Small-bowel imaging in Crohn's disease: a prospective, blinded, 4 ...

ORIGINAL ARTICLE: Clinical EndoscopySmall-bowel imaging in Crohn’s disease: a prospective, blinded,4-way comparison trialCraig A. Solem, MD, Edward V. Loftus Jr., MD, Joel G. Fletcher, MD, Todd H. Baron, MD,Christopher J. Gostout, MD, Bret T. Petersen, MD, William J. Tremaine, MD, Laurence J. Egan, MD,William A. Faubion, MD, Kenneth W. Schroeder, MD, PhD, Darrell S. Pardi, MD, Karen A. Hanson, RN, CNP,Debra A. Jewell, RN, John M. Barlow, MD, Jeff L. Fidler, MD, James E. Huprich, MD, C. Daniel Johnson, MD,W. Scott Harmsen, MS, Alan R. Zinsmeister, PhD, William J. Sandborn, MDRochester, Minnesota, USABackground: With the introduction of new techniques to image the small bowel, there remains uncertaintyabout their role for diagnosing Crohn’s disease.Objective: To assess the sensitivity and specificity of capsule endoscopy (CE), CT enterography (CTE), ileocolonoscopy,and small-bowel follow-through (SBFT) in the diagnosis of small bowel Crohn’s disease.Methods: Prospective, blinded trial.Setting: Inflammatory bowel disease clinic at an academic medical center.Patients: Known or suspected Crohn’s disease. Exclusion criteria included known abdominal abscess and nonsteroidalanti-inflammatory drug (NSAID) use. Partial small-bowel obstruction (PSBO) at CTE excluded patientsfrom subsequent CE.Interventions: Patients underwent all 4 tests over a 4-day period.Main Outcome Measurements: Sensitivity, specificity, and accuracy of each test to detect active small-bowelCrohn’s disease. The criterion standard was a consensus diagnosis based upon clinical presentation and all4 studies.Results: Forty-one CTE examinations were performed. Seven patients (17%) had an asymptomatic PSBO. Fortypatients underwent colonoscopy, 38 had SBFT studies, and 28 had CE examinations. Small-bowel Crohn’s diseasewas active in 51%, absent in 42%, inactive in 5%, and suspicious in 2% of patients. The sensitivity of CEfor detecting active small-bowel Crohn’s disease was 83%, not significantly higher than CTE (83%), ileocolonoscopy(74%), or SBFT (65%). However, the specificity of CE (53%) was significantly lower than the other tests(P ! .05). One patient developed a transient PSBO due to CE, but no patients had retained capsules.Limitation: Use of a consensus clinical diagnosis as the criterion standarddbut this is how Crohn’s disease isdiagnosed in practice.Conclusions: The sensitivity of CE for active small-bowel Crohn’s disease was not significantly different fromCTE, ileocolonoscopy, or SBFT. However, lower specificity and the need for preceding small-bowel radiography(due to the high frequency of asymptomatic PSBO) may limit the utility of CE as a first-line test for Crohn’sdisease. (Gastrointest Endosc 2008;68:255-66.)Abbreviations: CBC, complete blood count; CE, capsule endoscopy; CRP,C-reactive protein; CTE, CT enterography; NSAID, non-steroidal anti-inflammatorydrug; PEG, polyethylene glycol; PSBO, partial small-bowelobstruction; SB, small-bowel; SBFT, small bowel follow-through.Copyright ª 2008 by the American Society for Gastrointestinal Endoscopy0016-5107/$32.00doi:10.1016/j.gie.2008.02.017Modern clinicians have many options available to themfor small-bowel (SB) imaging in Crohn’s disease. Bariumstudies have traditionally been the conventional test forSB radiography in this setting. 1-6 Barium small-bowel follow-through(SBFT) examinations have previously beenshown to be preferable to barium enteroclysis studies ina prospective study involving patients with Crohn’swww.giejournal.org Volume 68, No. 2 : 2008 GASTROINTESTINAL ENDOSCOPY 255

Small-bowel imaging in Crohn’s diseaseSolem et alprocedure and the willingness to repeat the procedure inthe future.Consensus reference standardAfter the results of the 4 studies (CTE, ileocolonoscopy,CE, and SBFT) were recorded for each patient, the findingsof all of the studies were reviewed at the conclusionof the study period by a consensus panel of co-investigatorsrepresenting each diagnostic modality and a referringgastroenterologist, to establish a consensus global assessmentof SB Crohn’s disease. If there were discordantresults among studies, this discussion necessitated achange of opinion by one or more investigators about theinitial interpretation of the findings of one or more studies.More specifically, the investigator representing one of thediagnostic modalities had to concede that a mistake hadbeen made in the initial interpretation of the SB findingswith their modality. Thus, the reference standard for assessmentof Crohn’s disease activity was the final determinationof the consensus panel (ie, active, suspicious foractive, inactive, or absent). The consensus criterion standardhas previously been used in a prospective trial oftests evaluating perianal Crohn’s disease. 