small intestine contrast ultrasonography

ORIGINAL ARTICLE

Small Intestine Contrast UltrasonographyAn Alternative to Radiology in the Assessment of Small Bowel Disease

Nadia Pallotta, MD, PhD, Ernesto Tomei, MD, Angelo Viscido, MD, Emma Calabrese, MD,Adriana Marcheggiano, MD, Renzo Caprilli, MD, and Enrico Corazziari, MD

Background: Radiology and transabdominal ultrasonography(TUS) are used in the evaluation of the small bowel; however, theformer technique is limited by radiation exposure, and the latter by itsinability to visualize the entire small bowel.

Aim: To evaluate the diagnostic accuracy of small intestine contrastultrasonography (SICUS) to assess the presence, number, site, andextension of small bowel lesions.

Subjects and Methods: TUS, SICUS, and small bowel follow-through (SBFT) were performed in 148 consecutive patients (78women; age range, 12 to 89 yr), 91 with undiagnosed conditions, and57 with previously diagnosed Crohn’s disease (CD).

Results: In the undiagnosed patients, the sensitivity and specificityof TUS and SICUS were 57% and 100%, and 94.3% and 98%, re-spectively. In the CD patients, the sensitivity of TUS and SICUS was87.3% and 98%, respectively. In comparison with SBFT, the exten-sion of lesions was correctly assessed with SICUS and greatly under-estimated with TUS. The concordance index between SBFT andSICUS for the number and site of lesions was 1 and 1 (P < 0.001),respectively, in undiagnosed patients, and 0.81 and 0.83 (P < 0.001),respectively, in CD patients. Between SBFT and TUS, the concor-dance index was 0.28 and 0.27 (not significant), respectively, in un-diagnosed patients, and 0.28 and 0.31 (not significant), respectively,in CD patients.

Conclusions: The diagnostic accuracy of SICUS is comparable tothat of a radiologic examination, and is superior to that of TUS indetecting the presence, number, extension, and sites of small bowellesions. These findings support the use of noninvasive SICUS for aninitial investigation when small bowel disease is suspected and in thefollow-up of CD patients.

Key Words: contrast, Crohn’s disease, morphology, pathology,small intestine, ultrasonography

(Inflamm Bowel Dis 2005;11:146–153)

Radiologic assessment, with either small bowel follow-through (SBFT) or enema, is the most widely used inves-

tigation1–3 for the diagnosis of lesions located in the smallbowel that, except for the terminal ileum and duodenum, isnot accessible to routine endoscopy. Radiologic findings arethe reference standard for the diagnosis of Crohn’s disease(CD) when lesions are located only in the small intestine,4–11 acondition that occurs in nearly 40% of patients.6,8,12 In addi-tion, several assessments of the gastrointestinal tract may berequired periodically in CD patients over a lifetime because ofthe relapsing nature of the disease, which usually has its onsetin the second or third decade of life and progresses over theyears in an unpredictable manner to involve �1 segments ofthe gastrointestinal tract.8,11,12

Because the medical use of radiographs is a significantsource of radiation,13 the second most common exposure afternatural terrestrial radiation, radiologic exposure should beminimized, particularly in young people, women of child-bearing age, and those who may require repetitive assessmentsfor the follow-up of CD.

Transabdominal ultrasonography (TUS) has beenwidely used as a noninvasive tool in the diagnosis of intestinallesions and related complications of CD.14–24 However, TUShas inherent limitations in the assessment of small bowel le-sions because the virtual lumen and the presence of gas inthe intestinal loops do not allow the entire small bowel to bethoroughly visualized in detail. These limitations are high-lighted by the observations that TUS, even when performed intertiary referral centers for intestinal diseases, has a variable,and sometimes unsatisfactory, sensitivity (74% to 85%) in de-tecting small bowel lesions14,21–24 in undiagnosed CD pa-tients, and usually fails to detect small bowel lesions locatedoutside of the terminal ileum.22,24 We have recently reportedthat abdominal ultrasonography (US) performed after the in-gestion of an oral contrast medium (i.e., small intestine con-

Received for publication June 14, 2004; accepted October 22, 2004.From the Dipartimento di Scienze Cliniche, Università “La Sapienza,” Rome,

Italy.Supported, in part, by a grant from the Italian Ministero Università Ricerca

(MIUR), Faculty of Medicine, University of Rome, and by a grant fromANEMGI (Associazione per la Neurogastroenterologia e Motilità intesti-nale).

