Clinical Value of Technetium-99m-HMPAO-Labeled Leukocyte Scintigraphy and SpiralComputed Tomography in Active Crohn’s DiseaseTamas Molnar, M.D., Miklos Papo´s, M.D., Ph.D., Csaba Gyulai, M.D., Edit Ambrus, M.D., Ph.D.,Lilla Kardos, M.D., Ferenc Nagy, M.D., Ph.D., Andra´s Palko, M.D., Ph.D., Laszlo Pavics, M.D., Ph.D., andJanos Lonovics, M.D., Ph.D.First Department of Medicine, Faculty of Medicine, University of Szeged, Szeged, Hungary

OBJECTIVES: The diagnostic accuracy of technetium-99m-HMPAO-labeled leukocyte scintigraphy (LS) and spiral CTfor the detection of inflammatory activity was assessed; theextent of the inflammation and the complications were com-pared with the clinical and laboratory parameters and withthe endoscopic and radiological findings in patients withclinically active Crohn’s disease (CD).

METHODS: Twenty-eight patients (13 men, 15 women, av-erage age 32.5 yr, range: 18–59 yr) with an acute exacer-bation of CD were enrolled in the study. The disease be-havior type and the maximum extent of inflammation wereestablished by means of endoscopy (jejunoscopy andcolonoscopy) and enteroclysis. Nine patients with severecomplications (abscess and stenosis) underwent operation.The GI tract was divided into five segments (small bowel,ascending colon, transverse colon, descending colon, andrectosigmoid), the LS, CT, endoscopic, and radiologicalpictures of all segments were graded (range: 0–3) and thescores were summed and compared.

RESULTS: The investigations indicated that LS and CT hadsensitivities of 76.1% and 71.8%, specificities of 91.0% and83.5%, and accuracies of 82.6% and 77.5%, respectively,for detection of segmental inflammatory activity. With re-gard to the disease behavior type, the sensitivities of LS andCT were, respectively, 77% and 100% in the penetrating-fistulizing, 80% and 73% in the stricturing, and 68% and64% in the inflammatory form of CD. CT detected allabdominal abscesses, whereas the diagnostic value of LS forthe detection of the complications of CD was lower. Theinflammatory activity scores measured by LS displayed acloser correlation than that of CT with the Best index (r 50.71, p , 0.0005vs r 5 0.63, p , 0.001), the van Heesindex (r 5 0.61, p , 0.005vs r 5 0.59, p , 0.005), theserum fibrinogen level (r 5 0.67,p , 0.005vs r5 0.59,p ,0.005), or the C-reactive protein level (r 5 0.64,p , 0.005vs r 5 0.51,p , 0.01).

CONCLUSIONS: Both LS and CT are valuable noninvasivediagnostic methods in cases involving severe, active CD. LSseemed better for the detection of segmental inflammatory

activity, whereas CT displayed excellent suitability for therecognition of complications: abdominal abscesses werediagnosed with 100% efficiency. (Am J Gastroenterol 2001;96:1517–1521. © 2001 by Am. Coll. of Gastroenterology)

INTRODUCTION

Crohn’s disease (CD) is a chronic transmural granulomatousinflammatory bowel disease of unknown etiology, with al-ternating periods of symptomatic exacerbations and clinicalremissions. There are no typical specific characteristics ofan acute attack, because of the different clinical courses andthe heterogeneous localization of the disease. Therefore, thecorrect therapy of an exacerbation necessitates a knowledgeof the maximum extent of involvement, the behavior type(inflammatory, stricturing, or penetrating), the current clin-ical activity, and the presence of complications. Althoughendoscopic and radiological examinations are the gold stan-dard procedures for determination of the localization and theseverity of inflammation, these methods are not so valuablefor the detection of complications from CD, such as abdom-inal abscesses or suspected fistulas. Furthermore, bowelpreparations are relatively contraindicated in patients withsevere active disease. Imaging techniques, which do notimpose a great load on the patients, are therefore of highclinical relevance in the follow-up of CD patients, andespecially in those who undergo a severe active relapse.Noninvasive imaging methods, such as isotope-labeled leu-kocyte or granulocyte scintigraphy and spiral CT are usefulin cases involving a suspected complication (1, 2), but theaccuracy of these techniques in different types of active CDhas not yet been well established. Only one comparativeprospective study is to be found in the literature in which thetwo methods were performed simultaneously and the resultswere compared with endoscopic and operative findings inpatients with inflammatory bowel disease (3).