42 After the modifiedreadings were used to establish the consensus referencestandard for SB disease, the original interpretationof each study as recorded on the study forms, was usedto estimate the performance characteristics of the tests.In a prespecified primary analysis, the diagnoses of activeand suspicious SB Crohn’s disease were grouped together,and the diagnoses of inactive or absent diseasewere grouped together, for all reported analyses to detectactive SB Crohn’s. A prespecified secondary analysis wasalso performed for detecting the presence of SB Crohn’sdisease by combining the categories of active, suspicious,and inactive disease, and comparing this group to the absentcategory. (However, the latter results were consistentwith those of active SB Crohn’s, and are therefore not reportedin the Results.)Statistical analysisFor the primary analysis, the sensitivity of each test wascalculated by dividing the number of patients with positivetest results by the number of patients with active or suspiciousSB Crohn’s disease based on the consensus criterionstandard. The specificity of each test was calculated by dividingthe number of patients with negative tests resultsby the number of patients with inactive or absent SBCrohn’s disease based on the consensus criterion standard.The diagnostic accuracy of each test was calculatedby the number of patients with test results in agreementwith the consensus criterion standard divided by the numberof tests performed. Exact binomial 95% confidence intervalswere also reported.The operating characteristics of each SB examinationwere compared to those of CE in the subset of patientshaving both CE and another SB examination. The sensitivity,specificity, and accuracy of CTE, ileocolonoscopy, andSBFT were compared to CE by using McNemar’s test ofpaired proportions. The combined sensitivity, specificity,and accuracy of all pairs of tests that could be consideredwere also estimated against the consensus criterion standard.McNemar’s test was used to compare the performancesof pairs of tests (CE plus CTE vs ileocolonoscopyplus SBFT, CE plus ileocolonoscopy vs CTE plus SBFT,and CTE plus ileocolonoscopy vs CE plus SBFT). Comparisonsof pairs containing a common diagnostic test were notperformed. The P values were not adjusted for multiplecomparisons.RESULTSForty-two patients enrolled in the study (Fig. 1). Onepatient withdrew for personal reasons before any diagnosticstudies were performed and was not included in theanalysis. The median age of the 41 patients (81% female)participating in the study was 40.1 years (range 18.9-73.5years). Forty-one patients underwent CTE. One patientsubsequently withdrew for personal reasons and had nofurther studies. Ileoscopy was feasible in 36 of the 40 colonoscopies.Thirty-eight patients had SBFT examinationsafter 1 patient exited the study due to a stricture detectedon CTE, and 1 patient underwent surgery for a PSBO.Twenty-eight patients had CE studies after 10 patientswere excluded for PSBO or abscess, and 2 patients leftthe study for personal reasons. One capsule study was repeateddue to a technical malfunction with the initialexamination (no images were obtained), and one capsulewas retained in the stomach for the duration of batterylife, resulting in no SB images (but the capsule eventuallypassed spontaneously). Twenty-seven capsule studieswere then used for analysis. Table 1 details the combinationsof the 4 SB examinations patients received.The study co-investigators were able to reach a consensusopinion regarding the global assessment of SB Crohn’sdisease after unblinding at the conclusion of study in all 41patients who had one or more SB studies (Table 2). Thefinal clinical diagnoses for all patients, not limited to theSB, are reported in Table 3.Table 4 reports the sensitivity, specificity, and accuracyof the 4 SB tests individually, with the consensus criterionstandard used to classify patient status as either ‘‘active’’(active plus suspicious) or ‘‘inactive’’ (inactive plus absent).The sensitivity for the detection of active SB Crohn’sdisease was 83% for CE, 82% for CTE, 74% for ileocolonoscopy,and 65% for SBFT. The specificity for SB Crohn’sdisease was 100% for ileocolonoscopy, 94% for SBFT,89% for CTE, and 53% for CE. The overall accuracy forSB Crohn’s was 86% for ileocolonoscopy, 85% for CTE,79% for SBFT, and 67% for CE.The performance of CTE, ileocolonoscopy, and SBFTwere each compared with that of CE (Table 5). There258 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 2 : 2008 www.giejournal.org

