The first two authors contributed equally.
Detection and differentiation of inflammatory versus fibromatous Crohn's disease strictures: Prospective comparison of 18F-FDG-PET/CT, MR-enteroclysis, and transabdominal ultrasound versus endoscopic/histologic evaluation
Article first published online: 22 FEB 2012
Copyright © 2012 Crohn's & Colitis Foundation of America, Inc.
Inflammatory Bowel Diseases
Volume 18, Issue 12, pages 2252–2260, December 2012
How to Cite
Lenze, F., Wessling, J., Bremer, J., Ullerich, H., Spieker, T., Weckesser, M., Gonschorrek, S., Kannengieβer, K., Rijcken, E., Heidemann, J., Luegering, A., Schober, O., Domschke, W., Kucharzik, T. and Maaser, C. (2012), Detection and differentiation of inflammatory versus fibromatous Crohn's disease strictures: Prospective comparison of 18F-FDG-PET/CT, MR-enteroclysis, and transabdominal ultrasound versus endoscopic/histologic evaluation. Inflamm Bowel Dis, 18: 2252–2260. doi: 10.1002/ibd.22930
- Issue published online: 15 NOV 2012
- Article first published online: 22 FEB 2012
- Manuscript Accepted: 2 FEB 2012
- Manuscript Received: 12 JUN 2011
- Crohn's disease;
- positron emission tomography;
- computed tomography;
- magnetic resonance imaging;
Differentiation between inflammatory and fibromatous strictures in Crohn's disease (CD) is difficult but crucial for therapeutic decisions. The aim of this study was to assess the best noninvasive imaging method for the detection and differentiation of inflammatory and fibromatous stenoses in CD in comparison to endoscopic and histologic evaluation.
Patients with suspected CD strictures were included. According to a formalized endoscopic and histologic protocol, strictures were classified as inflammatory, mixed, and fibrostenotic. Strictures were further analyzed using fluorine 18-labeled fluoro-2-deoxy-D-glucose (18FDG) / positron emission tomography (PET) low-dose computed tomography (CT), magnetic resonance (MR) enteroclysis and transabdominal ultrasound using standardized scoring systems.
Thirty patients with 37 strictures were evaluated (inflamed n = 22; mixed n = 12, fibromatous n = 3). 18FDG-PET/CT detected 81%, MR-enteroclysis 81%, and ultrasound 68% of the strictures. Correct differentiation rates of strictures were 57% for MRE, 53% for 18FDG-PET/CT, and 40% for ultrasound. Differences of detection rates and differentiation rates were not statistically significant. When combining transabdominal ultrasound with 18FDG-PET/CT or MR-enteroclysis all strictures that required invasive treatment were detected.
Detection rates of the strictures were not significantly different between 18FDG-PET/CT, MR-enteroclysis, and ultrasound. Despite good stricture detection rates relating to our gold standard, 18FDG-PET/CT nor MR-enteroclysis nor ultrasound can accurately differentiate inflamed from fibrotic strictures. A combination of MR-enteroclysis and ultrasound as well as a combination of 18FDG-PET/CT and ultrasound resulted in a 100% detection rate of strictures requiring surgery or endoscopic dilation therapy, suggesting the combination of these methods as an alternative to endoscopy at least in the group of patients not able to perform an adequate bowel preparation. (Inflamm Bowel Dis 2012;)
The formation of intestinal strictures represents one of the most common complications of Crohn's disease (CD).1 Therapy for symptomatic CD strictures remains challenging.2 In clinical practice, due to different therapeutic approaches, we differentiate between inflammatory and fibrotic strictures. For a fibrostenotic stricture medical treatment would be ineffective and, therefore, surgical resection or endoscopic dilation therapy are the only reasonable treatment options. Although surgery is highly efficacious, one has to keep in mind the risk of stricture recurrence, as the risk of a second resection is close to 45% within 10 years following primary stricture resection.3–5 Therefore, in case of an inflammatory stricture medical treatment seems to be the optimal treatment strategy despite the potential risk of developing an ileus, requiring emergency surgery, and the risk of side effects due to immunosuppressive treatment, e.g., opportunistic infections and the increased risk of developing a malignant disease,6 while those risks would outweigh the benefit in case of fibrotic strictures.