The aim of the present study was to evaluate and tocompare the diagnostic accuracy of technetium-99m (99m-Tc)-HMPAO-labeled leukocyte scintigraphy (LS) and spiralCT on the basis of the clinical parameters and the endo-

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 96, No. 5, 2001© 2001 by Am. Coll. of Gastroenterology ISSN 0002-9270/01/$20.00Published by Elsevier Science Inc. PII S0002-9270(01)02312-7

scopic, radiological, and surgical findings in patients withclinically active CD.

MATERIALS AND METHODS

PatientsTwenty-eight patients who had suffered an acute relapse ora severe first attack of CD were eligible for inclusion in thestudy during the period from January 1998 to March 1999.Their CD was diagnosed on the basis of the standard clin-ical, radiological, endoscopic, and histological criteria (4).Jejunoscopy, small bowel enteroclysis, colonoscopy, LS,and CT were performed in all patients within 7 days ofadmittance. The median age of the patients was 32.5 yr(range: 18–59 yr). The median duration of the CD was 7.47yr (range: 0.1–25 yr).

Clinical ParametersAfter admission, a blood sample was taken for determina-tion of the sedimentation rate, Hb content, thrombocyte andleukocyte counts, C-reactive protein (CRP), and serum iron,albumin, and fibrinogen levels. Crohn’s disease activityindex (CDAI) (5) and van Hees index (6) were also calcu-lated as disease activity parameters. The above investiga-tions demonstrated the extent of involvement of the inflam-matory process in all patients. The maximum extents of CDwere as follows: seven patients with an isolated smallbowel, mainly ileal localization (five ileal, one jejunal andileal, one duodenal and ileal), 10 with a colonic localization,and 11 with an ileocolonic localization. The clinical char-acteristics are summarized in Table 1.

Endoscopy, Radiology, and OperationOne endoscopist performed all jejunoscopies and colonos-copies to assess the extent of CD and to classify the endo-scopic mucosal inflammatory changes. Four bowel seg-ments (the cecum and ascending colon, the transverse colon,the descending colon, and the sigmoid colon and rectum)and the duodenum–upper part of the jejunum were scoredby protocol of Mary and Modigliani (7) with some modifi-cation (05 normal mucosa, 15 edema, erythema, granu-larity of the mucosa, and aphthous lesions, 25 sporadic or

superficial ulcerations, and 35 extensive deep ulcerationsand stenosis). The severity of the inflammation in the smallbowel was established by enteroclysis in 23 patients,whereas in five patients the terminal ileum was also in-spected during colonoscopy. The radiological findings wereclassified by the same radiologist as normal (0), edema andgranularity of the mucosa (1), ulcerations (2), and stenosis(3) (8). The maximum count of the duodenojejunal and ilealscores was used to measure the small bowel activity. Withthese diagnostic methods, we were able to quantify theseverity of inflammation in each part of the GI tract. Lap-arotomy was performed on nine patients during their hos-pitalization, because of suspected abscess in five, and severestenosis in four cases. All patients with a suspected intra-abdominal abscess were operated on, and the intraoperativefindings served as gold standard.

Involved segments were divided into three groups accord-ing to the behavior of the inflammation: inflammatory (non-stricturing, nonpenetrating) type, stricturing type, and pen-etrating-fistulizing type (9).

Technetium-99m-HMPAO-Labeled LeukocyteScintigraphyIn vitro leukocyte labeling was carried out by a routinemethod (10), with some modification. Sixty milliliters ofvenous blood was collected in a syringe containing 6 ml of3.8% sodium citrate and 12 ml of 6% hydroxyethyl starchsolution. After spontaneous sedimentation of the erythro-cytes, the supernatant was centrifuged at 100g for 5 min,and the mixed leukocyte pellet was collected. 99m-Tc-HMPAO was formed by adding 99m-Tc in 1–1.5 ml toHMPAO (Leuco-Scint OSSKI, Budapest, Hungary). Mixedleukocytes were resuspended in 1 ml of 99m-Tc-HMPAO.After incubation with careful shaking for 10 min at roomtemperature, the unbound 99m-Tc-HMPAO was removedby centrifugation (450g, 5 min). After slow reinjection intothe patients (mean activity: 345 MBq; range: 208–614MBq), images were taken in the anterior view after 30 minand 2 h. The maximal radiation dose was 2.2 mSv. Theinflammatory process indicated by LS was localized in fivesegments: the small intestine, the ascending, the transverseand descending colon, and the rectosigmoid. The leukocyteuptake of each segment was scored relative to the normalbone marrow uptake (05 no uptake; 15 less than normalbone marrow uptake; 25 normal bone marrow uptake; 35bone marrow uptake) (11). The LS activity index was cal-culated by summing the segment scores.

