Double Balloon Endoscopy in Celiac Disease

Uncomplicated CD 

Several endoscopic signs have been described in CD. Most of the data concerning endoscopical signs date from the pre-DBE era and are limited to duodenum and proximal jejunum findings. Many of these signs can be encountered in the duodenum and are within the reach of a standard gastroscope. Sensitivity of these signs is around 50%, whereas specificity is nearly 100%.19

Flattened villi are the endoscopical depiction of villous atrophy. With the tip of the endoscope close to the mucosa, villi appear short and broadened (Fig. 1). A better view can be appreciated after rapid instillation of water, which allows the villous tips to float.20 If total villous atrophy is present, the mucosa can have a mosaic, or cobblestone-like appearance (Fig. 2). Scalloping of folds can be encountered throughout the small bowel (Fig. 3). The normally smooth bordered folds appear notched.21 This sign is best visualized in a maximally insufflated part of the small bowel. The scalloping appearance is probably the result of mucosal grooves, which can also be encountered between mucosal folds. Visualization of submucosal blood vessels has also been reported as a sign of mucosal atrophy in CD. We do not consider this a specific sign (Fig. 4).22

View Large Image Download to PowerPoint

Figure 1. Total and partial villous atrophy in the jejunum of a 70-year-old male with longstanding CD. Histopathological examination revealed Marsh IIIB, without evidence for RCD II or EATL.

View Large Image Download to PowerPoint

Figure 2. Cobblestone-like appearance of the jejunal mucosa in the same patient as in Figure 1. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 3. Scalloping of jejunal folds. Same patient as in Figure 1, Figure 2. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 4. Visualization of submucosal blood vessels in the jejunum of a 34-year-old male with iron deficiency anemia without a known cause. Pathological examination revealed no abnormalities. (Color version of figure is available at www.techgiendoscopy.com.)

A reduction or loss of folds in the jejunum is a typical sign in radiology of the small bowel, especially when combined with an increased number of ileal folds.23, 24, 25, 26, 27, 28, 29 In endoscopy, the reduction or loss of jejunal folds is defined as less than four folds in an endoscopic view of maximally insufflated jejunum.26 There are no data on increased ileal folds as an endoscopic sign of celiac disease. The decreased number of folds, combined with the disappearance of villi and mucosal atrophy can give the jejunal mucosa the aspect of colonic mucosa.

Dickey and coworkers described duodenal erosions in five patients with CD.27 In at least one patient, the erosions did not disappear after GFD, which could suggest RCD (Fig. 5). In our opinion, the presence of erosions should raise the suspicion of ulcerative jejunitis.29

View Large Image Download to PowerPoint

Figure 5. Solitary erosion in the proximal jejunum of a 64-year-old male patient, 2 years after the diagnosis of CD had been established. Pathology showed Marsh IIIB. T-cell flowcytometry showed 65% aberrant T-cells, consistent with RCD II. A more distal lesion proved to be EATL. (Color version of figure is available at www.techgiendoscopy.com.)

Once familiar with the variances in appearance of the above-mentioned signs, they can be appreciated in the small bowel of a great part of CD patients. Lack of these signs in the proximal small bowel of a CD patient in whom DBE has been indicated should, however, never give the false impression that no gross abnormalities can be encountered more distal.

RCD 

Endoscopy alone cannot distinguish between uncomplicated CD and RCD. Obtaining biopsy specimens is therefore of major importance in suspected RCD, not only to confirm RCD and exclude other forms of enteropathy, but also, if the diagnosis CD stands, to determine the type of RCD.

