Two Updates on Oesophagogastric Junction Adenocarcinoma from the Fifth WHO Classication of Tumors of the Digestive System: Alteration of Denition and Emphasis on HER2 Test.

Background: The incidence of oesophagogastric junction adenocarcinoma has increased rapidly but remains controversial over the last decades. There are two crucial updates of the fth World Health Organization(WHO) classication, including the alteration of its denition and the emphasize on human epidermal growth factor receptor 2 (HER2) test. Methods: A total of 566 clinicopathological samples from patients who were diagnosed with gastric adenocarcinoma were retrospectively analyzed. We comprehensively compared the clinicopathological features of oesophagogastric junction adenocarcinoma between fourth (V4.0) and fth (V5.0) WHO version. The clinicalpathological features among oesophagogastric junction, proximal and distal gastric tumors with fourth and fth edition were also compared, respectively. Besides, we discussed the correlation of the HER2-expression with clinicopathological features according to the V5.0. Results: The results showed that the difference was mainly between oesophagogastric junction and distal adenocarcinoma in V4.0, while some were found between proximal and distal adenocarcinoma in V5.0. Tumors invading oesophagus more than 3cm were still mainly oesophagogastric junction tumors. The expression of HER2 in oesophagogastric junction and proximal gastric adenocarcinoma was still higher than that in gastric body and distal site. Conclusions: The clinicopathological parameters of the oesophagogastric junction tumors changed to some extent in the updated WHO version. The proximal gastric tumors tended to be more invasive, rather than that located in oesophagogastric junction. But the latter with oesophageal invasion required additional management. The HER2-expression of oesophagogastric junction adenocarcinoma is the highest. The classication of the V5.0 is reasonable and worth a recommendation. The rabbit monoclonal antibodies included anti-CDX-2 (RMA-0631, Maxim, Fuzhou, China), anti-CK7 (Kit-0021, Maxim), anti-CK20 (Kit-0025, Maxim) anti-KI67 (MIB-1, Maxim), anti-C-erbB-2 (EP3, Maxim) and antibodies for mismatch repair protein (MMR:MLH1, PMS2, MSH2 and MSH6). All procedures were performed in the EnVision System by a Benchmark-ULTRA automatic immunohistochemical staining instrument (Asia-core, China). HER2 scoring system proposed by Hoffman et al. was set as the criteria [19]. HER2 status was considered negative (HER2-) with scores of 0 and 1(No membranous reactivity in <10% or faint or barely perceptible reactivity in ≥ 10% of tumor cells). HER2 with score of 3 (strong and complete basolateral membranous reactivity in ≥ 10 of tumor cells) was considered as HER2-amplication. HER2 status with score 2 was considered positive unless tested for gene amplication. The status of microsatellite was evaluated by four markers of mismatch repair protein (MMR), including microsatellite stability (MSS) with four positive markers and microsatellite instability of high frequency (MSI-H) with deciency of more than two markers [20]. software for Windows, Version 18. The clinicopathological parameters were collected according to a standardized protocol. The Mann-Whitney U test and Kruskal-Wallis test were performed to assess the difference of clinicopathological features among oesophagogastric junction adenocarcinoma, proximal and distal gastric tumors. The difference of clinicopathological features between oesophagogastric junction adenocarcinoma of V4.0 and V5.0, and the correlation between clinicopathological features and HER2-expression were evaluated with the Mann-Whitney U test and Fisher’s exact test. P-value <0.05 was considered signicantly different.