Spiral Computed TomographyEvery CT examination was performed with a helical CTscanner (Somatom Plus 4; Siemens, Erlangen, Germany)after the administration of oral contrast material. Patientsreceived 2 L of diluted sodium amidotrizoate (megluminamidotrizoate, Gastrografin; Schering AG, Germany) 2 hbefore the scanning, and drank it slowly up to the beginningof the examination. After an unenhanced study, 100 ml of

Table 1. Clinical Characteristics of Patients

Characteristics Data

No. of patients (F/M) 28 (15/13)CDAI (mean6 SD) 230.266 125.65Van Hees index (mean6 SD) 164.816 67.09ESR (mean6 SD), (mm/h) 43.286 30.50Hb (mean6 SD) (g/L) 11.986 2.31Thrombocyte count (mean6 SD) (3109/L) 411.286 158.58Leukocyte count (mean6 SD) (3109/L) 10.466 4.10Serum iron (mean6 SD) (mg/L) 7.936 6.54Serum albumin (mean6 SD) (g/L) 40.396 7.42CRP (mean6 SD) (mg/L) 63.646 81.01Fibrinogen (mean6 SD) (g/L) 5.386 1.64

CDAI 5 Crohn’s disease activity index; ESR5 erythrocyte sedimentation rate;CRP5 C-reactive protein.

1518 Molnar et al. AJG – Vol. 96, No. 5, 2001

Iopromide (Ultravist, Schering AG) was administeredi.v.The slice thickness was 10 mm, the pitch was 1, and thereconstruction interval was also 10 mm. The maximal radi-ation dose was 8 mSV. The images were reviewed by tworadiologists who were unaware of the clinical symptoms.The thickness, appearance, and enhancement of the bowelwall, and the mesenteric changes, lymph nodes, fistulas, andabscesses were assessed as indicators of the inflammatoryactivity, and were scored for each of the bowel segments ona 4-point scale, adopted from the study of Kolkmanet al.(3): 0 5 no thickening of the bowel wall, and a normalmesentery; 15 a thickened bowel wall, a homogenousaspect, no enhancement with intravenous contrast, and nodouble-halo sign; 25 a thickened bowel wall, enhancementwith intravenous contrast or a double-halo sign, ulcerationand mesenteric fibrofatty proliferation; and 35 a thickenedbowel wall, enhancement with intravenous contrast, ulcer-ation and mesenteric fibrovascular strands.

No clinical information was provided to the observers;only the gastroenterologists knew the clinical histories ofthe patients.

StatisticsData are expressed as means6 SD or as means and range.Correlation coefficients were calculated by using the Spear-man rank test.

RESULTS

All investigations were well tolerated and there were nocomplications. Endoscopy or enteroclysis gave a positiveresult in at least one segment of the GI tract in all patients.Both LS and CT gave positive results in 25 (but different)patients (25 of 28 [89.3%]). Endoscopy or enteroclysisrevealed positive results in 71 segments. LS gave positiveresults in 54 of these 71 segments, whereas CT did so in 51segments. The distribution of the positive segments is givenin Figure 1. Endoscopy and enteroclysis did not indicate anyabnormality in 67 segments. LS gave negative results in 60of these segments, whereas CT did so in 56 segments.

Table 2 presents the correlations between the LS and CT

scores and the endoscopic (and radiological) activity scores.Analysis of the above data demonstrated that the sensitivityand specificity of LS were 76.1% and 91.0%, whereas thoseof CT were 71.8% and 83.5%. The accuracy of LS was82.6%, and that of CT was 77.5%. The sensitivity andspecificity of LS and CT differ in the various anatomicallocalization of the involved segments: LS was better incolonic inflammation, whereas CT was better in small bowelprocesses (Table 3).

