Newly developed magnifying endoscopic classification of the Japan Esophageal Society to identify superficial Barrett’s esophagus-related neoplasms

Aim and methods The Japan Esophageal Society created a working committee group consisting of 11 expert endoscopists and 2 pathologists with expertise in Barrett’s esophagus (BE) and esophageal adenocarcinoma. The group developed a consensus-based classification for the diagnosis of superficial BE-related neoplasms using magnifying endoscopy. Results The classification has three characteristics: simplified, an easily understood classification by incorporating the diagnostic criteria for the early gastric cancer, including the white zone and demarcation line, and the presence of a modified flat pattern corresponding to non-dysplastic histology by adding novel diagnostic criteria. Magnifying endoscopic findings are composed of mucosal and vascular patterns, and are initially classified as “visible” or “invisible.” Morphologic features were evaluated for “visible” patterns, and were subsequently rated as “regular” or “irregular,” and the histology, non-dysplastic or dysplastic, was predicted. Conclusion We introduce the process and outline of the magnifying endoscopic classification.


Introduction
Barrett's esophagus (BE)-related neoplasms, including esophageal adenocarcinoma (EAC), are still rare in Japan [1,2]. However, the increasing number of patients with EAC because of the increasing prevalence of gastroesophageal reflux disease and BE is concerning [3]. Early detection is crucial for good quality of life and favorable prognosis of patients with EAC. Superficial BE-related neoplasms (SBN), including dysplasia and particularly EAC with flat macroscopic type, are often difficult to detect by white light endoscopy alone [4,5]. Studies showed the utility of magnifying endoscopy to detect SBN and proposed several magnifying endoscopic classifications based on narrow-band imaging (NBI) findings [6][7][8][9]. However, these classifications involve complicated and diverse criteria, making them difficult to use in clinical practice by general endoscopists.
We, the committee members of the Japan Esophageal Society (JES), developed a new magnifying endoscopic classification of the JES for predicting the histology of Barrett's epithelium, and named it as JES-BE classification. We introduce the process and outline of the JES-BE classification system.

Working group
The JES created a working committee group in 2012 consisting of 11 expert endoscopists and 2 pathologists with expertise in gastrointestinal neoplasms, including SBN, who are from 10 domestic institutions of a high-volume center or an academic university hospital. The working committee members were assembled during the 66th annual meeting 1 3 of the JES and convened nine times to proceed with the creation, assessment, and examination of the JES-BE classification system.

Development of consensus-based classification system
The JES working group collected high-definition magnification NBI (HM-NBI) images of 20 cases of non-dysplastic BE, including specialized intestinal metaplasia (SIM), and 25 cases of dysplastic BE (i.e., SBN, including low-and high-grade dysplasia, and flat-type superficial EAC invading up to the submucosa) from 10 domestic hospitals of the working group members. The working group discussed about mucosal and vascular patterns of the images and histologic findings at the mucosal site where the images had been obtained. The working group developed the JES-BE classification system based upon consensus among the working committee members. Table 1 shows the JES-BE classification system. The mucosal and vascular patterns in each HM-NBI image were initially classified as "visible" or "invisible". The "invisible" pattern cannot be subclassified. The "visible" mucosal and vascular patterns were subclassified as "pit" or "non-pit" and "net" or "non-net", respectively. The "pit" type mucosal pattern was marked by a circular pattern, and the "non-pit" was marked by tubular, linear, or ridged/villous patterns ( Table 3). The "net" type vascular pattern was marked by a network of vessel formation connected to each other, and the "non-net" type was marked by vessels without network formation [10]. Finally, each pattern was classified as either "regular" or "irregular" based on the diagnostic criteria agreed upon by the working group. The diagnostic criteria for regularity are listed in Table 2.