Background
Gastric cancer is the fth most common cancer worldwide and adenocarcinoma of oesophagogastric junction has drawn considerable attention, because of its remarkable increasing incidence [1]. Compared with oesophageal and gastric carcinoma, oesophagogastric junction adenocarcinoma requires different surgical procedure, as well as lymph nodes dissection. However, it remains unclear whether the tumor is of gastric or oesophageal origin. In particular, the computed tomography (CT) has a low speci city (44%) when diagnosing Siewert type II oesophagogastric junction adenocarcinoma [2,3]. The location of pathological anatomy was considered the most accurate after operation, but the de nition of oesophagogastric junction adenocarcinoma remains controversial in the last decades. Siewert et al has classi ed oesophagogastric junction adenocarcinoma into three types, including type I with its epicentre 1-5cm above the oesophagogastric junction, type II with its epicenter between 1 cm above and 2 cm below the junction, and type III with its epicenter 2-5 cm distal from the junction [4]. Meanwhile, Nishi et al. proposed that oesophagogastric junction tumor is located 2 cm above and 2 cm below the oesophagogastric junction, regardless of its different histological subtype [5]. Except Japan, the Siewert's classi cation has been widely used to distinguish the oesophagogastric junction adenocarcinoma and recommended in the fourth World Health Organization (WHO) classi cation (V4.0) of tumors of the digestive system.
In 2019, the fth WHO classi cation (V5.0) of tumors of the digestive system rede ned the oesophagogastric junction adenocarcinoma as having its epicenter from 5 cm to 2 cm, showing almost no difference with Nishi's classi cation [6]. The new de nition of oesophagogastric junction adenocarcinoma is similar with Siewert II. Efforts have been made to discuss the difference of clinicopathological features among the three Siewert subtypes. For example, a study conducted by Feith Met al demonstrated that patients with Siewert III tended to have advanced stage, with more extensive lymph node involvement and worse survival [7]. However, some data showed in Eastern, the outcome of Siewert II was similar to Siewert III [8]. Moreover, when compared with distal gastric tumors, oesophagogastric junction adenocarcinoma was associated with worse outcome [9]. As for the difference of clinicopathological features among oesophagogastric junction adenocarcinoma subtypes, there is no consistent conclusion. Therefore, a detailed and comprehensive analysis among the clinicopathological features is necessary, not only between the V4.0 and the V5.0 for oesophagogastric junction adenocarcinoma, but also between the gastric and oesophagogastric junction adenocarcinoma.
Besides, it is now increasingly clear that human epidermal growth factor receptor 2 (HER2) should be detected routinely to identify patients who may bene t from the target therapy of trastuzumb, which has been proposed in the V5.0 [6]. HER2 is a pro-oncogene encoded by erbB2 on chromosome 17 and its ampli cation may result in angiogenesis, tumorigenesis and excessive cell growth in several tissues such as colon, ovary, bladder, uterine and breast [10]. HER2 has been evaluated in many studies focusing gastric or oesophagogastric junction adenocarcinoma. With the change of scoring criteria and the de nition of oesophagogastric junction adenocarcinoma, the HER2-amplication rate exhibits a great discrepancy, which varies with an extremely wide range, from 4-53% (median, 20.2%) [11,12]. Previous studies showed that HER2-positivity rate was higher in oesophagogastric junction than gastric adenocarcinoma [13,14], while some studies suggested that the presence of HER2-expression was not associated with tumor location [15,16]. HER2-expression is heterogeneous in gastric or oesophagogastric junction adenocarcinoma, thus it's important to review the HER2 status and its correlating clinicopathological features according to the V5.0.
In the V5.0, the diagnosis of oesophagogastric junction adenocarcinoma and the emphasis on HER2 test were updated. In this study, we compared the clinicopathological features between gastric and oesophagogastric junction adenocarcinoma with V4.0 and V5.0, respectively. With that, we sought to assess the relationship between clinicopathological features and HER2-expression with V5.0, aiming to make an in-depth and comprehensive understanding about the V5.0 for oesophagogastric junction adenocarcinoma.

Case Selection
A retrospective analysis was conducted. A total of 566 patients were included, among of which, 464 were gastric adenocarcinoma and the remains were oesophagogastric junction adenocarcinoma (V4.0). All patients received radical resection of the tumor at Sichuan Cancer Hospital & Institute between 2016 and 2019. Patients who received neoadjuvant therapies were excluded. The study was approved by the ethics committee of Sichuan Cancer Hospital. .

Immunohistochemistry
The rabbit monoclonal antibodies included anti-CDX-2 (RMA-0631, Maxim, Fuzhou, China), anti-CK7 (Kit-0021, Maxim), anti-CK20 (Kit-0025, Maxim) anti-KI67 (MIB-1, Maxim), anti-C-erbB-2 (EP3, Maxim) and antibodies for mismatch repair protein (MMR:MLH1, PMS2, MSH2 and MSH6). All procedures were performed in the EnVision System by a Benchmark-ULTRA automatic immunohistochemical staining instrument (Asia-core, China). HER2 scoring system proposed by Hoffman et al. was set as the criteria [19]. HER2 status was considered negative (HER2-) with scores of 0 and 1(No membranous reactivity in <10% or faint or barely perceptible reactivity in ≥10% of tumor cells). HER2 with score of 3 (strong and complete basolateral membranous reactivity in ≥10 of tumor cells) was considered as HER2-ampli cation. HER2 status with score 2 was considered positive unless tested for gene ampli cation. The status of microsatellite was evaluated by four markers of mismatch repair protein (MMR), including microsatellite stability (MSS) with four positive markers and microsatellite instability of high frequency (MSI-H) with de ciency of more than two markers [20].

Fluorescence in situ hybridization (FISH)
The level of HER2 ampli cation were tested with the cases whose IHC score was two. FISH test of HER2-ampli cation was performed with PanthVysion kit (GSP, LBP, Guangzhou, China). The evaluation towards HER2-ampli cation was based on the ratio of HER2 to centromere 17 copy number, according to the guidelines of 2007 ASCO/CAP [21]. Cases were considered gene ampli ed for the HER2/CEP17 ratio of 2.2, equivocal with ratio less than 2.2 but more than 1.8 and negative with the ratio less than 1.8. The equivocal cases were not selected in our cohort.