Reprints: Enrico Corazziari, MD, Dipartimento di Scienze Cliniche, Univer-sità “La Sapienza” Policlinico “Umberto I,” Viale del Policlinico, 00161Rome, Italy (e-mail: [email protected])

Copyright © 2005 by Lippincott Williams & Wilkins

146 Inflamm Bowel Dis • Volume 11, Number 2, February 2005

trast US [SICUS]) enables the visualization of the entire smallbowel and gives the ability to dissociate 1 intestinal loop fromanother.25,26 SICUS has been shown to be comparable to radi-ology in detecting intestinal lesions in patients with undiag-nosed small bowel diseases.27,28 This study attempted to fur-ther assess the potential role of SICUS in the diagnosis of smallbowel disease. The primary aim of the present study was toevaluate the accuracy of SICUS in the assessment of the pres-ence, number, sites, and extension of small bowel lesions. Thesecondary aim was to compare the diagnostic accuracy ofhaving SICUS and TUS as reference standards to radiologicSBFT.

MATERIALS AND METHODS

SubjectsOne hundred forty-eight consecutive outpatients, who

were referred to our gastrointestinal (GI) outpatient clinic forthe diagnostic investigation of symptoms or signs suggestiveof small bowel pathology, for the recurrence of disease, or forthe follow-up of previously diagnosed CD, were prospectivelyevaluated. Pregnant women or women of child-bearing agewho were not using contraceptive measures were excludedfrom the study. The diagnosis of CD was based on the criteriaadopted in the European Collaborative Study on InflammatoryBowel Disease.29 The presence, site, and length of the smallbowel lesions were assessed at SBFT and, in those patientswho had undergone surgery, at the time that surgical findingswere evaluated. Medical history, abdominal and extraabdomi-nal complaints, and bowel habits were specifically investi-gated. Thirty-nine percent of patients with previously diag-nosed CD were referred or sought health care for disease re-currence, and 61% of patients were referred for follow-up. Inpatients with previously diagnosed CD, past surgery, the num-ber of recurrences, and current and previous medical treatmentwere asked about and the results reported. Patients with provenCD not affecting the small bowel were excluded from thestudy. Informed consent was obtained from each subject, andthe study protocol was approved by the local ethics committee.

Protocol of the StudyEach patient initially underwent a standardized clinical

interview and a physical examination, which was performedby 1 of 2 certified and experienced gastroenterologists (Drs.Viscido and Corazziari). After overnight fasting, patients con-secutively underwent TUS and SICUS, which were performedby an expert intestine-dedicated sonologist (Dr. Pallotta), onthe same day, on different days, and in random order; a radio-logic examination of the small bowel, which was performed bya GI radiologist (Dr. Tomei), and an endoscopic examinationof the entire large bowel and terminal ileum. Multiple mucosalbiopsy specimens were taken from the segments that had beeninvestigated by endoscopy. When deemed necessary, addi-tional investigations, including biochemistry, upper GI en-

doscopy, histology, computed tomography scan, or nuclearmagnetic resonance imaging were performed. The sonologistwas not aware of the clinical suspicion in the undiagnosed pa-tients who underwent the investigation.

In patients with proven CD, the sonologist and radiolo-gist were aware of the diagnosis, the history of bowel surgery,and the current clinical symptoms, but were blinded to the re-sults of the other investigations. At the end of the diagnosticinvestigation, each operator reported any small bowel abnor-malities on a standardized form, with particular reference tothe presence, anatomic site, and extension (in centimeters) ofintestinal wall and lumen abnormalities. The presence of fis-tulas and abscesses was checked and reported.

In the patients who had undergone surgery during thestudy period, US findings were compared with those of sur-gery.

US

Real-time US was performed using Toshiba Tosbee (To-kyo, Japan) equipment with a 3.5-MHz convex transducer anda 5-MHz linear array transducer. The sonologist had experi-ence with >7000 sonographic examinations of the whole ab-domen, and >3500 examinations of the bowel.

TUS

The criteria for the presence of small bowel lesions wereas follows14,18:1. The presence of a target appearance that was defined as a

strong echoic center surrounded by a sonolucent rim of>0.5 cm;

2. Increased wall thickness (i.e., �4 mm), measured both intransverse and longitudinal sections;

3. Stiff loop, which was defined as an intestinal loop with in-creased wall thickness, or a distended loop, which remainsunchanged during transabdominal compression;

4. Distended fluid-filled loops, which were defined as loopswith floating echogenic particles in the lumen;

5. The presence of a bowel stricture, which was identified bythe coexistence of a thickened and stiff bowel loop with theloss of wall layers and severe lumen narrowing, with orwithout prestenotic bowel dilatation; and

6. Abnormal intestinal motor activity, which was identified bythe absence of movements of echogenic particles and by theabsence of sonolucent disks, which change their diameter,disappear, and return shortly afterward.

SICUS

SICUS was performed according to a previously pub-lished25–27 method. Briefly, after the ingestion of contrast oralsolution (Promefarm, Milan, Italy) and after the contrast solu-tion was seen flowing through the ileum into the colon, a ret-rograde follow-through assessment of the entire small bowelwas performed visualizing, in a caudocranial sequence, thecontrast-filled ileal and jejunal loops. The position of patients

Inflamm Bowel Dis • Volume 11, Number 2, February 2005 Diagnostic Value of SICUS in Small Bowel Disease

© 2005 Lippincott Williams & Wilkins 147

was changed, and abdominal compression was used wheneverrequired to improve the visualization of any single loop and thedetection of intestinal abnormalities after the ingestion of theoral contrast solution.