It is therefore crucial to differentiate fibromatous from inflammatory strictures. In clinical practice a treatment decision is currently based on patient history, the results of endoscopic evaluation, the response to previous medical treatment, and the results of established imaging methods, such as transabdominal ultrasound, computed tomography (CT), or magnetic resonance (MR)-enteroclysis, although the sensitivity and specificity of these methods regarding this differentiation have not been well studied. In addition, new imaging techniques that also study tissue metabolism have evolved. For example, fluorine 18-labeled fluoro-2-deoxy-D-glucose (18FDG) / positron emission tomography (PET)/CT is widely used in oncology for tumor diagnostics and staging.7–9 18FDG-PET/CT combines metabolic information by visualizing glucose metabolism with anatomic data from CT scanning (Fig. 1). Our group first presented a retrospective analysis of the potential role of 18FDG-PET/CT in CD for differentiation between inflammatory and fibromatous strictures as an abstract during the Digestive Disease Week in 2005.10 Until now, only a few studies, mostly evaluating 18FDG-PET/CT for detecting inflammatory activity in inflammatory bowel disease (IBD), have been available.11–16 The aim of our present investigation was to elucidate the potential role of 18FDG-PET/CT in stricture detection as well as in stricture differentiation with respect to inflammation versus fibrosis in a prospective study comparing the results with those of established imaging methods, namely, MR-enteroclysis, endoscopy, and transabdominal ultrasound (Fig. 2).
MATERIALS AND METHODS
All patients with histologically proven CD and known or suspected symptomatic intestinal stricture were considered for participation in this prospective study. Patients were recruited from a single tertiary referral center (Department of Medicine B, Muenster University Hospital, Germany). Prior to study inclusion all patients gave informed consent. The protocol was accepted by the local Ethics Committee of the University of Muenster.
Exclusion criteria were age less than 18 years, pregnancy, and inability to confirm the use of valid contraceptive measures in females as well as general contraindications against the study examinations. Furthermore, patients with intestinal strictures requiring immediate surgery, e.g., in case of an ileus, were excluded. Patients with anal strictures were also excluded from the study due to the natural limitations of the imaging methods to be used.
Details concerning initial diagnosis, previous surgery, and current and previous medical history were documented. The Crohn's Disease Activity Index (CDAI) was calculated at the time of study inclusion.17 All patients underwent endoscopic evaluation of the stricture and biopsy samples were taken from the stricture. According to a standardized endoscopic and histologic protocol the strictures were classified into inflammatory, mixed, and fibrostenotic strictures (see below). Within 1 week all patients underwent endoscopic evaluation, transabdominal bowel ultrasound, MR-enteroclysis, and 18FDG-PET/CT. The investigators were aware of the diagnosis, the history of previous bowel resections, and the current clinical symptoms, but were blinded with respect to the stricture location and the results of all other investigations.
In all patients endoscopic examination of the stricture in combination with biopsy sampling was performed by one of two experienced endoscopists (H.U. or C.M.). In cases with known stricture the endoscopic approach of the most likely way to reach the stricture was chosen, for which standard video endoscopes were used (GIF 160 gastroscope and CF 160 colonoscope, Olympus, Hamburg, Germany; EN-450P5/20 or EN-450T5 double-balloon enteroscope, Fujinon Europe, Willich, Germany). A routine endoscopic evaluation of the entire bowel was not performed. Analgosedation was applied by intravenous administration of propofol alone or in combination with pethidine at a dosage needed for sedation. Patients were monitored by continuous recording of oxygen saturation and pulse rate as well as intermittent measurement of blood pressure.
The stricture was classified as inflammatory, mixed, or fibromatous according to a standardized scoring system considering five different stricture aspects (Table 1), which we created for this study. Each aspect was graduated into subcategories (predominantly fibromatous aspect = 1 point up to predominantly inflammatory aspects = 3 points). The rigidity was differentiated into rigid strictures (strictures were fixed = 1 point), mixed strictures (stricture were between fixed and soft = 2 points), and soft strictures (3 points). The endoscopist evaluated the rigidity during biopsy sampling from the stricture. The color of the stricture was differentiated into pale (1 point), normal (2 points), and reddened strictures (3 points). The mucosal aspect was divided into atrophic mucosa (1 point), moderately edematous mucosa (2 points), and extremely swelling mucosa (3 points). In terms of bleeding activity, mucosal friability was differentiated into not friable mucosa (no bleeding; 1 point), moderately friable mucosa (e.g., bleeding following contact with the endoscope; 2 points), and severely friable mucosa (e.g., spontaneous bleeding; 3 points). Furthermore, the absence of aphthae/ulcers (1 point), presence of aphthae (2 points), and ulcers (3 points) were considered. Accordingly, a stricture with a score from 5 up to 8 points was defined as fibromatous, a score greater than 8 points up to 12 points as mixed, and a score greater than 12 points as inflammatory. Following the endoscopic evaluation, biopsy samples from the stricture were taken for histological examination.
