Capsule endoscopy is cool. Like any technology, it has certain advantages and some shortcomings. Understanding that this technology has limited sensitivity and specificity is an important matter in clinical practice.
The suboptimal sensitivity is apparent in a recent study (Am J Gastroenterology, (10 January 2012) | doi:10.1038/ajg.2011.461). In this study titled, “Detection of Lesions During Capsule Endoscopy: Physician Performance Is Disappointing,” the authors compare the performance of 17 capsule endoscopists in reading clips from 24 capsule procedures. 18 of these clips were abnormal but only abnormal in <0.1% of frames. They compared different reading modes, speeds and the experience of the physicians. By using clips, they were able to truncate the sessions to a similar length as an individual capsule procedure.
Their findings showed detection rates of 69% for angioectasias, 38% for ulcers/erosions, 46% for masses/polyps, and 17% for blood. For all categories the detection rate was ~45% for SingleView15, QuadView20, and QuadView30; however, SingleView25 had a 25% detection rate. One surprising finding was how little experience mattered; there was no significant advantage in having more experience. In fact, the best performer who correctly identified 78% of the abnormalities had read only 45 capsule studies previously; the worst performer with a 17% score had read more than 11,000 previously.
In a typical capsule study, 40,000-60,000 images are generated. If there is not a diffuse disease, then the lesion may go undetected. The findings of these studies, though limited by the small number of interpreting physicians, confirm problems with sensitivity from other studies. One aspect of this study that deserves emphasis is that speed matters. As with colonoscopy, additional time reviewing the study does increase the likelihood of identifying abnormalities. Ultimately, technological innovation may improve the results with capsule endoscopy. With newer versions of the software, “suspected blood indicator” (SBI) quickly identified six of the 18 lesions and “QuickView” identified 11 of the 18 lesions. As with any test, though, if the clinical suspicion is high, having another physician review the study or using an alternative imaging tool may be necessary.
Another problem to keep in mind with capsule endoscopy (or wireless capsule endoscopy -WCE) is that especially in adults there is a significant background abnormality rate that may be clinically unrelated (Clin Gastro & Hepatol 2005; 3: 133-141. n=413. 13.8% with baseline mucosal injury).
- -IBD 2008; 14: 1219. Correlation of capsule & MRI enteroclysis. Capsule detected more mucosal lesions.
- -JPGN 2008; 47: 31. Capsule endoscopy may reclassify pediatric IBD.
- -Dig Liver Dis 2008; 40: 216-223. Study comparing capsule vs EGD in detecting varices.
- -JPGN 2008; 46:4. Review.
- -Gastroenterology 2007; 132: 855. Capsule outperforms push enteroscopy.
- -IBD 2007; 13: 331. Role of WCE in IBD.
- -Pediatrics 2006; 118: e904. Use of capsule endoscopy for HSP
- -Clin Gastro & Hep 2005; 3: 959. WCE should be 2nd step in obscure bleed, p egd/col.
- -Clin Gastro & Hep 2005; 3: 772. WCE outperforms enteroclysis in identifying Crohn’s sm bwl dz.
- -Gastroenterology 2005; 128: 1172. NSAID injury witnessed in 68% of healthy volunteers p 2 week course; 75% c increased fecal calprotectin.
- -Clin Gastro & Hepatol 2005; 3: 411. Practical applications & review.
- -Clin Gastro & Hepatol 2005; 3: 264-70. Study of 30 children; helpful in diagnosing obscure bleeding.
- -Endoscopy 2005; 37: 960-65. Results of capsule endoscopy trials
- -Clin Gastro & Hepatol 2005; 3: 133-141. WCE to assess NSAID injury, n=413. 13.8% with baseline mucosal injury. Celebrex with fewer lesions than Naproxen plus omeprazole.
- -Gortzak Y, Lantsberg L, Odes HS. Video capsule entrapped in a Meckel’s diverticulum. J Clin Gastro 2003; 37: 270-271.
- -Clin Gastro & Hep 2005; 3: 55. visible lesions c NSAIDs
- -IBD 2004;10: 278-285. WCE for Crohn’s (review) Capsule can help differentiate UC from Crohn’s.
- -Clin Gastro & Hepatology 2004; 2:xx. Dx of Peutz-Jeghers w WCE.
- -Gastro 2004; 126: 643-653. WCE performs well for occult bleeding, especially active bleeding (87% detection rate)
- -Clinical Gastro & Hepatology 2004; 2: 14-15, 31-40.
- -Gastroenterology 2003; 124 (suppl 1) A37. abnl small bowel findings common in healthy subjects.
- -JPGN 2003; 37: 332 (18A), n=58. Crohn’s, polyposis, occult bleeding
- -Gastroenterology 2003; 124: suppl 1, A-37 (284) 22% of normal pts c abnl wireless findings
- -Gastroenterology 2002; 123: 999-1005, 1385-88. capsule is superior to radiographs for suspected small bowel dz.
- -Gastro Endosc 2002; 56: 452-456. Algorithm suggests using this modality if negative upper & lower endoscopy & no acute overt bleeding. If capsule endoscopy is negative & patient is stable, observation appropriate.
- -Gastro Endosc 2002; 56: 621-24. ASGE guidelines.