Published online Mar 14, 2019. doi: 10.3748/wjg.v25.i10.1248
Peer-review started: January 2, 2019
First decision: January 18, 2019
Revised: February 14, 2019
Accepted: February 15, 2019
Article in press: February 16, 2019
Published online: March 14, 2019
Processing time: 74 Days and 12.2 Hours
Linked color imaging (LCI) emphasizes slight differences in the “red” color of the mucosa by image processing, including enhancement of differences in chroma and hue of the red mucosal color.
The utility of LCI for diagnosis of Helicobacter pylori gastritis has been well described, but there are few studies of LCI for diagnosis of gastrointestinal cancer and, to our knowledge, parallel analyses of LCI, white light imaging (WLI), and pathology samples have not been reported.
The objectives of this study are to evaluate the utility of LCI for endoscopic diagnosis of early gastric cancer and to examine if pathological findings can explain the changes in red shade in LCI.
Three endoscopists evaluated lesion recognition with WLI and LCI. Color values in Commission Internationale de l'Eclairage (CIE) 1976 L*a*b* color space were used to calculate the color difference (ΔE) between cancer lesions and non-cancer areas. After endoscopic submucosal dissection, blood vessel density in the surface layer of the gastric epithelium was evaluated pathologically. The identical region of interest was selected for analyses of endoscopic images (WLI and LCI) and pathological analyses.
LCI gave better contrast than WLI for the color difference between cancer lesions and surrounding non-cancer tissue; an a* cutoff ≥ 24 for the value in CIE 1976 L*a*b* color space had good sensitivity and specificity for diagnosis of early gastric cancer; and surface blood vessel density in cancer lesions was significantly higher than that in non-cancer areas.
LCI was useful for recognition of early gastric cancer lesions because this method provides good contrast in color differences between lesions and surrounding tissue. Blood vessel density from the surface to a depth of 350 μm was higher in cancer lesions than in non-cancer areas, and LCI clearly shows this feature as a change in redness.
If these color values are evaluable in real time, artificial intelligence may permit automatic recognition and diagnosis of lesions.