Bowel ultrasonography was performed by one of two experienced ultrasound examiners to detect and classify the strictures (T.K. or F.L.). The investigators were blinded to clinical details, results of previous investigations, and the final diagnosis. Ultrasonography was done after a fasting period of at least 6 hours with a 3.5–5 MHz convexed and/or a 7 MHz linear transducer depending on the resolution in each individual patient (Hitachi EUB 5000). A stricture was defined as a wall thickening greater than 4 mm and a fixed constriction of the intestinal lumen or a prestenotic dilation. For classification of inflammatory activity the vascular pattern and the echogenicity of the stricture were evaluated. Echogenicity was classified into hyper-, hypoechogenic, and mixed pattern. Strictures were examined by color ultrasound Doppler sonography and classified into Limberg indices 1–4 according to a score system defined by Limberg.18 A fibromatous stricture was defined as a hyperechogenic wall thickening and a Limberg index 1. A mixed stricture was classified as a mixed hypo- and hyperechogenic wall thickening and a Limberg index 2. An inflammatory stricture was defined as a hypoechogenic wall thickening with Limberg indices 3 or 4. All other findings were classified as mixed strictures, e.g., a hypoechogenic stricture with Limberg index 1.
After an overnight fast, a nasojejunal Teflon tube (length 1500 mm, outer/inner diameter 3.1/2.1 mm, Servoprax, Wesel, Germany) was positioned distal to the duodenojejunal junction with fluoroscopic guidance or endoscopically placed. During MRI, a minimum of 2000 mL 0.5% methylcellulose solution in water was infused through the tube at a flow rate of 120 mL/min with an infusion pump system (KMP 2000, Fresenius, Bad Homburg, Germany). We performed 1.5-T MRI (Gyroscan Intera, Philips, Best, the Netherlands) with use of a phased-array surface coil during infusion. The imaging protocol included axial and coronal breath-hold T1-weighted (FLASH 3D with fat saturation after intravenous application of 1 mmol/kg body weight gadolinium, TE user-defined, TR shortest, flip angle 50°, matrix reconstruction 512, slice thickness 6 mm, intersection gap 1 mm) and T2-weighted image acquisition (TSE, TSE factor 24, TE 70 ms, flip angle 90°, matrix reconstruction 512, slice thickness 6 mm, intersection gap 1 mm). Images were acquired with patients in the supine position with full abdominal coverage. The total imaging time per patient was ≈20 minutes.
A board-certified radiologist with more than 11 years experience in abdominal radiology reviewed the MR images (J.W.). The reviewer was blinded to clinical details, results of previous investigations, and the final diagnosis. The following MR-enteroclysis features were examined in general: mucosal or mural enhancement, intramural edema, transmural ulcers, mural thickening (> 3 mm), prestenotic dilation, extramural findings (comb sign, enlarged mesenteric lymph nodes, fistula, abscess). With regard to stenoses, fibrotic stenoses were assumed in case of missing hyperenhancement, lack of mural edema, lack of enlarged, and adjacent lymph nodes. An inflammatory stenosis was assumed in bowel segments with mural edema, mural hyperenhancement, and/or mural stratification, comb sign, enhancing and enlarged mesenteric lymph nodes (larger than 1 cm in diameter in their shortest axis), fistula or abscess adjacent to the stenotic bowel segment. A mixed type stricture was diagnosed when inflammatory as well as fibromatous criteria were found within one stricture.
A dual modality PET/CT scanner (Biograph Sensation 16; Siemens Forchheim, Germany, and Hoffmann Estates, IL) was used for data acquisition providing images that could be viewed separately or in fused mode combining morphological and functional image data. After a fasting period of at least 8 hours, a body mass-adapted 18F-FDG activity of 4 MBq/kg was injected intravenously. Images were acquired 1 hour later. Injected 18F-FDG activities and x-ray radiation doses of the CT scanner were reduced to the necessary minimum. Before 18F-FDG-PET/CT imaging acquisition, 0.5% methylcellulose solution in water was also infused as described for MR-enteroclysis if the examination was performed on another day than MR-enteroclysis. Otherwise, 18F-FDG-PET/CT was performed following MR-enteroclysis. All PET acquisitions were combined with low-dose CT scans. The combination of 18FDG-PET with low-dose CT had a maximum total effective radiation dose of up to 8 mSv (18FDG-PET below 7 mSv, low-dose CT = 1 mSv).