Wall thickness and lumen diameter were measured atseveral sites in the small bowel (i.e., proximal, middle, anddistal) at the level of maximally distended, and not contracting,intestinal loops.

The criteria for the presence of small bowel lesions andthe recurrence of CD were as follows14,25–27:1. The presence of a target appearance, which was defined as

a strong echoic center surrounded by a sonolucent rim of>0.5 cm;

2. Increased wall thickness (>3 mm) with a description of thepresence/absence of wall layers (Fig. 1B);

3. The presence and distribution of intestinal folds;4. Stiff loop, which was identified by the presence of an intes-

tinal loop with increased wall thickness that was not dis-tended by the contrast solution;

5. Bowel dilatation, which was defined as a lumen diameter of>2.5 cm;

6. The presence of a bowel stricture, which was defined as alumen diameter of <1 cm, measured at the level of the maxi-mally distended loop independent of the presence of pre-stenotic dilatation (Fig. 2); and

7. Abnormal motor activity, which was identified by the ab-sence of intermittent variation in the intestinal loop profileand a lumen diameter that was associated with a variation inwall thickness and/or a change in the loop axial plane.

RadiologyA barium SBFT procedure was performed according to

our standard protocol. In fasting condition, the patient drank350 to 450 mL of a barium sulfate suspension (Bario HD;Bracco, Milan, Italy) under the intermittent fluoroscopic guid-ance of a radiologist. Spot radiographs of the esophagus, stom-

FIGURE 1. Patient RS: CD of the ter-minal ileum. A, contrast barium ra-diology: the lumen appears irregularbecause of the presence of narrowedsegments alternating with normalsegments. The mucosa surface ap-pears nodular with ulcers; a fistuliza-tion at the level of the ileocecal valveis present. B, top, SICUS: reduced lu-men diameter (arrow). B, bottom,the presence of oral contrast mate-rial allows the measurement of thelumen diameter and wall thickness(arrows). Note the multilayeredstructure of the intestinal wall.

FIGURE 2. CD of the proximal ileum. Short segment of theluminal narrowing (arrows) in the absence of prestenotic dila-tation is shown. The presence of oral contrast medium allowsthe accurate measurement of luminal diameter and the exten-sion of a stricture.

Pallotta et al Inflamm Bowel Dis • Volume 11, Number 2, February 2005

148 © 2005 Lippincott Williams & Wilkins

ach, and duodenum were obtained with the patient in the up-right and recumbent position. Radiographs then were obtainedwith the patient in the prone position at intervals of 15 to 20minutes for the first hour, and then every half hour until theterminal ileum and right colon were well-visualized. These ra-diographs were immediately evaluated, and fluoroscopy withcompression spot radiographs was performed when deemednecessary. The terminal ileum was routinely studied. The cri-teria for the presence of small bowel lesions were as follows6:1. Lesion of the intestinal wall, which was defined by the pres-

ence of aphthous ulcers, thickened and/or nodular folds, ul-cerations, mesenteric border rigidity, cobblestone appear-ance, or filling defects;

2. Fold flattening or thickening;3. Jejunoileal fold pattern reversal;4. The presence of stiffened or hypotonic loops;5. The presence of a bowel stricture, which was defined as a

lumen diameter <1 cm in maximally distended loop (Fig.1A); and

6. Bowel dilatation, which was defined as a lumen diameter of>4 cm at the level of the jejunum and >3.5 cm in the ileum.

Diagnosis and Analysis of DataIn patients who had undergone surgery, intraoperative

findings and surgical specimens were used as a “gold stan-dard” for determining the final diagnosis. In those patients whohad not undergone surgery, the final diagnosis was establishedby experienced gastroenterologists who took into consider-ation the findings of all diagnostic investigations that weredeemed necessary, including radiologic studies and follow-updata, but unaware of and not taking into account the US find-ings. The sensitivity and specificity of TUS and SICUS in de-tecting the presence of lesions were assessed in comparisonwith final diagnosis.

The consistency among radiology, TUS, and SICUS indetecting the number and anatomic sites of small bowel lesionswas evaluated using the index of the percentage of agreement(observed and by chance) and the concordance index by � sta-tistics.30 The quantitative significance of � was calculated ac-cording to Landis and Koch.30 The comparative assessment ofthe bowel lesion extensions that were measured with the 3 pro-cedures was evaluated by analysis of variance. Furthermore,the correlation between the US and radiologic extent of thesmall bowel lesions was assessed by the Pearson correlationcoefficient.