The sum of the endoscopic and radiological activityscores was 159, that of the LS scores was 121, and that ofthe CT scores was 120. Endoscopic and radiological activityscores correlated significantly with LS activity scores (r 50.61, p , 0.005) and CT activity scores (r 5 0.58, p ,0.005), and also with the number of involved segmentsdetected by LS (r 5 0.61,p , 0.005) or by CT (r 5 0.59,p , 0.005). On the basis of the endoscopic, radiological, andoperative findings, the 72 involved segments were dividedinto three groups according to disease behavior. There were44 inflammatory, 15 stricturing, and 13 penetrating-fistuliz-ing segments. With regard to the disease behavior type, thesensitivities of LS and CT were 77% and 100% in thepenetrating-fistulizing, 80% and 73% in the stricturing, and68% and 64% in the inflammatory form of CD. LS detectedthe inflammatory and stricturing processes more correctly,whereas the CT exhibited an excellent sensitivity in thepenetrating-fistulizing form of inflammation.

Intraoperatively, five abscesses and four fistulas werefound. The rate of detection of abscesses with CT wasexcellent: all were detected without false-positive results,

Figure 1. Anatomic distribution of the positive segments (SB5small bowel; AC 5 ascending colon; TC5 transverse colon;DC 5 descending colon; RS5 rectum and sigmoid colon; E/E5endoscopy or entreoclysis; LS5 leukocyte scintigraphy.)

Table 2. Segmental Inflammatory Activity as Determined WithLS and CT Compared to the Endoscopic and HistologicalFindings

Endoscopy 0 1 2 367 11 32 28

LS 0 61 6 7 41 1 3 6 52 4 1 11 133 1 1 8 6

CT 0 56 7 8 51 5 3 8 52 4 1 12 73 2 0 4 11

LS 5 leukocyte scintigraphy; CT5 computed tomography.

Table 3. Sensitivity and Specificity of LS and CT in theDifferent Anatomic Localization

SB AC TC DC RS

Sensitivity (%)LS 72 80 85 75 70CT 83 80 71 58 65

Specificity (%)LS 90 100 100 81 73CT 90 75 86 94 82

LS 5 leukocyte scintigraphy; CT5 computed tomography; SB5 small bowel; AC5ascending colon; TC5 transverse colon; DC5 descending colon; RS5 sigmoidcolon and rectum.

1519AJG – May, 2001 Clinical Value of LS and CT in Crohn’s Disease

whereas three of four fistulas were diagnosed preopera-tively, again without false-positive results (sensitivity 100%and 75%, specificity 100%). LS indicated three of fiveabscesses and two of four fistulas, without false-positivefindings (sensitivity 60% and 50%, specificity 100%).

Correlations were calculated and analyzed between LSand CT activity scores and CDAI and both clinical activityindices and each laboratory parameter. Inflammatory activ-ity scores measured by LS displayed a closer correlationthan did CT with the Best index (r 5 0.71,p , 0.0005vsr 5 0.63,p , 0.001), the van Hees index (r 5 0.61,p ,0.005vs r 5 0.59,p , 0.005), the serum fibrinogen level(r 5 0.67,p , 0.005,vs r 5 0.59,p , 0.005), or the CRPlevel (r 5 0.64, p , 0.005,vs r 5 0.51, p , 0.01). Thesignificant correlations are presented in Table 4.

DISCUSSION

Provision of the correct therapy of an acute relapse of CDposes a serious problem for the therapist because of the greatvariability of the extent and the association with possiblysevere complications. A correct diagnostic method shoulddetermine the maximum spread of the inflammatory processand should dependably reveal fistulas and abscesses withoutimposing more than minimal strain on these severely illpatients. Either CT or LS is able to meet these requirements(12–14), but the radiation dose of CT is about 4-fold higher.The comparison of these techniques was necessary becauseonly 17 such CD patients had been subsequently reported, inwhom the two methods were performed simultaneously.The sensitivity that we observed for LS is similar to therange reported in previous studies: 79–92% (15, 16). How-ever, the sensitivity of CT as concerns the segmental in-flammatory activity was earlier assessed only by Kolkmanet al. They reported a sensitivity of 68%, but the diseasebehavior type of their patients was not stated. However, thisseems a very important factor. In our work the sensitivity ofCT in the different disease types ranged from 64% in theinflammatory form to 100% in the penetrating form of CD.It seems that different types of Crohn’s disease (e.g., in-

flammatory and stricturing) may occur simultaneously in asingle patient. The management of the patient is determinedby the complication (stenosis or fistula), so that classifica-tion of each segment is usually not necessary in clinicalpractice. On the other hand, a correct measurement of thevalue of different diagnostic methods does necessitate thebiological classification of each segment.

