Visibility of early gastric cancers by texture and color enhancement imaging using a high-definition ultrathin transnasal endoscope

Abstract

We evaluated whether texture and color enhancement imaging (TXI) using a high-definition ultrathin transnasal endoscope (UTE) improves the visibility of early gastric cancer (EGC) compared with white-light imaging (WLI). This study included 31 EGCs observed by TXI mode 2 using a high-definition UTE prior to endoscopic submucosal dissection. The first outcome was to compare the color differences based on Commission Internationale de l'Eclairage L*a*b* color space between EGCs and the surrounding mucosa by WLI and TXI using the UTE (objective appearance of EGC). The second outcome was to assess the visibility of EGCs by WLI and TXI using the UTE in an image evaluation test performed on 10 endoscopists (subjective appearance of EGC). Color differences between EGCs and non-neoplastic mucosa were significantly higher in TXI than in WLI in all EGCs (TXI: 16.0 ± 10.1 vs. WLI: 10.2 ± 5.5 [mean ± standard deviation], P < 0.001). Median visibility scores evaluated by 10 endoscopists using TXI were significantly higher than those evaluated using WLI (TXI: 4 [interquartile range, 4-4] vs. WLI: 4 [interquartile range, 3-4], P < 0.001). TXI using high-definition UTE improved both objective and subjective visibility of EGCs compared with WLI.

Figure 1

Study flowchart. EGC Early gastric cancer, ESD Endoscopic submucosal dissection, TXI Texture and color enhancement imaging, WLI White-light imaging.

Figure 2

Examples of actual endoscopic images used in the colorimetric evaluation process. A reddish depressed lesion is seen in the posterior wall of the middle gastric body. A total of 8 ROI were annotated on the surrounding non-neoplastic mucosa and EGC. ROIs were set at the same point on the WLI (A) and TXI (B). The yellow arrows indicate the EGC, and the white arrows indicate the surrounding non-neoplastic mucosa. All ROIs were standardized to 25 pixels (5 × 5).

Figure 3

A representative case showing improved color differences on TXI. (A) WLI displayed a slightly discolored, flat elevated lesion in the anterior wall of the lower gastric body. The demarcation line is slightly difficult to identify. The mean ΔE was 8.5 ± 1.6. (B) TXI displayed contrast enhancement between the lesion and the surrounding mucosa, resulting in an easily recognizable demarcation line. The mean ΔE was 12.7 ± 4.3.

Figure 4

A representative case showing improved visibility on TXI. (A)WLI displayed an isochromatic flat lesion in the anterior wall of the lower gastric body. Although there is a slight difference in surface structure, it is difficult to recognize. The median (IQR) visibility by 10 endoscopists was 2 (1–4). (B) TXI displayed enhancement of the structure of the EGC, and the mucosal atrophy in the background appeared more whitish. The median (IQR) visibility by 10 endoscopists was improved to 3 (2–4). The yellow arrows in (A) and (B) indicate the EGC.