Gastric cancer: Prevention, screening and early diagnosis

Table 1 Narrow-band imaging endoscopy for early gastric cancer/dysplasia diagnosis

Ref.	Endoscopic technique	Mucosal and vascular pattern for GC	Accuracy
Kaise et al[151]	NBI-ME for superficial depressed gastric lesions vs WLE	The triad: Absence of fine mucosal structure with microvascular dilation and heterogeneity	NBI-ME specificity (85%, theoretically calculated if all of the triad were positive), which was significantly (P < 0.001) superior to WLE general diagnosis (65%)
Kato et al[152]	NBI-ME vs WLE	The triad: Absence of fine mucosal structure with microvascular dilation and heterogeneity	NBI-ME sensitivity (93%) and specificity (95%)
Ezoe et al[153]	NBI-ME vs WLE	Irregular V pattern with a mucosal DL	NBI the diagnostic accuracy was significantly higher for than for WLI (79% vs 44%; P = 0.0001), as was its sensitivity (70% vs 33%; P = 0.0005). The diagnostic specificity of NBI (89%) was higher than that of WLI (67%), but the difference was not statistically significant
Capelle et al[154]	NBI without ME	Complete loss of architectural and mucosal pattern	The sensitivity, specificity, PPV and NPV for detection of premalignant lesions were 71%, 58%, 65% and 65% for NBI and 51%, 67%, 62% and 55% for WLE, respectively
Maki et al[155]	NBI-ME vs WLE to differentiate between cancer and adenoma in superficial elevated lesions of the stomach	WLE: Red coloring NBI-ME: An irregular V pattern with a DL, or irregular S pattern with a DL	The sensitivity, specificity, and accuracy of WLE vs NBI-ME were 64% (52%-76%) vs 95% (90%-100%), 94% (86%-100%) vs 88% (77%-99 %), and 74% (66%-83%) vs 92% (86%-98%), respectively
Tsuji et al[156]	NBI-ME	VS classification: (1) irregular V pattern with a DL between the lesion and the surrounding area; and (2) irregular S pattern with a DL between the lesion and the surrounding area	Sensitivity and specificity for carcinoma were 75.0% and 84.9%, respectively. PPV was 81.4%
Omori et al[157]	NBI-ME	Fine network (net-like appearance consisted of irregular shaped micro vessels), core vascular (clearly visible coiled or wavy vessels in the central area of the mucosal structure), and unclear patterns (micro vascular patterns is not observed)	Sensitivity 86.2%, specificity 97.0%
Wang et al[158]	NBI-ME vs CLE	NBI: “VS” classification system	Accuracy of the CLE and the NBI-ME diagnosis was 88% (95%CI: 78%-98%) and 81% (95%CI: 69%-93%), respectively
Kaise et al[159]	NBI-ME vs WLE	The triad: Disappearance of fine mucosal structure, microvascular dilation, and heterogeneity	The sensitivity and specificity for NBI-ME diagnosis using the triad (92.9% and 94.7%, respectively) were significantly better than those for WLE (42.9% and 61.0%, respectively)
Pimentel-Nunes et al[160]	NBI	“Irregular vessels and mucosa” (pattern C)	Accuracy 95%; 95%CI: 90%-99%; LR+ = 44.33

EGC: Early gastric cancer; GC: Gastric cancer; NBI-ME: Narrow-band imaging with or without magnification.

Table 2 Gastric cancer differentiation using narrow-band imaging with or without magnification

Ref.	Endoscopic technique	Differentiated-type EGC (D-EGC)	Undifferentiated-type EGC (UD-EGC)
Nakayoshi et al[161]	NBI-ME	Relatively regular fine network pattern	Relatively irregular, twisting or corkscrew pattern, with a relatively low density of microvessels
Endo et al[162]	NBI-ME	Grid network pattern with hypervascularity	Short twig or branch-like pattern with hypovascularity
Tamai et al[163]	NBI-ME describing depressed gastric adenomas vs protruding adenomas	Intramucosal carcinomas were more frequently found in depressed adenomas (reddish in color, a regular ultrafine network pattern of mucosal microvasculature) (25%) than in protruding adenomas (4.5%)
Yao et al[164]	NBI-ME	WOSa white substance within the neoplastic epithelium that may obscure the subepithelial microvascular pattern. More frequent in non-advanced neoplasia than in advanced carcinomas and that 100% of non- advanced lesions demonstrated a regular distribution of WOS
Yokoyama et al[165]	NBI-ME	Amongst the D-EGC lesions, fine-network pattern, intra-lobular loop pattern-1, intra-lobular loop pattern-2 and corkscrew pattern were observed in 15.7%, 59.6%, 24.2% and 0.5%, respectively. D-EGCs mainly exhibited fine-network pattern or intra-lobular loop pattern	In UD-EGC intra-lobular loop pattern-2 and corkscrew pattern were observed in 41.2% and 58.8%, respectively. Therefore, UD-EGCs were all classified as intra-lobular loop pattern-2 and corkscrew pattern

NBI-ME: Narrow-band imaging with or without magnification; EGC: Early gastric cancer.