In addition to visual analysis including all extraintestinal pathologic findings, all accumulations of 18FDG were semiquantitatively classified. Metabolic activity in all lesions detected was compared on the basis of SUVs (standardized uptake values) to mean liver activity and grouped into four levels (lower than liver, equal, slightly higher, and considerably increased) interpretable as levels of inflammation (no inflammation, low grade, moderate, and high grade).
Strictures with an SUV equal to or lower than the liver were classified as fibromatous. Strictures with an SUV slightly higher than the liver were classified as mixed. All strictures with a considerably increased SUV, i.e., two or more times higher than the mean liver SUV, were classified as inflammatory.
Furthermore, in a second analysis all strictures evaluated by 18FDG-PET/CT were classified employing a quantitative score. The score was calculated as maximum SUV of the stricture divided by the median SUV of the liver to objectively quantify the metabolic activity within the stricture (SUVmaxStricture/SUVmedLiver).
All 18FDG-PET/CT were analyzed by an experienced nuclear medicine specialist (J.B.). The reviewer was blinded to clinical details, results of previous investigations, and the final diagnosis. 18FDG-PET/CT is a new imaging technique for patients with IBD, with limited experience. Therefore, consensus reading was performed by a second experienced nuclear medicine specialist (M.W.). With consensus reading, one additional stricture, one additional entero-enteral fistula, as well as three additional inflammatory manifestations were detected. All other results were confirmed by the consensus reader.
Biopsy samples were evaluated by an experienced pathologist (T.S.) blinded to the results of the imaging studies. Paraffin tissue sections were evaluated with respect to the degree of inflammation as well as the degree of fibrosis with hematoxylin and eosin stains. The degree of inflammatory infiltration was graded into absent (0), mild to moderate (1), and severe inflammation (3) and the degree of fibrosis was graded into absent (0), mild to moderate (1), and severe fibrosis (3). Combining these two criteria, stricture biopsy samples were graded as inflammatory, mixed, or fibromatous.
Gold Standard Definition
The gold standard of the stricture was defined as a combination of endoscopic evaluation and histologic classification of biopsy samples (Table 2). Results of 18FDG-PET/CT, MR-enteroclysis and bowel ultrasound were compared with the gold standard.
|Definition of Gold Standard||Endoscopy|
Following diagnostic stricture evaluation, therapy of patients was determined according to the results of examinations and the patients' history based on the treating physician's individual decision and not according to the study protocol. Follow-up of patients was at least 6 months.
Standard procedures were applied to calculate sensitivity. For estimation of correlation, Spearman's rank correlation coefficient was determined and significance of PET scores was calculated using analysis of variance (ANOVA). The analysis was performed with Sigmastat Software v. 3.5 (Systat Software, Chicago, IL). Statistical significance was defined as P < 0.05. A 95% confidence interval (CI) was used.
Patients and Endoscopy/Histology
Thirty patients with symptomatic CD stricture were enrolled (17 females, 13 males; median age 35 years; range 22–63 years). In these 30 patients a total of 37 strictures were endoscopically detected and grouped relating to the gold standard into inflamed, mixed, or fibrotic lesions. All strictures were further evaluated by 18FDG-PET/CT, MR-enteroclysis and transabdominal ultrasound. The median time from first diagnosis of CD to study inclusion was 13.5 years (range 0–37 years). The median CDAI was 204 (range 89–555). Nineteen strictures were located in the ileocecal (51%), eight in the colonic region (22%), six within the small bowel other than ileocecal (16%), and four in the gastroduodenal region (11%) (Table 3).
|Location||All||Pyloric/Duodenal||Small Bowel*||Ileocecal/Term. Ileum||Colonic|
|Total (not passable)||37 (24)||4 (1)||6 (4)||19 (14)||8 (5)|
|18FDG-PET-CT detected||30/37 (81%)||1/4 (25%)||5/6 (84%)||17/19 (89%)||7/8 (88%)|
|MR detected||30/37 (81%)||1/4 (25%)||6/6 (100%)||16/19 (84%)||7/8 (88%)|
|US detected||25/37 (68%)||3/4 (75%)||3/6 (50%)||13/19 (68%)||6/8 (75%)|
As shown in Table 3, MR-enteroclysis detected 30 out of 37 strictures (81% CI 64%–92%), 18FDG-PET/CT 30 (81%, CI 64%–92%), and transabdominal ultrasound 25 strictures (68% CI 53%–84%). Sensitivity of stricture detection increased when two diagnostic procedures were combined: detection rate with 18FDG-PET/CT plus ultrasound was 92%, with MR-enteroclysis plus ultrasound 89%, and with MR-enteroclysis plus 18FDG-PET/CT 89%.