RESULTSA complete SICUS investigation was performed in all

patients 39 minutes (range, 12 to 90 min), on average, after theingestion of a mean (±SD) amount of 370 ± 108 mL of contrastsolution (range, 200 to 500 mL). TUS was performed in 15minutes, on average, and SBFT was performed in 2 hours

(range, 1 to 3 h), on average, after the ingestion of contrastmedium.

Diagnostic Accuracy of TUS and SICUS inPreviously Undiagnosed Patients

Ninety-one of the 148 enrolled patients (55 women; agerange, 12 to 89 yr) had not received any previous diagnosis.The diagnostic workup did not detect any small bowel lesionsin 56 patients. A final diagnosis of small bowel lesion wasmade in the remaining 35 patients (at surgery, 15 patients; atthe end of the diagnostic workup by gastroenterologists, 20patients).

The final diagnosis of all patients, and the sensitivity andspecificity of SICUS and TUS are reported in Table 1. TUSdetected at least 1 small bowel abnormality in 20 of the 35patients with an overall sensitivity and specificity for any smallbowel disease of 57% and 100%, respectively. SICUS detectedat least 1 small bowel abnormality in 33 of 35 patients with anoverall sensitivity and specificity of 94.3% and 98%, respec-tively. In 1 patient with celiac disease, and in 1 patient with anisolated duodenal polyp, SBFT and US did not detect any in-testinal lesion. At SBFT, 7 patients had �1 lesions localized inthe jejunum and proximal ileum, and 19 patients in the termi-nal ileum. Seven patients had involvement of the entire smallbowel. In the assessment of the number and site of lesions, thepercentage observed, and the chance proportions of agreementbetween radiology and SICUS findings were 1 and 0.52, re-spectively, with a � value of 1 (P < 0.001); the proportion ofagreement between radiology and TUS findings were 0.60 and0.44, respectively, with a � value of 0.28 (not significant). Six-teen patients met the diagnosis of CD. The sensitivity of TUSand SICUS in the correct detection of small intestine CD was68.7% and 100%, respectively. In patients with CD, the meanextension of the diseased small bowel was 20.9 ± 10.6 cm atSBFT, 20.3 ± 11.9 cm at SICUS (not significant versus SBFT),and 13.2 ± 4 cm at TUS (P < 0.0001 versus SICUS and SBFT).In the assessment of lesion extension, the correlation coeffi-cient between SBFT and TUS was 0.80 (P < 0.01), and be-tween SBFT and SICUS it was 0.90 (P < 0.001).

Diagnostic Accuracy of TUS and SICUS inPatients With Previously Diagnosed CD

At study enrollment, 57 of the 148 patients (23 women;age range, 18 to 88 yr) were known to have CD, with 36 pa-tients having isolated ileal disease, 5 having jejunal and proxi-mal ileal disease, 2 having colonic disease, and 14 having il-eocolonic disease (Table 2). The diagnosis of CD before studyenrollment was made at surgery in 27 of 57 cases, and at en-doscopy and histology performed on multiple biopsy speci-mens and at radiology in the remaining 30 patients. One patientwith previously diagnosed ileocolonic disease did not have asmall bowel recurrence of CD. In patients who had not yetbeen operated on, the mean follow-up time after initial diag-nosis was 6 years (range, 1.5 to 21 yr).



TUS detected at least 1 small bowel abnormality in 48 ofthe 55 CD patients with an overall sensitivity of 87.3%. SICUSdetected at least 1 small bowel abnormality in 54 of the 55 CDpatients with an overall sensitivity of 98.2% (Table 2).

Accuracy of TUS and SICUS in the Assessmentof Extension, Number, and Sites of SmallBowel Lesions in Patients With CD

In patients with CD, the mean extension of the diseasedsmall bowel was 27.7 ± 16.6 cm at SBFT, 28.6 ± 16.5 cm atSICUS (not significant versus SBFT), and 17 ± 8.9 cm at TUS(P < 0.0001 versus SICUS and SBFT). In the assessment oflesion extension, the correlation coefficient between SBFTand TUS was 0.59 (P < 0.001), and between SBFT and SICUSwas 0.88 (P < 0.001).

At SBFT, 12 patients had at least two jejunal and proxi-mal ileal CD involvements. SICUS detected the same lesionsin 10 patients, and TUS in 5 patients. SICUS missed a duode-nal lesion that was confirmed at endoscopy in 1 patient,whereas SBFT missed a jejunal lesion that was detected atSICUS and confirmed at surgery in another patient. SICUSand SBFT agreed in the assessment of the site of the lesions inall but 3 patients with a percentage observed, and chance pro-portion of agreement 0.94 and 0.65, respectively, with a �value of 0.83 (P < 0.001). In the assessment of the site of le-sions, the percentage observed and chance proportion of agree-ment between TUS and SBFT were 0.73 and 0.61, respec-tively, with a � value of 0.31 (not significant).