Table 3 Studies evaluating the horizontal extent (DL) of early gastric cancer on narrow-band imaging with or without magnification

Ref.	Endoscopic technique	Aim of the study	Results
Okada et al[166]	NBI-ME	Assessment the comparative relationship between NBI-ME images and histopathological findings in patients with UD-type EGCs prior to either ESD or surgery	NBI-ME images of UD-type EGCs proved to be very closely related to the histopathological findings
Nonaka et al[167]	NBI-ME	Estimating a DL on NBI-ME in comparison with biopsy findings as a gold-standard	The DL that could be recognized at 2 points on the orifice and anal sides of each lesion during ME-NBI was consistent with the pathological findings in 22 patients with 0-IIc lesions, 7 with 0-IIb lesions, and 2 with 0-IIb + IIc lesions, showing an accuracy of 100%
Kiyotoki et al[168]	NBI-ME vs ICC	Evaluated the usefulness of NBI-ME for determining the tumor margin compared with ICC (indigocarminechromoendoscopy)	The rate of accurate marking of the ME-NBI group was significantly higher than that of the ICC group (97.4% vs 77.8%, respectively; P = 0.009)
Nagahama et al[169]	NBI-ME vs CE	To investigate the usefulness and limitations of NBI-ME when CE is unsuccessful for determining the horizontal extent of EGC	The proportion of cancers showing unclear margins using CE was 18.9% (66/350). Of these, 62 of 66 cancers were examined using ME with NBI, with the entire margins successfully delineated in 72.6% (45/62) of the lesions that had shown unclear margins using CE. The success rate was 0% for undifferentiated cancers, significantly lower than that for differentiated lesions (P < 0.00001)

UD-type EGCs: Undifferentiated-type EGCs; ESD: Endoscopic submucosal dissection; ICC: Indigo carmine chromoendoscopy; NBI-ME: Narrow-band imaging with or without magnification.

Table 4 Studies evaluatingthe horizontal extent (DL) of early gastric cancer on flexible spectral imaging color enhancement

Ref.	Endoscopic technique	Aim of the study	Results
Jung et al[170]	FIME vs WLME	Discrimination of non-neoplastic lesion, adenoma, and cancer of the stomach	The proportion of agreement and the degree of agreement between endoscopic and pathological diagnosis by WLME were 0.85 and 0.76, respectively, and those by FIME were 0.91 and 0.86, respectively
Mouri et al[171]	FICE vs WLE	78 differentiated, 22 undifferentiated EGC were analyzed before an endoscopic or surgical resection	The score of the FICE observation improved in 46 cases (46%), was unchanged in 54 cases (54%), and decreased in no cases (0%)
Tanioka et al[172]	FICE with ultraslim endoscopy vs WLE	Endoscopy focusing on the enhanced contrast between tumor and non-tumor lesions	Visibility with FICE was superior to WLE in 54% of the observations and comparable to WLE in 46% of the observations
Osawa et al[173]	small-caliber FICE vs WLE	Evaluate median color differences between malignant lesions and the surrounding mucosa	Greater median color differences were present in FICE images compared with WLE, resulting in images with better contrast (27.2 vs 18.7, P < 0.0001)
Osawa et al[174]	OBI(without magnification and with 40-fold magnification) vs WLE	Delineating the depressed-type EGC	DL of the depressed-type EGC was easily identified by OBI without magnification in 26 of 27 cases (96%)
Yoshizawa et al[175]	OBI vs WLE	The identification of the DLs of an elevated-type EGC without Magnification and the rate of success in identifying the abnormal surface structure of GC by using low-magnified OBI images	DLs were easily identified in OBI images, even without magnification
Jung et al[170]	FIME vs WLME	The discrimination of non-neoplastic lesion, adenoma, and cancer of the stomach	The proportion of agreement and the degree of agreement between endoscopic and pathological diagnosis by WLME were 0.85 and 0.76, respectively, and those by FIME were 0.91 and 0.86, respectively
Dohi et al[176]	I-FICE vs WLE, FICE and CE	To evaluate the usefulness of I-FICE in EGC demarcation	The median ranking score for I-FICE images was significantly higher than that obtained from the other methods

FIME: Magnifying endoscopy with flexible spectral imaging color enhancement system; WLME: White light magnifying endoscopy; FICE: Flexible spectral imaging color enhancement; OBI: Optimal band imaging; DL: Demarcation line I-FICE- FICE with indigo carmine; CE: Chromoendoscopy (indigo carmine).