According to the gold standard, strictures were grouped as inflammatory (n = 22, 60%), mixed (n = 12, 32%), and fibromatous lesions (n = 3, 8%). MR-enteroclysis correctly differentiated 17 out of 30 detected strictures into the inflammatory, mixed, and fibromatous type (57%). 18FDG-PET/CT categorized 16 of 30 detected strictures (53%). With transabdominal ultrasound, 10 of 25 detected strictures were allocated correctly (40%). MR-enteroclysis and 18FDG-PET/CT differentiated 94% and 83%, respectively, of the inflammatory strictures correctly, while transabdominal ultrasound diagnosed 33% of these strictures properly. None of the mixed and fibromatous strictures detected were diagnosed correctly by MR-enteroclysis. 18FDG-PET/CT categorized one of the mixed and fibromatous strictures correctly. Notably, all three fibromatous strictures were correctly assessed by transabdominal ultrasound (Table 4).
|Detected||30/37 (81%)||18/22 (82%)||9/12 (75%)||3/3 (100%)|
|Correct diagnosis||16/30 (53%)||15/18 (83%)||1/9 (11%)||0/3 (0%)|
|Diagnosis inflammatory||15/18 (83%)||8/9 (89%)||2/3 (67%)|
|Diagnosis mixed||3/18 (17%)||1/9 (11%)||1/3 (33%)|
|Diagnosis fibromatous||0/18 (0%)||0/0 (0%)||0/3 (0%)|
|Detected||30/37 (81%)||18/22 (82%)||9/12 (75%)||3/3 (100%)|
|Correct diagnosis||17/30 (57%)||17/18 (94%)||0/9 (0%)||0/3 (0%)|
|Diagnosis inflammatory||17/18 (94%)||7/9 (78%)||3/3 (100%)|
|Diagnosis mixed||0/18 (0%)||0/9 (0%)||0/3 (0%)|
|Diagnosis fibromatous||1/18 (6%)||2/9 (22%)||0/3 (0%)|
|Detected||25/37 (68%)||15/22 (68%)||7/12 (58%)||3/3 (100%)|
|Correct diagnosis||10/25 (40%)||5/15 (33%)||2/7 (29%)||3/3 (100%)|
|Diagnosis inflammatory||5/15 (33%)||1/7 (14%)||0/3 (0%)|
|Diagnosis mixed||7/15 (47%)||2/7 (29%)||0/3 (0%)|
|Diagnosis fibromatous||3/15 (20%)||4/7 (57%)||3/3 (100%)|
Quantitative evaluation of strictures by 18FDG-PET/CT using the SUVmaxStricture/SUVmedLiver score resulted in median values of 3.8 for inflammatory strictures (range 1.7–10.6), 2.7 for mixed strictures (range 1.8–5.4), and 2.2 for fibromatous strictures (range 2.1–4.3). These differences, however, were not statistically significant (P = 0.23).
MR-enteroclysis described five additional strictures that were not evaluated endoscopically; two of those were not detected by 18FDG-PET/CT. Furthermore, three additional inflammatory manifestations of CD and one enteroenteric fistula were diagnosed. 18FDG-PET/CT also described five additional strictures, two of which were not detected by MR-enteroclysis. As these potentially additional strictures had been located beyond endoscopically not passable strictures, they could not be confirmed and assessed by endoscopy. The enteroenteric fistula seen by MR-enteroclysis was also diagnosed with 18FDG-PET/CT.
Ten additional inflammatory manifestations were diagnosed by 18FDG-PET/CT (gastric ulcer [n = 1] / gastric aphthosis [n = 1] / enhanced intestinal activity [n = 8]). Extraabdominal findings of 18FDG-PET/CT were of different clinical significance (thyroid adenoma [n = 1; Fig. 3], Hashimoto thyroiditis [n = 1], pneumonia [n = 1], steroid acne [n = 1], nonspecific pulmonary nodules [n = 10]).