In the assessment of the number of lesions, the percent-age observed and chance proportion, and of agreement be-tween radiology and SICUS were 0.92 and 0.59, respectively,with a � value of 0.81 (P < 0.001); those between radiology andTUS were 0.71 and 0.60, respectively, with a � value of 0.27(not significant).

Accuracy of TUS and SICUS in the Assessmentof the Extension, Number of Lesions, and theSites of Small Bowel Lesions in CD PatientsUndergoing Surgery

During the study period, 15 patients with CD underwentsurgery. The mean extension of small bowel lesions was 28.5 ±14.2 cm at surgery, 29.4 ± 11 cm at SICUS (not significantversus surgery), and 14.5 ± 6.3 cm at TUS (P < 0.01 versussurgery and SICUS). In the assessment of lesion extension, thecorrelation coefficient between surgery and TUS was 0.20 (notsignificant), and between surgery and SICUS it was 0.85 (P <

TABLE 1. Sensitivity and Specificity of SICUS and TUS in Detecting Small Bowel Lesionsin Undiagnosed Patients*

Diagnosis No. (%)

SICUS TUS

TP FN TN FP TP FN TN FP

IBS 36 (39.6) — 0 35 1 — 0 36 0CD 16 (17.6) 16 0 — 0 11 5 — 0UC 7 (7.7) — 0 7 0 — 0 7 0Malignant tumors 6 (6.6) 6 0 — 0 3 3 — 0Celiac disease 5 (5.5) 4 1 — 0 2 3 — 0Indeterminate colitis 4 (4.4) — 0 4 0 — 0 4 0Polyps 4 (4.4) 3 1 — 0 0 4 — 0Others 13 (14.3) 4† 0 9‡ 0 4† 0 9‡ 0Total 91 (100) 33 2 55 1 20 15 56 0

*IBS, irritable bowel syndrome; UC, ulcerative colitis; TP, true positive; FN, false negative; TN, true negative;FP, false positive.

†One patient with isolated ileal ulcer (operative finding), 1 patient with intestinal tuberculosis (operative find-ing), 1 patient with idiopathic segmental dilatation of the ileum, and 1 patient with idiopathic pseudoobstruction(operative finding).

‡Patients met the final diagnosis as follows: obscure bleeding, 6 patients; appendicitis, 1 patient; gastric lym-phoma, 1 patient; constipation and fecaloma, 1 patient.

TABLE 2. True-Positive and False-Negative Findings of SICUSand TUS in Detecting Small Bowel Lesions in Patients WithKnown CD*

Lesion No. (%)

SICUS TUS

TP FN TP FN

Terminal ileum 18 (32.7) 17 1 14 4Ileum-jejunum 23 (41.8) 23 0 20 3Ileocolonic 14 (25.5) 14 0 14 0Total 55 (100) 54 1 48 7

*TP, true positive; FN, false negative.



0.001). Five patients had jejunal and proximal ileal lesions.SICUS correctly detected the sites of lesions in all patients, andTUS did so in 2 of them.

DISCUSSIONRadiologic examination of the small bowel is still con-

sidered to be the reference investigation in the diagnosis andfollow-up of small bowel lesions.1–11 However, due to radia-tion exposure, radiology is not indicated in children andwomen of child-bearing age, and is not to be frequently re-peated in the same subjects. Furthermore, it is not feasible toperform conventional contrast barium radiology in patients inwhom there is a suspicion of intestinal obstruction, or who aregreatly debilitated or giving poor cooperation, and it provideslimited information concerning the extent of the transmuraland perivisceral extension of disease such as CD or tumors.Because of these considerations, and because small bowel pa-thology is relatively infrequent, radiology of the small bowel isnot considered as a screening tool, and oftentimes not even asan early investigation in the diagnostic workup of patients inwhom an intestinal disease is suspected. The observation thatabout 70% of all radiologic investigations of the small bowelshow normal findings2,21,23,27 and that CD, the most prevalentchronic inflammatory bowel disease, is diagnosed 3 years afterthe onset of the symptoms,8 are indicative of the limited diag-nostic usefulness of radiologic assessment of the small bowelin general, and in the early symptomatic phase of CD. Further-more, due to the chronic recurrent course, patients with CDrequire frequent clinical monitoring, and radiology should beavoided at short time intervals.

In the last decade, TUS, which is a noninvasive, widelyavailable, and inexpensive technique, has been frequently per-formed as first-level investigation of small bowel disease. Todate, only 5 studies14,21–24 have evaluated the diagnostic accu-racy of TUS as a screening technique to detect undiagnosedsmall bowel disease. The sensitivity of TUS performed in ded-icated centers by qualified investigators with specific expertise

did not exceed 85%.24 It should be noted, however, that thisreported level of diagnostic sensitivity of TUS was achieved ininvestigations that were performed in populations of patientswho were mainly affected by CD, and false-negative findingsin celiac patients were not considered.24 Furthermore, whenthe sensitivity of TUS in detecting CD lesions was evaluatedaccording to the site of the disease, the sensitivity fell to 28.9%for proximal small bowel localization.

