Improved visibility of lesions of the small intestine via capsule endoscopy with computed virtual chromoendoscopy

Gastrointest Endosc. 2011 Feb;73(2):299-306. doi: 10.1016/j.gie.2010.10.016.

Abstract

Background: We can now enhance video capsule endoscopy (CE) images in real time by means of a flexible spectral imaging color enhancement (FICE) digital processing system. Reports on the clinical usefulness of this system are few.

Objective: To clarify whether visualization of lesions of the small intestine is improved by FICE image analysis.

Design: A retrospective study.

Setting: Academic medical center.

Methods: Five physicians compared FICE images with corresponding conventional images of 145 lesions obtained from 122 patients who underwent video CE at our hospital. The lesions were classified as angioectasia (n=23), erosion/ulceration (n=45), or tumor (n=75), and 3 different sets of FICE images were viewed (ie, at 3 different wavelength settings). Physicians rated the visibility of the lesions on FICE images as follows: +2 (improved visibility), +1 (somewhat improved visibility), 0 (visibility equivalent to that of conventional video CE visibility), -1 (somewhat decreased visibility), and -2 (decreased visibility). Scores for each lesion were totaled (per FICE setting) and evaluated. Intraobserver agreement was also examined.

Results: With FICE setting 1 (red 595 nm, green 540 nm, blue 535 nm), improvement was achieved for 87% (20/23) of angioectasia images, 53.3% (26/47) of erosion/ulceration images, and 25.3% (19/75) of tumor images. With setting 2 (red 420 nm, green, 520 nm, blue 530 nm), improvement was achieved for 87% (20/23), 25.5% (12/47), and 20.0% (15/75), respectively. With setting 3, only equivalence was achieved. Intraobserver agreement was good to satisfactory at 5.4 or higher.

Limitations: Single-center study.

Conclusions: CE-FICE improves visibility of small-bowel angioectasia, erosion/ulceration, and tumor.

Publication types

  • Comparative Study

MeSH terms

  • Capsule Endoscopy / methods*
  • Humans
  • Image Enhancement / methods*
  • Image Processing, Computer-Assisted*
  • Intestinal Diseases / diagnosis*
  • Intestine, Small*
  • ROC Curve
  • Retrospective Studies