In a recent study, Hadithi and coworkers found loss of jejunal folds and scalloping of jejunal folds in all 14 patients with RCD.28 Nodularity was present in 9 patients (64%) and visible submucosal vessels in 2 patients (14%). Jejunal erosions were encountered in 1 patient (7%). These signs could not discriminate between RCD I and RCD II. In 6 patients, small bowel histology improved in the middle or distal part of the small bowel. In 7 patients, flow-cytometry showed an increased population of aberrant T-cells, suggestive of RCD II. No ulceration (predefined as mucosal defects of at least 5 mm in diameter) or stenosis was found in this group.28 Recently, DBE-assisted chromoendoscopy using indigo carmin was used to characterize the mucosal lesions in a patient with RCD.30

Ulcerative Jejunitis and EATL 

Although histological features of ulcerative jejunitis are often nonspecific, there is immunopathological and molecular evidence that most of such cases represent early T-cell lymphomas.11 Ulcerations of various sizes can be encountered isolated, clustered, or diffusely throughout the jejunum. They can appear as small, white, usually only slightly depressed, circular, or confluent lesions, or as larger, deeper ulcers (Figure 6, Figure 7). In our opinion, these findings indicate additional imaging of the part of the small bowel that has not been reached during antegrade DBE.

View Large Image Download to PowerPoint

Figure 6. Deep ulceration in the presence of villous atrophy in the jejunum of a 55-year-old female with longstanding CD. Histopathology of this lesion showed Marsh IIIC, with over 90% aberrant T-cells.

View Large Image Download to PowerPoint

Figure 7. Multiple ulcerations in the jejunum with edematous folds. Same patient as in Figure 6. Biopsy specimens from this site were consistent with RCD II, without the presence of lymphoma. (Color version of figure is available at www.techgiendoscopy.com.)

In our experience, isolated or multiple stenoses most often are encountered in patients with EATL. Biopsy specimens are, however, frequently nonspecific. Stenoses could reflect severe, circumferential ulceration. Stenoses can be short, resembling webs, or up to 1 or 2 centimeters long (Figure 8, Figure 9). Although no reports of perforation of stenosis during DBE have been published, we usually try to avoid balloon insufflation near such strictures. However, dilation by the balloon of either the endoscope or the overtube may occur unintentionally (Fig. 10). Only once has balloon dilation of a CD-related stricture been performed in our center, to pass the stricture with the endoscope and obtain biopsy specimens of a distally located suspicious mass, suspected to be an EATL-localization. Severe stenosis often is associated with food impaction and prestenotic dilation (Fig. 11).

View Large Image Download to PowerPoint

Figure 8. Stenosis in the presence of total villous atrophy in a 66-year-old male patient in whom CD was diagnosed 1 year before. Pathological examination showed a monoclonal T-cell population, consistent with EATL. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 9. Short, web-like stenosis with ulceration in the jejunum of the same patients as in Figure 6. Histopathology showed no localization of EATL at this site. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 10. Unintentional dilation of a web-like stricture due to the balloons of the endoscope or overtube in a 65-year-old female with a 40-year history of celiac disease. Pathological examination showed Marsh IIIA, without aberrant T-cells, consistent with RCD I. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 11. Ulcerative pinpoint stenosis with food impaction in the jejunum of the same patient as in Figure 6. Histological examination showed the localization of EATL. (Color version of figure is available at www.techgiendoscopy.com.)

Endoscopically, EATLs can present as small circular or confluent ulcers, or as large ulcerative protruding masses, with an irregular and vulnerable surface, often with necrosis or stenosis (Figure 12, Figure 13, Figure 14). Villi are usually absent. In the study by Hadithi and coworkers, five patients were diagnosed with EATL and two with ulcerative jejunitis.28 Ulceration was found in all patients with EATL, whereas in only one patient stenosis was encountered. In two patients with ulcerative lesions, no EATL could be found, despite the many biopsy specimens that had been taken. None of the findings in patients with EATL or ulcerative jejunitis could have been found with conventional EGD.

View Large Image Download to PowerPoint

Figure 12. Irregular, ulcerative mass preceding a stenosis of the distal jejunum. Same patient as in Figure 6. Histological examination showed the localization of EATL. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 13. Isolated oval lesion, with slightly elevated irregular borders and a central, ulcerative indentation in the jejunum of the same patient as in Fig. 9. Pathological examination showed this lesion to be EATL. (Color version of figure is available at www.techgiendoscopy.com.)

View Large Image Download to PowerPoint

Figure 14. Irregular circular ulcerative lesions, with loss of folds and prestenotic dilation the jejunum of the same patient as in Figure 7. Pathological examination showed these lesions to be EATL. (Color version of figure is available at www.techgiendoscopy.com.)