In our prospective comparative study LS more accuratelydetected the extent of GI involvement caused by CD. CTgave a better result only when there was small bowel in-volvement and in segments with penetrating behavior. LSpresented a relatively low sensitivity to detect abscesses(three from five). Technetium-99m has a relatively shorthalf-life of 6 h, but at neutral pH Tc-99m and HMPAO forma lipophilic complex that is rapidly incorporated into theleukocytesin vitro. Therefore, Tc-HMPAO-labeled leuko-cytes migrate to the site of inflammation within 30 min, butthey disappear from there rapidly. In a number of studiesmade with this isotope technique, it was reported that thescintigrams were prepared in 1 and in 2–4 h. Later scanningcan be performed if indium-111 is used, because this isotopehas a half-life of 65 h. This method might well have resultedin superior results as it concerns the detection of abscesses.However, indium is more expensive and not easily avail-able, and a large degree of self-radiation has been encoun-tered.

Although both methods somewhat underestimated theseverity of the inflammation and the length of the involve-ment, the accuracy of both diagnostic procedures was suf-ficient for adequate therapy. CT displayed excellent suit-ability for the recognition of extraintestinal complications.

The CT and LS scores and the numbers of involvedsegments detected by CT and LS in our study correlatedwell, not only with the endoscopic results and laboratoryparameters, but also with the CDAI and the van Hees index.Although several researchers claim that the most frequentlyused indices correlate only poorly with the endoscopic find-ing (17), and have little clinical relevance, our results sug-gest just the opposite.

In conclusion, our results indicate that, in a severe first

Table 4. Correlation Between Activity Indices, Laboratory Parameters, and Endoscopic, LS, and CT Scores

Van Hees CDAI Fibrinogen CRP SedimentSe

Albumin Se Iron Thr. Count

EAS r 5 0.87 r 5 0.78 r 5 0.78 r 5 0.71 R5 0.80 r 5 20.68 r 5 20.78 r 5 0.73p , 0.00001 p , 0.0001 p , 0.0001 p , 0.0005 p , 0.00001 p , 0.0005 p , 0.00001 p , 0.0005

LSAS r 5 0.61 r 5 0.71 r 5 0.67 r 5 0.64 R5 0.59 r 5 0.57 r 5 20.50 r 5 0.59p , 0.005 p , 0.0005 p , 0.0005 p , 0.001 p , 0.005 p , 0.005 p , 0.01 p , 0.005

LSEX r 5 0.56 r 5 0.66 r 5 0.59 r 5 0.56 R5 0.60 r 5 20.46 r 5 20.52 r 5 0.53p , 0.005 p , 0.001 p , 0.005 p , 0.005 p , 0.005 p , 0.05 p , 0.01 p , 0.01

CTAS r 5 0.57 r 5 0.63 r 5 0.59 r 5 0.51 R5 0.51 r 5 20.59 r 5 20.52 NSp , 0.005 p , 0.001 p , 0.005 p , 0.01 p , 0.01 p , 0.005 p , 0.01

CTEX r 5 0.52 r 5 0.63 r 5 0.58 r 5 0.49 R5 0.56 r 5 20.49 r 5 20.56 NSp , 0.01 p , 0.001 p , 0.005 p , 0.05 p , 0.005 p , 0.05 p , 0.005

EAS5 endoscopic activity score; LSAS5 leukocyte scintigraphy activity score; LSEX5 leukocyte scintigraphy extension (number of involved segments detected by leukocytescintigraphy); CTAS5 computed tomography activity score; CTEX5 computed tomography extension (number of involved segments detected by CT; CDAI5 Crohn’s diseaseactivity index; CRP5 C 5 reactive protein; Sediment5 sedimentation; Thr5 thrombocyte.

1520 Molnar et al. AJG – Vol. 96, No. 5, 2001

attack of CD, when the radiological and endoscopic proce-dures are relatively contraindicated, both LS and CT areconvenient for the patient without any complications. Al-though both methods have a significant diagnostic value inactive CD, the accuracy of these investigations is not thesame in the different disease behavior types.