The JES-BE classification system
Regularity of mucosal pattern is evaluated based on form, size, arrangement, density, and white zone. The white zone has been used in the M-NBI diagnosis of the early gastric cancer. Yagi et al. [11] initially mentioned a white zone that can be seen as a whitish edge of mucosal pattern of the early gastric cancer under magnification NBI observation. They described the hypothesis of a mechanism of white zone visibility. They suggested that cancerous lesions often showed indistinct or invisible white zone that would correspond to absent surface pattern and disappearance of the fine mucosal structure in the previous studies on the early gastric cancer [12,14]. We, the working group members, reached a consensus that clearly visible and indistict or invisible white zones would be suggestive of non-cancerous and cancerous lesions, respectively.
Regular mucosal patterns were marked by showing similar forms and/or homogeneous sizes, regular arrangement, and clearly visible white zone. Irregular mucosal patterns were marked by various forms, heterogeneous sizes, irregular arrangement, high density, and indistinct or invisible white zone [11][12][13].
Regular vascular patterns were marked by their location between or along mucosal ridges with similar forms, gently bending, regularly branching, and gradual changes in the vessel's caliber. Irregular vascular patterns were marked by not following mucosal structures with various forms, steeply bending, irregularly branching, and abrupt changes in the vessel's caliber [6,12,13].
If magnifying endoscopic findings would show both of regular and irregular patterns, the irregular pattern was representative of the endoscopic findings to prevent form missing the chance of biopsy for a suspected dysplastic lesion. "Unclassified" mucosal or vascular pattern that could not be classified into "regular" or "irregular" was rated as an "irregular" pattern because tissue samples from mucosal sites with "unclassified" pattern should be obtained for biopsy, similar to that with "irregular" pattern ( Table 2).
Flat pattern was originally defined as flat mucosa (none of pits and villi: i.e., absent pattern) with normal-appearing long-branching vessels in a previous study [6]. The original flat pattern was classified into a "regular" pattern and suggested to be of non-dysplastic BE, including SIM. The flat mucosal pattern and the long-branching vessels will mimic an absent mucosal pattern and often coexisted with steeply bending or tortuous vessels appearing an irregular vascular pattern, respectively. Both latter M-NBI patterns were suggestive of the early gastric cancer [12,14]. Thus, interpretation of the flat pattern as a regular pattern that corresponds to non-dysplastic (i.e., non-cancerous) histology will be difficult, particularly among endoscopists with expertise in the M-NBI diagnosis of the early gastric cancer. A previous study showed that a distinct demarcation line was a primarily important diagnostic criterion for the early gastric cancer [14]. Our study demonstrated that none of non-dysplastic lesions with flat pattern have a distinct demarcation line [15]. In addition, greenish thick vessels are frequently seen in non-dysplastic lesion with flat pattern. Following these results, "no clear demarcation" and "a greenish thick vessel" were incorporated in the definitions of a modified flat pattern in this classification. We classified the modified flat pattern independently as "regular" mucosal/vascular patterns ( Table 2).

Predicted histology based on JES-BE classification
Visible mucosal/vascular patterns allow histology prediction, whereas invisible mucosal/vascular patterns do not. Regular and irregular mucosal/vascular patterns are suggestive of non-dysplastic and dysplastic histology, respectively. Unclassified mucosal/vascular patterns are suggestive of dysplastic histology. Flat pattern that meets modified criteria of this classification is rated as a regular pattern and suggestive of non-dysplastic histology. When the mucosal/ vascular patterns are visible and graded differently (i.e., one regular and other irregular or invisible), predicted histology is determined non-dysplastic or dysplastic based on a comprehensive diagnosis of regular or irregular (Fig. 1). Figure 1 shows a diagnostic flowchart based on this classification. If a lesion was found, we initially assess a mucosal pattern that would be visible under moderately to fully magnifying observation. We then assess a vascular pattern that would be visible under fully magnifying observation. Based on mucosal and vascular pattern, we finally make comprehensive diagnosis and predict histology.

Relationships between JES-BE and other classifications (Table 3)
Recently, published NBI classification of BE classified absent mucosal pattern and normal-appearing long-branching vessels as "irregular" and "regular", respectively [13]. Both patterns were the original criteria for a flat pattern [6]. To resolve the discrepancy, a flat pattern is independently classified as "regular" mucosal/vascular patterns by the modified diagnostic criteria in the JES-BE classification (Tables 2, 3). The JES-BE classification will be simpler than well-known traditional classifications [6, 7,9]. Table 3 Relationships between currently known and newly developed classifications of surface patterns for predicting histology of Barrett's epithelium ABV abnormal vessels, BING Barrett's international narrow-band imaging group a Including normal-appearing long-branching vessels that is an original diagnostic criterion for a flat pattern b Flat pattern defined by modified criteria c Regularity marked by diagnostic criteria ( Table 2) Kara et al. [6] Sharma et al. [7] Anagnostopoulos et al. [

Conclusion
We developed the JES-BE classification and introduced the process and outline of the classification system. Further studies are needed to validate the diagnostic validity and reliability of the magnifying endoscopic classification.