All patients were followed for 6 months. Therapeutic decisions were based on the individual option of the referring clinician and were not defined by the study protocol. Fifty-seven percent of the strictures were treated surgically or endoscopically within 6 months (21/37 strictures) (Table 5). The median time from study inclusion to surgery/dilation therapy was 19 days (range 1–140 days). 18FDG-PET/CT and MR-enteroclysis detected all but one (duodenal) strictures which required surgery or endoscopic dilation therapy (95%). Transabdominal ultrasound did not reveal five strictures that required surgery or endoscopic dilation therapy (24%, ileocolonic n = 3, small bowel stricture n = 2). However, when combining transabdominal ultrasound with 18FDG-PET/CT or MR-enteroclysis, all strictures requiring surgery or endoscopic dilation therapy were detected.
|Strictures||Total||Surgery*||Endoscopic Dilation||Medically Refractory#||(Sub-)Ileus#|
|Total||21/37 (57%)||16/21 (76%)*||5/21 (24%)||18/21 (86%)||3/21 (14%)|
|Inflammatory||12/22 (55%)||10/16 (63%)*||2/5 (40%)||10/12 (83%)||2/12 (17%)|
|Mixed||6/12 (50%)||3/16 (19%)||3/5 (60%)||5/6 (83%)||1/6 (17%)|
|Fibromatous||3/3 (100%)||3/16 (19%)||0/5 (0%)||3/3 (100%)||0/3 (0%)|
As the European Crohn's and Colitis Organization (ECCO) states in their consensus on the treatment of CD, “Differentiation between inflammatory and fibrostenotic bowel stenosis would be very helpful, but current techniques do not permit an accurate distinction,”19 the aim of our study was to compare current standard imaging methods with 18FDG-PET/CT regarding the ability to differentiate between inflamed and fibrotic strictures.
18FDG-PET/CT is widely used in oncology for tumor diagnostics and staging.7, 8, 20 18FDG-PET/CT combines metabolic information by visualizing glucose metabolism with the anatomic data of CT scanning. So far, only a couple of studies have looked at 18FDG-PET/CT for assessing generally inflammatory activity in IBD. For example, a pilot study of 12 patients with IBD including seven patients with CD by Meisner et al12 showed a high correlation of PET activity with endoscopic activity determined by colonoscopy. A second study by Louis et al13 looking at 22 patients with CD showed an overall sensitivity of 18FDG-PET/CT for endoscopically detected lesions of 73%, while the sensitivity for severe lesions, defined as deep ulcers and strictures, was 100%. In our study specifically investigating CD strictures, we also found a high sensitivity regarding the detection of inflammatory activity within the strictures (82% of detected inflammatory strictures were categorized correctly by 18FDG-PET/CT). When comparing our results with the studies by Louis et al and Meisner et al, one has to take into consideration that the latter two exclusively looked at inflammation in the colon/ileocolonic region, while we studied CD strictures in the entire gastrointestinal tract. Although the number in the subgroup is small, 18FDG-PET/CT as well as MR-enteroclysis appear to be less sensitive in detecting strictures in the duodenum/proximal jejunum, as they missed three of four endoscopically detected strictures in this area.
None of the studies quoted above compared the results of 18FDG-PET/CT with other imaging methods, whereas in our study we did so relating to MR-enteroclysis and transabdominal ultrasound. Detection rates of strictures were found not to be significantly different. A recently published retrospective study assessing the diagnostic value of 18FDG-PET combined with CT-enterography in active CD showed that PET scanning added to CT-enterography did not identify additional abnormal segments when compared with CT-enterography alone.16 Results of that study and our present investigation are in accordance with a meta-analysis of 33 prospective studies comparing ultrasound, MR-enteroclysis, leukocyte scintigraphy, CT, and 18FDG-PET (without CT), which did not find significant differences of the diagnostic accuracy in IBD in general (not only strictures) among these imaging techniques.21
Although in our study the detection rate of CD strictures was high, neither 18FDG-PET/CT nor the other established methods showed a reliable sensitivity and specificity with respect to the primary aim of this study, namely, differentiation of inflamed from fibrotic strictures according to our gold standard. One potential reason may be that even in mainly fibrotic strictures some inflammation is present, as has been described in a surgical specimens study by Jacene et al.15 Differences in the degree of inflammation might therefore not be high enough with respect to the sensitivity of the imaging methods investigated. According to this assumption, we found in the PET/CT analysis a tendency of higher SUV values for inflammatory strictures than for mixed and fibromatous lesions. However, these differences between the evaluated strictures were not statistically significant. Furthermore, according to our gold standard only three strictures were categorized as fibromatous. Therefore, conclusions for pure fibrotic strictures are limited.