Solem et alSmall-bowel imaging in Crohn’s diseaseFigure 1. A graphic representation of patient movement throughout the course of the study.TABLE 1. Combinations of the 4 SB examinationsundergone by 42 participants initially enrolled inthe studyCE(n Z 27)Ileocolonoscopy(n Z 36)SBFT(n Z 38)CTE(n Z 41)No.patientsNo No No No 1No No No Yes 1No No Yes Yes 3No Yes No Yes 2No Yes Yes Yes 8Yes No Yes Yes 1Yes Yes Yes Yes 26CE, Capsule endoscopy; SBFT, small bowel follow-through; CTE,computed tomographic enterography.were 12 patients in each of 3 sensitivity comparisons, andfor specificity there were 15 patients, except to compareCE versus SBFT, where only 14 received both examinations.The sensitivity of CE (83%) was not significantly differentfrom that of CTE (67%, P Z .63), ileocolonoscopy(67%, P Z .69), or SBFT (50%, P Z .22). However, thespecificity of CE (53%) was significantly lower than thatfor either CTE (100%, P Z .02) or SBFT (100%, P Z.02). The specificity of CE (43%) was also significantlylower among the 14 patients having ileocolonoscopy(100%, P Z .03). The lower specificity of CE was due toTABLE 2. Final consensus diagnosis for small-bowelCrohn’s disease in 41 patients who underwent at least 1diagnostic test, and in 26 patients who underwent all 4testsSmall-bowel Crohn’sdisease diagnosisAll patients,n (%)*Patients withall 4 tests, n (%)yActive 21 (51) 11 (42)Absent 17 (42) 12 (46)Inactive 2 (5) 2 (8)Suspicious 1 (2) 1 (4)Total 41 (100) 26 (100)*Percentage of the group undergoing at least 1 SB imaging test.yPercentage of the group undergoing all 4 tests.the fact that, in 7 patients, there were findings that ledthe CE reader to deem the examination either suspicious(n Z 5) or ‘‘definite’’ (n Z 2) for active Crohn’s disease,but according to the consensus criterion standard, thesepatients had either no evidence of Crohn’s disease (n Z5) or only inactive disease (n Z 2).The performance characteristics of pairs of SB examinationsagainst the consensus criterion standard were estimated(Table 4). A patient was classified positive foractive SB Crohn’s disease if either of the 2 examinationsindicated active disease. The sensitivity for combinationsof examinations ranged from a high of 100% for the CEand ileocolonoscopy combination to a low of 78% forwww.giejournal.org Volume 68, No. 2 : 2008 GASTROINTESTINAL ENDOSCOPY 259

Small-bowel imaging in Crohn’s diseaseSolem et alTABLE 3. Final overall clinical diagnosis (not limited tothe small intestine) in all 41 patients who underwent atleast 1 diagnostic test, and in 26 patients whounderwent all 4 testsClinicaldiagnosisAllpatients,n (%)*Patientswith all 4tests, n (%)yCrohn’s disease 30 (73%) 19 (73%)Indeterminate abdominal pain 3 (7%) 3 (11%)Irritable bowel syndrome 2 (5%) 2 (8%)Ulcerative colitis 1 (2%) 0 (0%)Indeterminate colitis 1 (2%) 0 (0%)Carcinoid 1 (2%) 0 (0%)Submucosal mass, small bowel 1 (2%) 1 (4%)Gastroparesis 1 (2%) 1 (4%)Indeterminate diarrhea 1 (2%) 0 (0%)Total 41 (100%) 26 (100%)*Percentage of the group undergoing at least 1 SB imaging test.yPercentage of the group undergoing all 4 tests.the ileocolonoscopy and SBFT combination. The specificityof examination combinations against the criterion standardranged from a high of 100% for SBFT plusileocolonoscopy to a low of 53% for both the CE plusSBFT and CE plus CTE combinations. The overall accuracyof examination combinations against the criterion standardranged from 89% for both the CTE plus ileocolonoscopyand CTE plus SBFT combinations to 70% for theCE plus CTE and CE plus SBFT combinations.No pair of SB examinations was significantly more sensitivethan the combination containing the other two examinations(among the 4 SB examinations), P O .05(Table 6). However, each combination of examinationsthat included CE had a significantly lower specificity(57%) than did the opposite pair of examinations (P Z.03 for each of the 3 comparisons, unadjusted).Patients completed a questionnaire at the end of thestudy. All patients swallowed the capsule easily. Ninetytwopercent of patients did not experience any pain duringthe CE study, significantly better than colonoscopy(P ! .05), but no better than CTE or SBFT (P O .05).Ninety-three percent of patients stated they would undergoa CE study again, which was not significantly differentthan the other tests (P O .05).Even though all patients with obstructive symptomswere excluded from study