We suggest that in a severe active CD patient with un-known disease behavior type and localization, LS and CTshould be performed simultaneously. In a patient with anacute flare-up and a known disease behavior type, one ofthese procedures is sufficient, preferably LS in the inflam-matory and stricturing form, and CT in the penetrating formof CD.

Reprint requests and correspondence:Tamas Molnar, M.D.,First Department of Medicine, Faculty of Medicine, University ofSzeged, H-6720 Szeged, Kora´nyi fasor 8, Hungary.

Received May 30, 2000; accepted Oct. 30, 2000.

REFERENCES

1. Gore RM, Cohen MI, Vogelzang RL, et al. Value of computedtomography in the detection of complications of Crohn’s dis-ease. Dig Dis Sci 1985;30:701–9.

2. Scholmerich J, Schmidt E, Schu¨michen C, et al. Scintigraphicassessment of bowel involvement and disease activity inCrohn’s disease using Technetium 99m-hexamethyl propyleneamine oxine as leukocyte label. Gastroenterology 1988;95:1287–93.

3. Kolkman JJ, Falke THM, Ross JC, et al. Computed tomogra-phy and granulocyte scintigraphy in active inflammatorybowel disease. Dig Dis Sci 1996;41:641–50.

4. Lennard-Jones JL. Classification of inflammatory bowel dis-ease. Scand J Gastroenterol 1989;24(suppl):2–6.

5. Best WR, Becktel JM, Singleton JW, et al. Development of aCrohn’s disease activity index (National Cooperative Crohn’sDisease Study). Gastroenterology 1976;70:439–44.

6. van Hees PAM, van Elteren PM, van Lier HJ, et al. An indexof inflammatory activity in patients with Crohn’s disease. Gut1980;21:279–86.

7. Mary JY, Modigliani R. Development and validation of anendoscopic index of the severity for Crohn’s disease: A pro-spective multicentre study. Gut 1989;30:983–9.

8. Orel SG, Rubesin SE, Jones B, et al. Computed tomography vsbarium studies in the acutely symptomatic patient with Crohndisease. J Comput Assist Tomogr 1987;11:1009–16.

9. Sachar DB, Andrews HA, Farmer RG. Proposed classificationof patient subgroups in Crohn’s disease. Working team report.Gastroenterol Int 1992;5:141–54.

10. Schu¨michen C, Scho¨lmerich J. Tc 99 m HM-PAO labeling inleukocytes for detection of inflammatory bowel disease. NuclComp 1986;17:274–6.

11. Papo´s M, Nagy F, Na´rai G, et al. Antigranulocyte immu-noscintigraphy and 99mTC-hexamethilpropilenamine-oxim-labelled leukocyte scintigraphy in inflammatory bowel dis-ease. Dig Dis Sci 1996;41:412–20.

12. Fishman EK, EJ Wolf, Jones B, et al. CT evaluation ofCrohn’s disease. Effect on patient management. AJR Am JRoentgenol 1987;148:537–40.

13. Goldberg HI, Gore RM, Margulis AR, et al. Computed to-mography in the evaluation of Crohn’s disease. AJR Am JRoentgenol 1983;140:277–82.

14. Sciaretta G, Furno A, Mazzoni M, et al. Technetium-99mhexamethyl propylene amine oxime granulocyte scintigraphyin Crohn’s disease. Diagnostic and clinical relevance. Gut1993;34:1364–9.

15. Lannto E, Ja¨rvi IK, Krekela I, et al. Technetium-99m hexa-methyl propylene amine oxine leucocytes in the assessment ofdisease activity in inflammatory bowel disease. Eur J NuclMed 1993;20:766–9.

16. Gibson P, Lichtenstein M, Salehi N, et al. Value of positivetechnetium-99m-leucocyte scans in predicting intestinal in-flammation. Gut 1991;32:1502–7.

17. Brignola C, Iannone P, Pasquali S, et al. Clinical significanceand prognostic value of 111 In-labelled leukocyte scanning inCrohn’s disease: A prospective study. Eur J GastroenterolHepatol 1990;2:451–4.

1521AJG – May, 2001 Clinical Value of LS and CT in Crohn’s Disease