The results of the present study confirm a low sensitivityof 57% for TUS in previously undiagnosed patients when con-sidering patients with any disease, and a sensitivity of 68.7%when considering only patients with CD. In a previous pro-spective study22 that was performed in a series of 227 patientswith suspected or known disease of the upper and lower GItract, TUS images were considered to be of poor quality in35% of the patients due to obesity or intestinal gas. Further-more, the ability of TUS to identify the site of the small bowelwhere the lesion was located was satisfactory only in the ter-minal ileum. In the present study, low TUS sensitivity (Table1) was due mainly to the inability of TUS to identify and assessthe characteristic distribution of the valvulae conniventes inceliac patients, and to its inability to distinguish intestinal wallendoluminal structures such as polyps. Likewise, false-negative results of TUS occurred in patients with CD involve-ment that was limited to the mucosa and submucosa, and inpresenting with erosions at endoscopy or inflammatory infil-trates (Fig. 3) that caused a slight thickening of the intestinalwall. Because of the low negative predictive value22,24 and thelimited ability to identify the site of the lesions, TUS per-formed as an initial examination must be followed by smallbowel radiology.

The use of oral contrast medium overcomes the inherentlimitation of TUS because it allows the optimal transmission ofUS and, by distending the intestinal lumen, enables the visu-alization of the small bowel wall along its entire length. Dif-ferent from the method used in a previous study31 in which asimilar contrast medium was administered through a nasojeju-

FIGURE 3. Patient GB: CD of the ter-minal ileum. A, ileal biopsy showinga disproportionate inflammatory in-filtrate in the submucosa (black ar-rows), whereas the overlying mu-cosa is almost normal. B, SICUS: thethickness of the intestinal wall ap-pears slightly increased with a mul-tilayered structure (white arrows).



nal tube directly into the small bowel, the use of an orally in-gested contrast material offers a high diagnostic yield with anoninvasive imaging method. In this study, in which the finaldiagnosis was used as reference standard, the diagnostic sen-sitivity of SICUS in detecting small bowel lesions was 94%,with a specificity of 98%. SICUS and SBFT failed to demon-strate any intestinal lesion in 1 celiac patient with a history ofmild hypochromic anemia and in 1 patient with an isolatedduodenal polyp. When only the diagnosis of CD is considered,the sensitivity of SICUS reached 100%. It should be empha-sized that the patients who were evaluated in the present studywere outpatients and that a relevant number of them met thecriteria for a final diagnosis of irritable bowel syndrome,which is different from the findings of previous investiga-tions.22,24 The high diagnostic accuracy of SICUS and its levelof concordance with radiology findings in ruling out the pres-ence of small bowel lesions seem to be of clinical relevance,because an initial negative SICUS finding could help patientsavoid radiation exposure by not having the patients undergosmall bowel radiology.

TUS is considered a useful technique in the follow-up ofpatients with CD. The sensitivity of TUS in detecting intestinallesions in patients who have previously received diagnoses ofCD has been reported to vary from 80.8% to 95%.15–17,32–33

The present study had a comparable sensitivity (87%). Thehighest values for TUS sensitivity reported in previous studiescan be explained mainly by the absence19,32 of lesions in thejejunum and proximal ileum, which can be missed with TUS,in the patients investigated, as confirmed by the present study.In patients with previously diagnosed CD, the sensitivity ofSICUS in the assessment of lesions in the small intestine was98% with no false-positive findings, indicating that the oralcontrast material also greatly enhances the diagnostic accuracyof the US technique in the follow-up of CD patients.

Furthermore, from the direct comparison of TUS andSICUS with SBFT, only SICUS had good agreement in thedetection of the number and sites of small bowel lesions inpatients with CD of the small bowel. The agreement and �statistics in the comparison between SICUS and SBFT can beconsidered substantial in accordance with guidelines of Landisand Koch.30

In planning a surgical intervention in CD patients, it isrelevant to know the extension of the lesions, in addition to thenumber and sites of localization. To date, few studies have as-sessed the accuracy of TUS in the assessment of the extensionof the CD intestinal involvement, reporting variable results onthe correlation of TUS with radiology and intraoperative find-ings.19,24,34–35 Two recent studies19,24 that were performed inthe same group of patients reported a significant correlation(r = 0.51) between TUS and small bowel enema in the assess-ment of CD lesion extension. The present study, although notcontradicting such a correlation, has shown that when an ap-propriate test of comparative analysis is employed (i.e., analy-

sis of variance) the accuracy of TUS in the assessment of theextension of small bowel lesions is lower than that observedwith SICUS and confirmed by SBFT and surgical findings.

A limitation of the present study is that both TUS andSICUS were performed by the same operator, who might haveeither overestimated or underestimated the presence of lesionswith the former technique. However, this seems unlikely be-cause the sensitivity of TUS in the group of CD patients whohave previously received a diagnosis was comparable to thatreported in previous studies. In addition, despite the fact thatthe sonologist was not aware of the clinical suspicion in theundiagnosed patients, the specificity of TUS was 100% in thisgroup. It should also be pointed out that by performing TUSfirst, the sensitivity of SICUS could have been inadvertentlyincreased because the results of TUS were already known. De-spite this protocol limitation, the detection with SICUS of sev-eral lesions in areas other than those with lesions identifiedwith TUS is supportive of the high diagnostic yield of SICUS.In addition, the diagnostic yield of SICUS is not merely limitedto detecting the presence of lesions, but also to obtaining amore accurate definition of the number, sites, and extensionsof lesions. Nonetheless, it would be appropriate to assess theinterobserver variability of SICUS in further studies.

An additional limitation of this study is that US is anoperator-dependent technique, and the investigation of thesmall bowel by means of US requires relevant skills. The op-erator in the present study was an experienced sonologist withspecific expertise in small bowel investigation; however, aseparate study by our group (data in press) has indicated thatthe use of oral contrast medium by an inexperienced sonogra-pher makes SICUS more accurate than TUS performed by anexperienced sonographer.

Compared with TUS, SICUS is a time-consuming tech-nique, with an average duration of examination of 40 minutes.However, the comprehensive detailed information offered bySICUS, which is at least comparable to that offered by radiol-ogy and is definitely superior to that from TUS, suggests thatthis technique should be used as a first-level investigation, thushelping many patients to avoid subsequent radiology and sav-ing time, costs, and radiation exposure.

Confirming the result of our previous study,27 the smallamount of oral contrast medium that was used, which did notexceed 500 mL, could be ingested by all patients and did notcause discomfort or side effects in any of them.

In conclusion, the diagnostic accuracy of SICUS is com-parable to that of a radiologic examination and is superior tothat of TUS in detecting multiple lesions and those lesions lo-cated in the jejunal and proximal ileal loops. The results of thepresent study support the use of the noninvasive, widely avail-able, inexpensive, and probably cost-effective SICUS as aninitial investigation in patients with suspected small bowel dis-ease and in the follow-up of patients with CD. In the latter



group, SICUS can offer accurate information that can be usedwhen planning a surgical intervention.

ACKNOWLEDGMENTSThe authors thank Dr. A. Buffone and Dr. G. Milite for

assistance in the analysis of the data.

REFERENCES1. Ott DJ, Chen VM, Gelfand DW, et al. Detailed per-oral small bowel ex-

amination vs entereroclysis. Radiology. 1985;155:29–34.2. Toms AP, Barltrop A, Freeman AH. A prospective randomised study

comparing enteroclysis with small bowel follow-trough examinations in244 patients. Eur Radiol. 2001;11:1155–1160.

3. Hara AK, Leighton JA, Sharma VK, et al. Small bowel: preliminary com-parison of capsule endoscopy with barium study and CT. Radiology.2004;230:260–265.

4. Bernstein CN, Boult IF, Greenberg HM, et al. A prospective radomizedcomparison between small bowel enteroclysis and small bowel follow-through in Crohn’s disease. Gastroenterology. 1997;113:390–398.

5. Magliante DDT, Chernish SM, Kelvin FM, et al. Crohn disease of thesmall intestine: accuracy and relevance of enteroclysis. Radiology. 1992;184:541–545.

6. Goldberg HI, Caruthers SB, Nelson JA, et al. Radiographic findings of thenational cooperative Crohn’s disease study. Gastroenterology. 1979;77:925–937.

7. Chernish SM, Magliante DDT, O’Connor KW. Evaluation of the smallintestine by enteroclyis for Crohn’s disease. Am J Gastroenterol. 1992;87:696–701.

8. Podolsky DK. Inflammatory bowel disease. N Engl J Med. 1991;325:928–937.

9. Halligan S, Nicholls S, Beattle RM, et al. The role of small bowel radiol-ogy in the diagnosis and management of Crohn’s disease. Acta Paediatr.1995;84:1375–1378.

10. Hallagn S, Saunder BP, Williams CB, et al. Adult Crohn’s disease: canileoscopy replace the small bowel radiology? Abdom Imaging. 1998;23:117–121.

11. Loftus EV, Schoenfeld P, Sandborn WJ. The epidemiology and naturalhistory of Crohn’s disease in population-based patient cohorts from NorthAmerica: a systematic review. Aliment Pharmacol Ther. 2002;16:51–60.

12. Farmer RG, Whelan G, Fazio VW. Long-term follow-up of patients withCrohn’s disease. Gastroenterology. 1985;88:1818–1825.