enrollment, we detected anasymptomatic PSBO in 7 patients (17%) on CTE (Fig. 2),excluding them from CE. Soon after capsule ingestion,one patient with a very long segment of active diseasebut without PSBO on CTE (3.4%) experienced a highlysymptomatic PSBO requiring hospitalization, but the capsulepassed within a few days following medical therapywith corticosteroids and infliximab, with symptoms resolving(Fig. 3). Unlike the higher capsule retention rates seenin other studies, we had no retained capsules (confirmedby radiograph) using our diagnostic algorithm.Chyme significantly limited study quality in 4 patientswho underwent CE (14%) despite efforts to cleanse with1 L PEG solution prior to the study. The terminal ileumwas examined completely in only 75% of studies. One capsulestudy had a technical malfunction resulting in no images.In a single patient with suspected Crohn’s disease,CE demonstrated multiple SB erosions when the other 3SB tests were normal (Fig. 4).DISCUSSIONIn this prospective, blinded trial, we compared 4 primarysmall-intestinal imaging modalities in the diagnosisof SB Crohn’s disease. In contrast to the other studies examiningthe utility of CE in patients with known or suspectedCrohn’s disease, 22,27,29,33-40 we believe this is theonly prospective, blinded trial comparing all 4 primarySB imaging modalities over a short time period, avoidingboth treatment and natural history effects. Unlike the ‘‘diagnosticyield’’ reported in most published CE studies, weestablished a surrogate criterion standard, allowing us tocalculate the sensitivity, specificity, and accuracy of eachtest for diagnosing SB Crohn’s disease. We found thatthere was no significant difference among CTE, CE, ileocolonoscopy,and SBFT for the sensitivity or accuracy of diagnosingSB Crohn’s disease. CTE and CE demonstratedsimilarly high sensitivity for SB Crohn’s, which was numericallysuperior to SBFT and ileocolonoscopy. Although thenumerical superiority of CTE and CE was not statisticallysignificant, this may have been due to the limited samplesize and potential lack of power to detect a significant difference.Combining diagnostic tests may improve sensitivity,and this is commonly used in clinical practice. Wefound that the specificity of CE for SB Crohn’s diseasewas significantly lower than other SB imaging modalities,which lowered its overall accuracy. Our study also indirectlysuggests that SB radiologic imaging may be neededprior to CE in suspected or known Crohn’s disease patients,due to the high frequency (17%) of asymptomaticPSBO. When we used an algorithm of performing CTEprior to CE and using a finding of stricture with proximalbowel dilation to preclude ingestion of the endoscopycapsule, we observed no capsule retention.The finding of a lower specificity of CE for SB Crohn’sdisease is not entirely unexpected. ‘‘Diagnostic yield’’ asa measure of the operating performance of a diagnostictest is less than idealdit merely reports the prevalenceof abnormalities detected by a particular imaging modalityand does not attempt to determine whether those260 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 2 : 2008 www.giejournal.org

Solem et alSmall-bowel imaging in Crohn’s diseaseTABLE 4. Individual performance estimates of 4 diagnostic tests and all pairs of these tests for active small-bowel Crohn’s disease*SB test Patients tested, n Sensitivity (%) 95% CI Specificity (%) 95% CI Accuracy (%) 95% CIIndividual testCE 27 10/12 (83) 52-98 8/15 (53) 27-79 18/27 (67) 46-83CTE 41 18/22 (82) 60-95 17/19 (89) 67-99 35/41 (85) 71-94Ileocolonoscopy 36 14/19 (74) 49-91 17/17 (100) 80-100 31/36 (86) 71-95SBFT 38 13/20 (65) 41-8 17/18 (94) 73-100 30/38 (79) 63-90Pairs of TestsCE & CTE 27 11/12 (92) 62-100 8/15 (53) 27-79 19/27 (70) 50-86CE & ileocolonoscopy 26 12/12 (100) 74-100 8/14 (57) 29-82 20/26 (77) 56-91CE & SBFT 27 11/12 (92) 62-100 8/15 (53) 27-79 19/27 (70) 50-86SBFT & ileocolonoscopy 34 14/18 (78) 52-94 16/16 (100) 79-100 30/34 (88) 73-97CTE & ileocolonoscopy 36 16/19 (84) 60-97 16/17 (94) 71-100 32/36 (89) 74-97CTE & SBFT 38 17/20 (85) 62-97 17/18 (94) 73-100 34/38 (89) 75-97CTE, CT enterography; CE, capsule endoscopy; SBFT, small-bowel follow-through; 95% CI, 95% confidence intervals.*Criterion standard is based on consensus diagnosis of active Crohn’s disease or suspicious for Crohn’s disease.TABLE 5. Comparison of sensitivity and specificitybetween capsule endoscopy and 3 other diagnostictests for active small-bowel Crohn’s disease*SB testSensitivity(%)P-valueySpecificity(%)P-valueyCE 10/12 (83) ref 8/15 (53) refCTE 8/12 (67) .63 15/15 (100) .02Ileocolonoscopy 8/12 (67) .69 14/14 (100) .03SBFT 6/12 (50) .22 15/15 (100) .02CTE, CT enterography; CE, capsule endoscopy; SBFT, small-bowelfollow-through; 95% CI, 95% confidence intervals.