SG (2004) Multidetector CT urography: techniques, clinical applications, and pitfalls. Semin Ultrasound CT MR 25: National Hospital Ambulatory Medical Care Survey, Emergency Department Summary for U.S. Department of Health and Human Services, Center for Disease Control and Prevention, National Center for Health Statistics 8. Kamel IR, Goldberg SN, Keogan MT, Rosen MP, Raptopolous V (2000) Right lower quadrant pain and suspected appendicitis: nonfocused appendiceal CT review of 100 cases. Radiology 217: Malone AJ, Wolf CR, Malmed AS, Melliere BF (1993) Diagnosis of acute appendicitis: value of unenhanced CT. AM J Roentgenol 160: Lane MJ, Katz DS, Ross BA et al (1997) Unenhanced helical CT for suspected acute appendicitis. Am J Roentgenol 168: Wise SW, Labuski MR, Kasales CJ et al (2001) Comparative assessment of CT and sonographic techniques for appendiceal imaging. Am J Roentgenol 176: Lane MJ, Lui DM, Huynh MD, Jeffrey RB, Mindelzun, Katz DS (1999) Suspected acute appendicitis: nonenhanced helical CT in 300 consecutive patients. Radiology 213: Guillerman RP, Brody AS, Kraus SJ (2002) Evidence-based guidelines for pediatric imaging: the example of the child with possible appendicitis. Pediatr Ann 31: Huyn LN, Coughlin BF, Wolfe J, Blank F, Lee SY, Smithline HA (2004) Patient encounter time intervals in the evaluation of emergency department patients requiring abdominopelvic CT: oral contrast versus no contrast. Emerg Radiol 10: Kirkpatrick RH, Wittenberg J, Schaffer, DL, Black EB, Hall DA, Braitman BS, Ferrucci JR JT (1978) Scanning techniques in computed body tomography. Am J Roentgenol 130:

8 Stephens DH, Hattery RR, Sheedy P II (1976) Computed tomographyof the abdomen early experience with the EMI body scanner. Radiology 119: Kandis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33: Altman DG (1991) Practical statistics for medical research. In: Altman DG (ed) Chapman and Hall, London 19. Mitchell DG, Parker L, Sunshine JH, Levin DC (2002) Body MR imaging and CT volume:variations and trends based on an analysis of medicare and fee for service health insurance databases. Am J Roentgenol 179: Trzeciak S, Rivers EP (2003) Emergency department overcrowding in the United States: an emerging threat to patient safety and public health. Emerg Med J 20: Schull MJ, Szalai JP, Schwartz B, Redelmeier DA (2001) Emergency department overcrowding following systematic hospital restructuring: trends at twenty hospitals over ten years. Acad Emerg Med 8: McCabe JB (2001) Emergency department overcrowding: a national crisis. Acad Med 76: Leslie A, Jones AJ, Goddard PR (2000) The influence of clinical information on the reporting of CT by radiologists. Br J Rad 73: Bechtold RE, Chen MY, Ott DJ et al (1997) Interpretation of abdominal CT: analysis of errors and their causes. J Comput Assist Tomo 21: Yoon LS, Haims AH, Brink JA, Rabinovici R, Forman HP (2002) Evaluation of an emergency radiology quality assurance program at a level I trauma center:abdominal and pelvic CT studies. Radiology 224: Freed KS, Paulson EK, Frederick MG et al (1998) Interobserver variability in the interpretation of unenhanced helical CT for the diagnosis of ureteral stone disease. J Comput Assist Tomogr 22: Garg K, Kemp JL, Russ PD, Baron AE (2001) Thromboembolic disease: variability of interobserver agreement in the interpretation of CT venography with pulmonary angiography. Am J Roentgenol 176: Ruiz Y, Caballero P, Caniego JL et al (2003) Prospective comparison of helical CT with angiography in pulmonary embolism: global and selective vascular territory analysis. Interobserver agreement. Eur Radiol 13: Lal NR, Uwada MM, Eldevik P, Desmond JS (2000) Clinical consequences of misinterpretations of neuroradiologic CT scans by on-call radiology residents. Am J Neuroradiol 21: Alfaro D, Levitt MA, English DK, Williams V, Eisenberg R (1995) Accuracy of interpretation of cranial computed tomography scans in an emergency medicine residency program. Ann Emerg Med 25: Fletcher BD, Glicksman AS, Gieser P (1999) Interobserver variability in the detection of cervical thoracic Hodgkin s disease by computed tomography. J Clin Oncol 17: