Esophagogastric junction adenocarcinoma shares characteristics with gastric adenocarcinoma: Literature review and retrospective multicenter cohort study

Abstract The incidence of esophagogastric junction (EGJ) adenocarcinoma has been gradually increasing in Asia, just like in Western countries a few decades ago. Despite recent advances in next‐generation sequencing and multimodal treatments, EGJ adenocarcinoma is still an aggressive malignancy with poor outcomes. Clinically, EGJ adenocarcinoma can be separated into Barrett's adenocarcinoma and cardiac adenocarcinoma, with frequent similarities observed. Barrett's adenocarcinoma is likely to be of gastric origin in terms of its premalignant background, risk factors, and stem cell regulators. Recent comprehensive genomic analyses suggest that immunotherapy may be essential for high‐level microsatellite instability (MSI‐H)‐ and Epstein‐Barr virus (EBV)‐associated subtypes, and against the immunosuppressive phenotype in genomically stable (GS) subtypes, in the treatment of EGJ and gastric adenocarcinoma. Although the chromosomal instability (CIN) subtype dominates EGJ adenocarcinoma, there is still a need to investigate the other molecular subtypes and their targets. Because of the distinctive characteristics of tumor location of EGJ adenocarcinoma, we also described the results of a multicenter cohort study of EGJ adenocarcinoma, comparing Siewert type I (distal esophagus), II (cardia of the stomach), and III (subcardia) tumors. We show that type I tumors were frequently accompanied by Barrett's esophagus (78%, P < .0001), with a significantly unfavorable outcome (multivariate EGJ‐cancer‐specific mortality hazard ratio = 1.81, 95% CI, 1.06‐2.97; P = .031). In addition, over half (56%) of these cases experienced disease recurrence in the lymph nodes. Our findings suggest that Barrett's adenocarcinoma may be an aggressive phenotype of EGJ adenocarcinoma due to the potential risk of tumor spread through the complex lympho‐vascular network of the esophagus.


| BARRE T T ' S AND C ARD IAC ADENOC ARCINOMA
There has been a gradual increase in the incidence of esophagogastric junction (EGJ) adenocarcinoma in Asian countries, 1-4 including Japan. 5 Despite the recent advances in comprehensive genetic analyses as well as the progress in multimodal treatments, EGJ adenocarcinoma is still an aggressive malignancy with poor outcomes. EGJ adenocarcinoma includes Barrett's adenocarcinoma (also known as esophageal adenocarcinoma, or adenocarcinoma in distal esophagus) and cardiac adenocarcinoma (adenocarcinoma of gastric cardia) with esophageal invasion. 6 Barrett's and cardiac adenocarcinomas have been increasing in parallel since the late 1970s in Western countries, and are recognized as common upper gastrointestinal cancers. 7 There are numerous similarities between Barrett's and gastric adenocarcinomas in terms of tumor characteristics and background, with few differences highlighted to date.
In histological examinations, Barrett's adenocarcinoma and intestinal-type gastric adenocarcinoma have a common premalignant background of intestinal metaplasia caused by chronic inflammation.
Intestinal-type gastric adenocarcinoma in the body and antrum of the stomach is mainly related to Helicobacter pylori-induced chronic gastritis. 8 Helicobacter pylori infection can also induce both chronic inflammation and metaplasia in the gastric cardia and cardiac adenocarcinoma. [9][10][11][12][13] In addition, in terms of risk factors, both Barrett's and cardiac adenocarcinoma are associated with obesity in Western countries. [14][15][16][17] Adipokines produced from adipose tissue in association with metabolic syndrome can influence the development of chronic inflammation and cancer progression. 18,19 Finally, experimental data from transgenic mouse models suggest that Barrett's metaplasia may arise from gastric cardia progenitor cells in response to bile acid-mediated inflammation via LGR5 expression and IL-1β-IL-6 signaling. 20 Cholecystokinin 2 receptor (CCK2R, also known as CCKBR), which regulates gastric stem cells in the cardia or antrum regions of the stomach, is also upregulated in Barrett's esophagus and in esophageal adenocarcinoma. 21,22 Although a few studies have proposed the esophageal submucosal gland as the origin of Barrett's adenocarcinoma, considering previous perspective studies, Barrett's adenocarcinoma is likely to be of gastric origin. 23,24