*Criterion standard is based on consensus diagnosis of activeCrohn’s disease or suspicious for Crohn’s disease.yUsing Mc Nemar’s test for paired proportions, testing performancecompared to the CE examination, unadjusted for multiplecomparisons.abnormal findings are clinically significant. Most studiesreporting ‘‘diagnostic yield’’ do so because the test isnot being compared against a reference standard. CE isan extremely sensitive tool for direct visualization ofthe SB mucosa. As in most diagnostic tests, improvedsensitivity typically comes at the cost of decreased specificity.43 Goldstein et al 44 found that 22% of asymptomaticcontrol patients who denied recent NSAID ingestion hadSB abnormalities at CE, with an average of 4 lesions persubject. As several CE findings, such as mucosal erythema,are nonspecific, and do not indicate a diagnosisof active Crohn’s disease by themselves, their inclusionTABLE 6. Selected comparisons of sensitivity andspecificity between pairs of small-bowel examinations*CE &ileocolonoscopySensitivity(%)P- Specificityvaluey (%)12/12 (100) 8/14 (57)P-valueyCTE & SBFT 9/12 (75) .25 14/14 (100) .03CE & CTE 11/12 (92) 8/14 (57)SBFT &ileocolonoscopy9/12 (75) .63 14/14 (100) .03CE & SBFT 11/12 (92) 8/14 (57)CTE &ileocolonoscopy9/12 (75) .63 14/14 (100) .03CTE, CT enterography; CE, capsule endoscopy; SBFT, small-bowelfollow-through; 95% CI, 95% confidence intervals.*Criterion standard is based on consensus diagnosis of activeCrohn’s disease or suspicious for Crohn’s disease.yUsing McNemar’s test for paired proportions, unadjusted formultiple comparisons.into reference standard assessments of disease activityin SB imaging trials may result in confirmation bias favoringCE.Buchman et al 36 used SBFT as a method to detect SBstrictures that may lead to retained capsules, and founda 29% incidence of SB strictures in Crohn’s patients, whichprecluded capsule endoscopy. By using SBFT as a screeningexamination, they observed 2 episodes of capsule retention(7%). Voderholzer et al 27 compared CE and CTEwww.giejournal.org Volume 68, No. 2 : 2008 GASTROINTESTINAL ENDOSCOPY 261

Small-bowel imaging in Crohn’s diseaseSolem et alFigure 2. A, B, C, Computed tomography enterography in a 43-year-old woman with Crohn’s disease, depicting asymptomatic PSBO due to multiplesegments of actively inflamed jejunum, precluding ingestion of endoscopy capsule. Axial and 2D multiplanar reformatted images depict both luminalnarrowing caused by active inflammation (with segmental hyperenhancement mural stratification, arrows) and prestenotic bowel dilation(arrowheads).in 56 patients with known Crohn’s disease. 27 CTE was performedprior to CE, and 27% of patients were precludedfrom CE because of strictures. 27 They used stricter criteriafor excluding patients from CE (ie, stricture !1-cm luminal diameter with no proximal SB dilation required),but still experienced one episode of capsuleretention (2%) requiring removal from the jejunum withpush enteroscopy, and one episode of symptomaticPSBO requiring high-dose corticosteroid therapy beforespontaneous passage.Our results are somewhat at variance with other studiescomparing CE and CTE in patients with known or suspectedCrohn’s disease. 22,27,29 A meta-analysis of these 3studies suggested that the prevalence of abnormalities detectedon CE was 38% higher than that of CTE. 41 However,this value was significantly higher than CTE only for thesubgroup of patients with known Crohn’s disease. We believethe differences in study results are due to a numberof factors. In one of these studies, 22 the enteric contrastused apparently was a positive agent (30% Telebrix), not262 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 2 : 2008 www.giejournal.org

Solem et alSmall-bowel imaging in Crohn’s diseaseFigure 3. A, B, C, D, SB ulceration and stenosis in the patient who experienced a highly symptomatic PSBO shortly after capsule ingestion, resulting inhospitalization for several days. E, CTE demonstrated approximately 50 cm of actively inflamed ileum with mural enhancement and stratification (arrows),but without any proximal SB dilation (arrowhead). The symptoms resolved and the capsule eventually passed after steroid and infliximabtreatment.a negative/neutral contrast agent, as used in our study. Useof a positive contrast would diminish the operating characteristicsof CTE, as it would not take advantage of thedifferences in attenuation between the bowel lumen andbowel wall. Furthermore, our study was the only one ofthese that met both of the following criteria: there existeda range of diagnostic uncertainty (recruitment of bothknown and suspected Crohn’s), and the readers of eachwww.giejournal.org Volume 68, No. 2 : 2008 GASTROINTESTINAL ENDOSCOPY 263