13. IARC Programme on the Evaluation of Carcinogens Risks to Humans.Ionizing Radiation: Part I. X- and Gamma (�)-Radiation, and Neutron.Lyon, France: International Agency for Research on Cancer; 1999:75.

14. Sonnenberg A, Erckenbrecht J, Peter P, et al. Detection of Crohn’s diseaseby ultrasound. Gastroenterology. 1982;83:430–434.

15. Pera A, Cammarota T, Comino E, et al. Ultrasonography in the detectionof Crohn’s disease and in the differential diagnosis of inflammatory boweldisease. Digestion. 1988;41:180–184.

16. Papi C, Iscaro D, Salvatori V, et al. Sonographic evaluation of Crohn’sdisease. Ital J Gastroenterol. 1989;21:257–262.

17. Schwerk WB, Bech K, Raith M. A prospective evaluation of high resolu-tion sonography in the diagnosis of inflammatory bowel disease. Eur JGastroenterol Hepatol. 1992;4:173–182.

18. Gashe C, Moser G, Turetschek K, et al. Transabdominal bowel sonogra-phy for the detection of intestinal complications in Crohn’s disease. Gut.1999;44:112–117.

19. Parente F, Maconi G, Bollani S, et al. Bowel ultrasound in assessment ofCrohn’s disease and detection of related small bowel strictures: a prospec-tive comparative study versus x ray and intraoperative findings. Gut.2002;50:490–495.

20. Maconi G, Carsana L, Fociani P, et al. Small bowel stenosis in Crohn’sdisease: clinical, biochemical and ultrasonographic evaluation of histo-logical features. Aliment Pharmacol Ther. 2003;18:749–756.

21. Sheridan MB, Nicholson DA, Martin DF. Transabdominal ultrasonogra-phy as the primary investigation in patients with suspected Crohn’s dis-ease or recurrence: a prospective study. Clin Radiol. 1993;48:402–404.

22. Hollerbach S, Geissler A, Schiegl H, et al. The accuracy of abdominalultrasound in the assessment of bowel disorders. Scand J Gastroenterol.1998;33:1201–1208.

23. Astegiano M, Bresso F, Cammarota T, et al. Abdominal pain and boweldysfunction: diagnostic role of intestinal ultrasound. Eur J GastroenterolHepatol. 2001;13:927–931.

24. Parente F, Greco S, Molteni M, et al. Role of the early ultrasound in de-tecting inflammatory intestinal disorders and identifying their anatomicallocation within the bowel. Aliment Pharmacol Ther. 2003;18:1009–1016.

25. Pallotta N, Baccini F, Corazziari E. Contrast ultrasonography of the nor-mal small bowel. Ultrasound Med Biol. 1999;25:1335–1340.

26. Pallotta N, Baccini F, Corazziari E. Small intestine contrast ultrasonog-raphy. J Ultrasound Med. 2000;19:21–26.

27. Pallotta N, Baccini F, Corazziari E. Small intestine contrast ultrasonog-raphy (SICUS) in the diagnosis of small intestine lesions. Ultrasound MedBiol. 2001;27:335–341.

28. Cittadini G, Giasotto V, Garlaschi G, et al. Transabdominal ultrasonog-raphy of the small bowel after oral administration of a non-absorbableanechoic solution: comparison with barium enteroclysis. Clin Radiol.2001;56:225–230.

29. Shivananda S, Lennard-Jones J, Logan R, et al. Incidence of inflammatorybowel disease across Europe: is there a difference between north andsouth? Results of the European Collaborative Study on InflammatoryBowel Disease (EC-IBD). Gut. 1996;39:690–697.

30. Kramer SM, Feinstein AR. Clinical biostatistics LIV: the biostatistics ofconcordance. Clin Pharmacol Ther. 1981;29:111–123.

31. Folvik G, Bjerke-Larssen T, Odegaard S, et al. Hydrosonography of thesmall intestine: comparison with radiological barium study. Scand J Gas-troenterol. 1999;34:1247–1252.

32. Solving J, Ekberg O, Lindgren S, et al. Ultrasound examination of thesmall bowel: comparison with enteroclysis in patients with Crohn disease.Abdom Imaging. 1995;20:323–326.

33. Maconi G, Parente F, Bollani S, et al. Abdominal ultrasound in the assess-ment of extent and activity of Crohn’s disease:clinical significance andimplication of bowel wall thickening. Am J Gastroenterol. 1996;91:1604–1609.

34. Pedersen Hojlund B, Gronvall S, Dorph S, et al. The value of dynamicultrasound scanning in Crohn’s disease. Scand J Gastroenterol. 1986;21:969–972.

35. Lim Hoon J, Tae Ko Y, Lee Ho D, et al. Sonography of inflammatorybowel disease: findings and value in differential diagnosis. AJR Am JRoentgenol. 1994;163:343–347.



small intestine contrast ultrasonography

Documents