| C AND IDATE MOLECULE S OF THER APEUTI C TARG E TS
Previously, gastroesophageal tumors were predominantly classified by pathological classification; however, in recent years -in the era of next-generation sequencing technology -a molecular taxonomy has emerged. [25][26][27][28] The Cancer Genome Atlas (TCGA) Network has shown that gastroesophageal adenocarcinoma can be categorized into four molecular subtypes: Epstein-Barr virus (EBV)-associated, high-level microsatellite instability (MSI-H), genomically stable (GS), and chromosomal instability (CIN) tumors. These subtypes are classified using a range of techniques, including somatic copy number aberration, whole-genome and whole-exon sequencing, RNA sequencing, methylation assays, and proteomics analysis. 26,28 In the multiomic data of EGJ adenocarcinoma acquired from TCGA from cBioPortal for Cancer Genomics (https://www.cbiop ortal.org), there were a total of 172 cases which were classified as gastroesophageal junctional categories, including seven cases (4.1%) of MSI-H, six cases (3.5%) of EBV, 11 cases (6.4%) of GS, and 148 cases (86%) of CIN. 26,28 In this section, we discuss the potential therapeutic targets of adenocarcinoma from esophagogastric junction (EGJ) as well as gastric adenocarcinoma, according to these four molecular subtypes. immunogenic, and the host will often activate anti-tumor immunity against these neoantigens. 30 In addition, MSI-H tumors can upregulate immunological checkpoints such as programmed death receptor-1 (PD-1), programmed death ligand-1 (PD-L1), or PD-L2, in order to escape the host anti-tumor immunity. In a randomized controlled clinical trial, Nivolumab treatment was shown to exert a significant survival benefit to patients with metastatic gastric or EGJ adenocarcinoma. 31,32 Nivolumab is an immune checkpoint inhibitor that blocks PD-1. It may therefore be more effective for an MSI-H population.
Epstein-Barr virus-associated tumors show hypermethylation, CDKN2A silencing, frequent mutations in PIK3CA and ARID1A, and immune reactivity. 26,28 EBV-associated tumors seem to occupy only a small fraction of EGJ adenocarcinoma in TCGA data. Drugs that inhibit the PI3K pathway or methylation may be potentially beneficial for these tumors. Recent comprehensive genomic analyses revealed a significant mRNA expression of PD-L1 and PD-L2 in EBV-associated gastric adenocarcinoma, suggesting that, like MSI-H tumors, these tumors are also sensitive to immune checkpoint inhibitors, such as Nivolumab. 28,31 Genomically stable tumors are associated with diffuse histology, and present with frequent mutations in CDH1 and RHOA, and the CLDN18-ARHGAP26 fusion gene. 26,28 Besides these major alterations, mutations in BRCA1, BRCA2, RAD51C, PALB2, and CTNNA1 are detected in diffuse-type gastric adenocarcinoma. 33  Note: (%) Indicates the proportion of cases with specific clinicopathological features for each Siewert classification group. * P < .05, **P < .01, ***P < .001.

| CLINIC AL AND PROG NOS TI C D IFFEREN CE S BY TUMOR LO C ATI ON
Esophagogastric junction adenocarcinoma extends across the thorax and abdomen to varying degrees. For patients with EGJ adenocarcinoma, tumor location is specified by Siewert classification as follows: type I, defined as tumors of the distal esophagus, in which the epicenter of the tumor is located 1-5 cm above the anatomical EGJ; type II, true junctional tumors, in which the epicenter is located 1 cm above and 2 cm below the EGJ; and type III, gastric tumors that infiltrate into the esophagus, for which the epicenter is located between 2 and 5 cm below the EGJ. 50 Although there does not seem to be any difference in terms of the carcinogenic origin of Barrett's and cardiac adenocarcinoma, as mentioned earlier, this anatomical classification may differentiate these two tumors in terms of clinicopathological or prognostic characteristics. Considering that Barrett's esophagus involves a replacement of normal squamous epithelium with metaplastic mucosa in response to gastroesophageal reflux disease (GERD), Barrett's adenocarcinoma is likely to be located more proximally (Siewert type I) than cardiac adenocarcinoma. 51   Cancer Control (UICC) classification of esophageal cancer, which is applicable to Siewert type III tumors.

| Clinicopathological and survival analysis in terms of tumor location
The baseline characteristics according to Siewert classification are shown in Table 2. Type I tumors were observed in 59 cases (15%), type II in 280 cases (71%), and type III in 56 cases (14%). Type I tumors were associated with recent cases (P < .0001), more frequently accompanied by Barrett's esophagus (78%, P < .0001), less frequently associated with Helicobacter pylori infection (P < .0001), and showed a less-advanced disease stage as compared with type II or III tumors (P = .0002). Because surgeons often select a transthoracic approach for type I tumors (P < .0001), operative time was the longest for type I tumors (P < .0001), with less blood loss and a successful R0 resection rate (P = .0041). Patients with type III tumors exhibited a significantly larger size (P < .0001), because the radius of the tumor in type III tumors must be at least 2 cm to invade into the  Note: The multivariate Cox regression model included the same set of covariates selected as in Table 2.
esophagus. Accordingly, type III cases included more advanced disease stage, longer operative time with more blood loss, and a lower R0 resection (Table 2).
In the survival analysis, there were 192 deaths, including 135 EGJ-cancer-specific deaths, over a median follow-up of 5.3 years (interquartile range, 5.0-6.6 years) for censored cases. Kaplan-Meier analyses according to disease stage are provided in Figure 2A-C.  Figure 2A). Kaplan-Meier analyses according to Siewert classification (N = 395, Figure 3A-C) showed that the 5-year EGJ-cancerspecific mortality of Siewert type III tumors was the worst (50.4%, Figure 3A) (Tables   S2 and S3), there was no significant association between Barrett's esophagus and patient outcome in multivariate survival analysis (Tables S4 and S5).

ACK N OWLED G EM ENTS
We would like to express our sincere appreciation for data acquisition to Ikumi Haraguchi (Cancer Institute Hospital).

CO N FLI C T O F I NTE R E S T S
None of the authors has any conflict of interest related to this study.