Small-bowel imaging in Crohn’s diseaseSolem et alFigure 4. A, B, C, D, SB ulcers on capsule endoscopy in the single patient found to have active SB Crohn’s disease on CE despite all other imagingstudies being normal (3.4%).modality were blinded to the results of the other modalities.Furthermore, in our study, all patients underwent allimaging modalities after informed consent had been obtained;in other words, the decision to perform certaintests, or even enroll patients, was not dependent uponthe results of other tests.The primary limitation of this study was the use ofthe consensus panel clinical diagnosis as the referencestandard assessment. Because no criterion standard forthe diagnosis of Crohn’s disease exists, we believe the consensusclinical diagnosis best approximates how this diagnosisis actually made in clinical practice. Crohn’s diseaseis a clinical diagnosis based upon a comprehensive evaluationincluding clinical presentation, SB radiographic imaging,endoscopic findings, and histology. The conceptof using a consensus diagnosis when no criterion standarddiagnostic test exists has been used previously in a studyexamining the sensitivity of diagnostic tests (magneticresonance imaging, examination under anesthesia, andendoscopic ultrasound) for perianal fistulas in Crohn’sdisease. 42 This method unfortunately requires that the resultsof all imaging tests will be included in the determinationof the consensus clinical diagnosis with which it isbeing compared. However, we believed this was necessarybecause clinical symptoms and serologic markers cannotreliably estimate or predict the presence, severity, or locationof active Crohn’s disease. Submitting all patients toexploratory laparoscopy or laparotomy for a criterion standardconfirmation is obviously invasive and impractical.Moreover, a meta-analysis of several studies suggests thatCE may be more sensitive at diagnosing SB Crohn’s diseasethan other imaging modalities. 41 This unique situationwarrants inclusion of CE in the consensus clinicaldiagnosis, to minimize incorrectly labeling lesions seenon CE but not the other imaging modalities as false positives.Neither laparotomy nor laparoscopy are routinelyused for diagnosing Crohn’s disease in standard clinicalpractice, but rather these interventions are reserved forcomplications of Crohn’s that are not amenable to medicaltherapy. Additionally, surgery may detect gross findings ofCrohn’s disease but may miss mucosal lesions that couldbe clinically significant.264 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 2 : 2008 www.giejournal.org

Solem et alSmall-bowel imaging in Crohn’s diseaseAs a result of our study, we have adopted the followingdiagnostic algorithm for evaluating SB Crohn’s disease.CTE and ileocolonoscopy are first-line combined tests,with a combined sensitivity of 84% and specificity of94% for active Crohn’s disease. CTE has high sensitivity,similar to CE, but with superior specificity. Furthermore,CTE detects extraluminal intestinal complications (eg, fistulasand abscesses) and may help distinguish inflammatoryfrom fibrostenotic strictures, facilitating therapeuticdecisions while avoiding potential risks of retained capsules.Ileocolonoscopy stages the extent of disease andobtains necessary histology. SBFT should be consideredas an alternative to CTE based upon local expertise orother extenuating circumstances such as iodine allergyor renal insufficiency. CE should subsequently be performedif the clinical suspicion for Crohn’s disease remainshigh despite a negative evaluation.DISCLOSUREThe following authors report that they have no disclosuresrelevant to this publication: C. A. Solem,T. H. Baron, C. J. Gostout, B. T. Petersen, W. J. Tremaine,L. J. Egan, W. A. Faubion, K. W. Schroeder, D. S. Pardi,K. A. Hanson, D. A. Jewell, J. M. Barlow, J. E. Huprich,W. S. Harmsen, A. R. Zinsmeister, W. J. Sandborn. Thefollowing authors report actual or potential conflicts:E. V. Loftus, Jr, received consulting fees from Given Imagingin the 12 months prior to implementation of the trial.J. G. Fletcher and C. D. Johnson receive grant supportfrom GE Healthcare, the manufacturer of the multidetectorCT scanner and the iodinated intravenous contrastused for CT enterography