Jejunal gastrointestinal stromal tumor missed by three capsule endoscopies

We read with interest the case report by Jovanovic et al. about the ineffectiveness of capsule endoscopy and total double-balloon enteroscopy (DBE) to elicit the cause of obscure overt gastrointestinal bleeding in a case of gastrointestinal stromal tumor (GIST) [1].

We encountered a similar case of a 65-year-old man with a history of recurrent obscure gastrointestinal bleeding for which no diagnosis had been made, despite two upper and lower endoscopies, abdominal computed tomography (CT), and two capsule endoscopies (PillCam SB2; Given Imaging, Yoqneam, Israel). Repeat capsule endoscopy was unremarkable. A CT enterography was performed and this revealed an irregular enhancement of a 3-cm segment of the proximal jejunum ([Fig. 1]).

Fig. 1 Computed tomography (CT) enterography revealed an irregular contrast enhancement (arrow) of a proximal jejunal segment.

A retrospective review of all capsule endoscopy videos showed no abnormality. DBE (Fujinon, Saitama, Japan) disclosed a submucosal lesion with a central depression, 1.5 m distal to the ligament of Treitz ([Fig. 2]).

Fig. 2 Double-balloon enteroscopy (DBE) disclosed a submucosal lesion of the proximal jejunum, with a central depression.

The patient underwent single-port laparoscopy and the involved segment was resected ([Fig. 3 a, b]).

Fig. 3 a The lesion was easily identified at laparoscopy because of its predominantly extraluminal component. b Intraluminal aspect of the surgical specimen. The lesion protruded slightly at the mucosa, had a central depression (arrow) and was ulcerated (arrowhead). Endoscopic submucosal injection of Indian ink had been done near the lesion to facilitate intraoperative location (asterisks).

Histopathological analysis, including immunochemistry for CD117 (c-kit), confirmed a low risk GIST ([Fig. 4 a, b]).

Fig. 4 a Well-circumscribed homogeneous parietal tumor (T) (hematoxylin and eosin [H&E]; × 20). b The lesion consisted of interlacing bundles of regular spindle cells (H&E × 200). The diagnosis of gastrointestinal stromal tumor (GIST) was confirmed by positive immunostaining (inset at right; CD117; × 100).

Nowadays, capsule endoscopy is considered to be the first-line diagnostic tool for evaluation of the small bowel in patients with obscure gastrointestinal bleeding [2]. However, recently published studies with DBE have clearly demonstrated that capsule endoscopy can miss even large malignant masses [3] [4] [5], with a miss rate of up to 18.9 % [3]. It is well recognized that the duodenum and proximal jejunum are the least well visualized by capsule endoscopy, probably because of rapid capsule transit, bile and/or bubble artifact, and relatively poor luminal distension [4]. In our case the tumor was missed by three capsule endoscopies, probably because of its predominantly extraluminal rather than intraluminal position.

In conclusion, one or more negative capsule endoscopies do not exclude the presence of significant lesions. If there is a high index of suspicion, alternative imaging modalities such as DBE, CT angiography or enterography, and magnetic resonance enterography should be considered.

References

• 1 Jovanovic I, Krivokapic Z, Menkovic N et al. Ineffectiveness of capsule endoscopy and total double-balloon enteroscopy to elicit the cause of obscure overt gastrointestinal bleeding: think GIST!.  Endoscopy. 2011;  43 (Suppl 2) 91-92
• 2 Fisher L, Lee K rinsky, Anderson M A et al. Endoscopy in obscure GI bleeding.  Gastrointest Endosc. 2010;  72 471-479
• 3 Lewis B S, Eisen G M, Friedman S. A pooled analysis to evaluate results of capsule endoscopy trials.  Endoscopy. 2005;  37 960-965
• 4 Postgate A, Despott E, Burling D. Significant small-bowel lesions detected by alternative diagnostic modalities after negative capsule endoscopy.  Gastrointest Endosc. 2008;  68 1209-1214
• 5 Chong A K, Chin B W, Meredith C G. Clinically significant small-bowel pathology identified by double-balloon enteroscopy but missed by capsule endoscopy.  Gastrointest Endosc. 2006;  64 445-449