Another problem might be the gold standard used in our study, consisting of endoscopic inspection and histologic biopsy assessment of the strictures. As endoscopy does not allow a view below the mucosal layer and biopsy samples were often taken from only one side of the stricture in case of endoscopic impassability, a misjudgment of some strictures cannot be entirely excluded. A more accurate gold standard would apparently be the histologic evaluation of resected specimens at the time of study evaluation. However, following CD treatment guidelines surgical resection due to the high recurrence rate within the years after primary surgery should mainly be limited to fibrotic strictures, as inflamed ones often can successfully be treated by intensified medical therapy.2 We, therefore, although knowing the above mentioned potential limitations, thought the gold standard used to be the best option, as by all means we wanted to include highly inflamed strictures as well.
When looking at the detection rate of CD strictures in general, 18FDG-PET/CT in our hands was very sensitive. However, one has to keep in mind that 18FDG-PET/CT was the only method in our study using ionizing energy. A combination of 18FDG-PET with low-dose CT had a maximum total effective radiation dose of up to 8 mSv (18FDG-PET below 7 mSv, low-dose CT = 1 mSv). The total radiation dose is therefore close to a conventional abdominal CT (10 mSv) and less than a barium enema (15 mSv).22 Patients who have complicated CD, such as patients who need repeated surgery due to strictures, and especially young patients, are at risk of exposure to significant amounts of diagnostic radiation energy, with a significant increase in the last decade.23 As CD patients often belong to the group of younger patients, it is important to keep in mind that these patients have a higher lifetime estimated risk of radiation-induced cancer.22 On the other hand, as in our study the combination of MR-enteroclysis/ultrasonography or 18FDG-PET/CT/ultrasonography detected all endoscopically proven CD strictures, one also has to keep in mind potential sedation-associated as well as endoscopy-associated complications. In patients with CD, colonic perforations during colonoscopy are reported in up to 0.8% of procedures.24 In addition, colonoscopy requires bowel preparation, which is especially difficult up to impossible for patients with strictures. As strictures often can be evaluated only from one side by endoscopy due to the inability to pass through the stricture with the scope, one of the two nonendoscopic diagnostic methods might be an alternative option for at least a subgroup of CD patients.
In our study the combination of MR-enteroclysis and ultrasound as well as the combination of 18FDG-PET/CT and ultrasound resulted in a 100% detection rate of CD strictures requiring surgery or endoscopic dilation therapy, suggesting the combination of these methods as an alternative to endoscopy, at least in the group of patients not being able to perform an adequate bowel preparation. Due to its easy application, ultrasound should be preferably used as a first imaging method in experienced centers next to an at least initial endoscopic evaluation. In case of nonconclusive findings, or suspected complications such as fistula or abscess development, MR-enteroclysis or CT should be additionally performed. According to a recent meta-analysis, the sensitivity and specificity rates for the latter two methods are comparable, although as suggested by Panes et al,25 MR-enteroclysis should be preferred due to its lack of radiation. For follow-up analysis, particularly regarding the treatment response, ultrasound is probably the most feasible and cost-effective method. 18FDG-PET/CT having similar detection rates as an MR-enteroclysis might serve as an alternative method. By means of this algorithm, CD strictures will very likely be detected. Nevertheless, even when using 18FDG-PET/CT the above-cited ECCO statement still holds true that current techniques do not yet permit an accurate distinction between inflamed and fibrotic strictures.
The authors thank K. Menzel, S. Huesmann, and A. Mähler for excellent technical assistance. Guarantors of the article: C. Maaser and F. Lenze.
- 10Pet-Ct — a new diagnostic tool for differentiation between inflammatory and fibromatous stenosis in Crohn's disease. Gastroenterology. 2005; 128: A–310., , , et al.
- 25Systematic review: the use of ultrasonography, computed tomography and magnetic resonance imaging for the diagnosis, assessment of activity and abdominal complications of Crohn's disease. Aliment Pharmacol Ther. 2011; 34: 125–145., , , et al.