exams. J. L. Fidler has an NIHgrant, in which GE Healthcare is an external collaborator.Supported in part by Given Imaging and the MayoFoundation for Medical Education and Research.REFERENCES1. Goldberg H, Caruthers S, Nelson J, et al. Radiographic findings of theNational Cooperative Crohn’s disease study. Gastroenterology1979;77:925-37.2. Nolan D, Gourtsoyiannis N. Crohn’s disease of the small intestine: a reviewof the radiological appearances in 100 consecutive patients examinedby a barium infusion technique. Clin Radiol 980;31:597–603.3. Nolan D, Piris J. Crohn’s disease of the small intestine: a comparativestudy of the radiological and pathological appearances. Clin Radiol1980;31:591-6.4. Diner W, Hoskins E, Navab F. Radiologic examination of the small intestine:review of 402 cases and discussion of indications andmethods. South Med J 1984;77:68-74.5. Carlson H. Perspective: the small bowel examination in the diagnosisof Crohn’s disease. AJR Am J Roentgenol 1986;147:63-5.6. Bernstein C, Boult I, Greenberg H, et al. A prospective randomizedcomparison between small bowel enteroclysis and small bowel follow-throughin Crohn’s disease. Gastroenterology 1997;113:390-8.7. Coremans G, Rutgeerts P, Geboes K, et al. The value of ileoscopy withbiopsy in the diagnosis of intestinal Crohn’s disease. Gastrointest Endosc1984;30:167-72.8. Lorenz-Meyer H, Malchow H, Miller B, et al. European CooperativeCrohn’s disease study (ECCDS): colonoscopy. Digestion 1985;31:109-19.9. Gomes P, Boulay C, Smith C, et al. Relationship between disease activityindices and colonoscopic findings in patients with colonic inflammatorybowel disease. Gut 1986;27:92-5.10. Modigliani R, Mary J. Reproducibility of colonoscopic findings inCrohn’s disease: a prospective multicenter study of interobserver variation.Dig Dis Sci 1987;32:1370-9.11. Pera A, Bellando P, Caldera D, et al. Colonoscopy in inflammatorybowel disease. Gastroenterology 1987;92:181-5.12. Mary J, Modigliani R. Development and validation of an endoscopicindex of the severity for Crohn’s disease: a prospective multicentrestudy. Gut 1989;30:983-9.13. Landi B, Anh T, Cortot A, et al. Endoscopic monitoring of Crohn’s diseasetreatment: a prospective, randomized clinical trial. Gastroenterology1992;102:1647-53.14. Marshall J, Barthel J. The frequency of total colonoscopy and terminalileal intubation in the 1990s. Gastrointest Endosc 1993;39:518-20.15. Robinson A, Wolfe P. A prospective evaluation of successful terminalileum intubation during routine colonoscopy. Gastrointest Endosc1994;40:544-6.16. Leighton JA, Shen B, Baron TH, et al. ASGE guideline: endoscopy in thediagnosis and treatment of inflammatory bowel disease. GastrointestEndosc 2006;63:558-65.17. Reittner P, Goritschnig T, Petritsch W, et al. Multiplanar spiral CT enterographyin patients with Crohn’s disease using a negative oral contrastmaterial: initial results of a noninvasive imaging approach. EurRadiol 2002;12:2253-7.18. Turetschek K, Schober E, Wunderbaldinger P, et al. Findings at helicalCT-enteroclysis in symptomatic patients with Crohn disease: correlationwith endoscopic and surgical findings. J Comput Assisted Tomogr2002;26:488-92.19. Doerfler OC, Ruppert-Kohlmayr AJ, Reittner P, et al. Helical CT of thesmall bowel with an alternative oral contrast material in patientswith Crohn disease. Abdom Imaging 2003;28:313-8.20. Hassan C, Cerro P, Zullo A, et al. Computed tomography enteroclysisin comparison with ileoscopy in patients with Crohn’s disease. Int JColorectal Dis 2003;18:121-5.21. Wold PB, Fletcher JG, Johnson CD, et al. Assessment of small bowelCrohn disease: noninvasive peroral CT enterography compared withother imaging methods and endoscopy–feasibility study. Radiology2003;229:275-81.22. Eliakim R, Suissa A, Yassin K, Katz D, Fischer D. Wireless capsule videoendoscopy compared to barium follow-through and computerised tomographyin patients with suspected Crohn’s disease - final report. DigestLiver Dis 2004;36:519-22.23. Buchman AL, Miller F, Wallin A, et al. Videocapsule endoscopy versusbarium contrast studies for diagnosis of Crohn’s disease [abstract].Gastroenterology 2004;126:A459.24. Hara AK, Leighton JA, Sharma VK, et al. Small bowel: Preliminary comparisonof capsule endoscopy with barium study and CT. Radiology2004;230:260-5.25. Schreyer AG, Seitz J, Feuerbach S, et al. Modern imaging using computertomography and magnetic resonance imaging for inflammatorybowel disease (IBD) AU1. Inflammatory Bowel Dis 2004;10:45-54.26. Sailer J, Peloschek P, Schober E, et al. Diagnostic value of CT enteroclysiscompared with conventional enteroclysis in patients with Crohn’sdisease. AJR Am J Roentgenol 2005;185:1575-81.27. Voderholzer WA, Beinhoelzl J, Rogalla P, et al. Small bowel involvementin Crohn’s disease: a prospective comparison of wireless capsuleendoscopy and computed tomography enteroclysis. Gut 2005;54:369-73.28. Bodily KD, Fletcher JG, Solem CA, et al. Crohn disease: mural attenuationand thickness at contrast-enhanced CT enterography –www.giejournal.org Volume 68, No. 2 : 2008 GASTROINTESTINAL ENDOSCOPY 265