Small Bowel Adenocarcinoma 

Small bowel adenocarcinoma of the small bowel is histologically identical to colorectal adenocarcinoma. Endoscopically, they often appear as irregular, ulcerative lesions, often leading to stenosis and prestenotic dilation (Fig. 15). It is not certain whether the adenoma–carcinoma sequence observed in colorectal cancer also occurs in the small bowel. The growth pattern of adenocarcinoma of the small bowel usually tends to be more exophytic than in lymphoma. However, no certain distinction can be made without extensive sampling for histopathological examination.

View Large Image Download to PowerPoint

Figure 15. Primary adenocarcinoma of the jejunum. Note the irregular, tumorous, ulcerative lesion which can often leading to stenosis and prestenotic dilation (photo courtesy of Lucia C. Fry, MD, University of Magdeburg, Germany). (Color version of figure is available at www.techgiendoscopy.com.)

Biopsy 

To allow an accurate histopathological evaluation, multiple biopsy specimens should be obtained from multiple locations. Using spiked biopsy forceps, we usually take eight biopsies per site, so enough material for flow cytometric analysis is collected. It is important not only to take biopsy specimens from ulcerative, stenotic, or tumorous tissue, but also from relatively normal-looking mucosa, since on many occasions biopsies from evident lesions contain too many cell debris or granulomatous tissue to allow for flow cytometry. In the presence of EATL, intraepithelial lymphocytes in relatively normal mucosa often show an abnormal immunophenotype identical to that of the lymphoma.29

Tattooing 

Although EATL is a condition requiring systemic treatment, surgical debulking of large tumor localizations can be indicated to prevent small bowel perforation during chemotherapy. Since debulking is only necessary for large tumors, which can usually be palpated by the surgeon without the need for an explorative enterotomy to decide on the resection margins, tattooing these lesions is often not indicated. However, if there are multiple tumors, or if strictures proximal or distal to a solitary EATL are present, tattooing of the most distal and most proximal EATL location and/or stricture allows the surgeon to choose wider resection margins which will include all relevant pathological lesions.

VCE 

VCE is a minimally invasive method to evaluate the small bowel. Its use in complicated disease, however, is limited due to the inability to obtain biopsy specimens.30, 31 Although ulcerative jejunitis and EATL can be detected with VCE, histological confirmation is mandatory. In a study on the yield of VCE in RCD by Daum and coworkers, total small bowel examination was achieved in only 9 of 14 patients. VCE detected one case of EATL, which was confirmed using DBE.31 The risk of capsule retention due to strictures should be considered before performing VCE in patients with suspected complicated CD.

Computed Tomography Enteroclysis 

A study from Mallant and coworkers shows computed tomography enteroclysis (CTE) a useful tool in discriminating between uncomplicated CD/RCDI and RCDII/EATL.32 RCD II and EATL showed more bowel wall thickening, lymphadenopathy, and intussusception, less increase in number of small mesenteric vessels, and a smaller splenic volume compared with CD and RCD I. Additionally, the presence of hyposplenism, which is more often encountered in RCDII/EATL, can be evaluated.

Magnetic Resonance Enteroclysis 

Magnetic resonance enteroclysis (MRE) is a promising tool to evaluate the small bowel in patients with suspected RCD or EATL. Not only does MRE allow visualization of jejunal and ileal fold abnormalities, stenosis, dilation, and intraluminal masses, it also informs on bowel wall thickness, mesenterial fat infiltration, lymph node enlargement, and splenic volume. Recent research from our center shows MRE to have a sensitivity of 83% and a specificity of 94% in diagnosing RCD II or EATL.33 In our center, MRE is used as a screening tool to determine which patients should undergo DBE, and as the first modality of choice to visualize the distal small bowel that has not been reached during DBE because of non-passable strictures.

Conventional Enteroclysis 

In our opinion CTE and MRE, with their ability to evaluate not only intraluminal but also extraluminal abnormalities, have replaced conventional enteroclysis. Therefore, there is no place for conventional small bowel radiology in the diagnosis or management of CD.