Small-bowel imaging in Crohn’s diseaseSolem et alcorrelation with endoscopic and histologic findings of inflammation.Radiology 2006;238:505-16.29. Hara AK, Leighton JA, Heigh RI, et al. Crohn disease of the small bowel:preliminary comparison among CT enterography, capsule endoscopy,small-bowel follow-through, and ileoscopy. Radiology 2006;238:128-34.30. Rajesh A, Maglinte DDT. Multislice CT enteroclysis: technique and clinicalapplications. Clin Radiol 2006;61:31-9.31. La Seta F, Buccellato A, Tese L, et al. Multidetector-row CT enteroclysis:indications and clinical applications. Radiol Med 2006;111:141-58.32. Paulsen SR, Huprich JE, Fletcher JG, et al. CT enterography as a diagnostictool in evaluating small bowel disorders: review of clinical experiencewith over 700 cases. Radiographics 2006;26:641-57, discussion657–62.33. Eliakim R, Fischer D, Suissa A, et al. Wireless capsule video endoscopyis a superior diagnostic tool in comparison to barium follow-throughand computerized tomography in patients with suspected Crohn’sdisease. Eur J Gastroenterol Hepatol 2003;15:363-7.34. Fireman Z, Mahajna E, Broide E, et al. Diagnosing small bowel Crohn’sdisease with wireless capsule endoscopy. Gut 2003;52:390-2.35. Herrerias JM, Caunedo A, Rodriguez-Tellaz M, et al. Capsule endoscopyin patients with suspected Crohn’s disease and negative endoscopy.Endoscopy 2003;35:564-8.36. Buchman AL, Miller FH, Wallin A, et al. Videocapsule endoscopy versusbarium contrast studies for the diagnosis of Crohn’s disease recurrenceinvolving the small intestine. Am J Gastroenterol 2004;99:2171-7.37. Ge ZZ, Hu YB, Xiao SD. Capsule endoscopy and push enteroscopy inthe diagnosis of obscure gastrointestinal bleeding. Chin Med J2004;117:1045-9.38. Chong AKH, Taylor A, Miller A, et al. Capsule endoscopy vs. push enteroscopyand enteroclysis in suspected small-bowel Crohn’s disease.Gastrointest Endosc 2005;61:255-61.39. Marmo R, Rotondano G, Piscopo R, et al. Capsule endoscopy versusenteroclysis in the detection of small-bowel involvement in Crohn’sdisease: a prospective trial. Clin Gastroenterol Hepatol 2005;3:772-6.40. Golder SK, Schreyer AG, Endlicher E, et al. Comparison of capsule endoscopyand magnetic resonance (MR) enteroclysis in suspected smallbowel disease. Int J Colorectal Dis 2006;21:97-104.41. Triester SL, Leighton JA, Leontiadis GI, et al. A meta-analysis of theyield of capsule endoscopy compared to other diagnostic modalitiesin patients with non-stricturing small bowel Crohn’s disease. Am JGastroenterol 2006;101:954-64.42. Schwartz D, Wiersema M, Dudiak K, et al. A comparison of endoscopicultrasound, magnetic resonance imaging, and exam under anesthesiafor evaluation of Crohn’s perianal fistulas. Gastroenterology 2001;121:1064-72.43. Lashner BA. Sensitivity-specificity trade-off for capsule endoscopy inIBD: is it worth it? Am J Gastroenterol 2006;101:965-6.44. Goldstein JL, Eisen GM, Lewis B, et al. Video capsule endoscopy to prospectivelyassess small bowel injury with celecoxib, naproxen plusomeprazole, and placebo. Clin Gastroenterol Hepatol 2005;3:133-41.Received September 20, 2007. Accepted February 4, 2008.Current affiliations: Division of Gastroenterology and Hepatology (E.V.L.,T.H.B., C.J.G., B.T.P., W.J.T., W.A.F., K.W.S., D.S.P., K.A.H., D.A.J., W.J.S.),Department of Radiology (J.G.F., J.M.B., J.L.F., J.E.H., C.D.J.), and Division ofBiostatistics (W.S.H, A.R.Z.), Mayo Clinic College of Medicine, Rochester,Minnesota, Park-Nicollet Clinic (C.A.S.), Saint Louis Park, Minnesota, USA,Clinical Pharmacology Unit (L.J.E.), University College Hospital, Galway,Ireland, Department of Radiology (C.D.J.), Mayo Clinic, Scottsdale, Arizona,USA.Presented in part at the 106th Annual Meeting of the AmericanGastroenterological Association, Digestive Disease Week, May 14-19,2005, Chicago, Illinois (Gastroenterology 2005;128[Suppl 2]:A74).Reprint requests: Edward V. Loftus, Jr., MD, Division of Gastroenterologyand Hepatology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905.266 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 2 : 2008 www.giejournal.org