Statistical analysis
Data was analyzed with SPSS software for Windows, Version 18. The clinicopathological parameters were collected according to a standardized protocol. The Mann-Whitney U test and Kruskal-Wallis test were performed to assess the difference of clinicopathological features among oesophagogastric junction adenocarcinoma, proximal and distal gastric tumors. The difference of clinicopathological features between oesophagogastric junction adenocarcinoma of V4.0 and V5.0, and the correlation between clinicopathological features and HER2-expression were evaluated with the Mann-Whitney U test and Fisher's exact test. P-value <0.05 was considered signi cantly different.

Results
Clinicopathological features of gastric and oesophagogastric junction adenocarcinoma with V4.0 The Clinicopathological features of gastric (proximal gastric, gastric body and distal gastric)) and oesophagogastric junction adenocarcinoma (V4.0) were summarized in Table 1. There was no signi cant difference in patient age, histological type, degree of differentiation, Mclassi cation and HER2-status between gastric and oesophagogastric junction adenocarcinoma. Compared with distal gastric cancer, oesophagogastric junction adenocarcinoma was associated with larger tumor (P < 0.001), higher T classi cation (P < 0.001), more frequent lymphovascular invasion (P = 0.022) and more frequent nodal metastases (P = 0.011). In addition, oesophagogastric junction adenocarcinoma had more lymph node metastasis than gastric body (P = 0.015) and distal gastric tumors (P = 0.006). There was no difference in other parameters among oesophagogastric junction, proximal gastric and gastric body adenocarcinoma. HER2 status was merely different between proximal and distal gastric (P = 0.001). According to the new version (V5.0), the clinicopathological features of gastric and oesophagogastric junction adenocarcinoma were compared in Table 2. There was no signi cant difference in patient age, histological type, N classi cation, M classi cation, sites of lymph node metastasis and HER2 status between gastric and oesophagogastric junction adenocarcinoma. Compared with proximal gastric, oesophagogastric junction adenocarcinoma had higher degree of differentiation (P = 0.029). Besides, although oesophagogastric junction tumors were smaller than proximal gastric (P = 0.041) and gastric body (P = 0.032), it had advanced T classi cation than distal gastric (P = 0.012) tumors. The comparison of HER2 status was consistent with that in V4.0, suggesting signi cant difference was only detected among proximal and distal gastric (P = 0.021).   Comparison of clinicopathological features of oesophagogastric junction adenocarcinoma between V4.0 and V5.0.
In order to have a more comprehensive understanding of the impact of the change in the de nition of oesophagogastric junction adenocarcinoma, we compared the clinicopathological features between the fourth and fth edition of oesophagogastric junction adenocarcinoma. All cases of oesophagogastric junction adenocarcinoma were divided into two groups: group 1was diagnosed with V4.0 criteria, within 2 cm to 5 cm from oesophagogastric junction, while group 2 was diagnosed with V5.0 criteria, within 2 cm from oesophagogastric junction. There were 45 (44%) cases of group 1 and 57 (56%) cases of group 2. Candidates in group 2 showed smaller tumor (P < 0.001), earlier T classi cation (P = 0.021) and N classi cation (P = 0.035). There was no signi cant difference in other parameters between the two groups.

Correlation between HER2 status and clinicopathological characteristics
We evaluated the association between HER2 status and clinicopathological features among all 566 cases and the cases were divided according to the criteria ofV5.0. The incidence of HER2-expression in oesophagogastric junction (25%) and proximal gastric (22%) were dramatically higher than gastric body (13%) and distal gastric (11%) tumors (P=0.001). HER2-positivity was more common in low-moderate differentiation cases than the highers (P<0.001). Moreover, tumors of HER2-expression were distinctly associated with larger tumor (P<0.001) and more advanced M classi cation (P=0.039). Interestingly, there was no statistically difference in T classi cation and N classi cation. Although no difference in MMR status, there was merely one case of MSI-H in HER2-expression tumors.

Discussion
The Siewert's classi cation is widely used for surgical procedure of oesophagogastric junction adenocarcinoma, while the Nishi's classi cation from Japan is almost the same as the Siewert II. The comparation among Siewert subtypes has been well studied [22], as well as those between gastric and oesophagogastric junction adenocarcinoma. In V5.0, the de nition of oesophagogastric junction adenocarcinoma was the same as Nishi's classi cation in addition to emphasis on adenocarcinoma. The change in the de nition of oesophagogastric junction tumors is signi cantly important for clinicopathological assessment and clinical management, such as the surgical dissection procedure.
In this study, we compared the clinicopathological features between gastric and oesophagogastric junction adenocarcinoma with V4.0 and V5.0, respectively. The oesophagogastric junction adenocarcinoma had more advanced T-classi cation than distal gastric adenocarcinoma with the criteria of both the V4.0 and V5.0. This was concurred by previous studies, that proximal (oesophagogastric junction and cardia) tumors were associated with poor outcomes [23,24]. However, there were still some changes between the clinicopathological features of V4.0 and V 5.0.When compared with V4.0, the differences of N-classi cation and the sites of lymph node metastasis were mainly focused on oesophagogastric junction and distal gastric adenocarcinoma. But these differences were mainly found on proximal gastric and distal gastric adenocarcinoma with the V5.0. This was perhaps because the oesophagogastric junction tumors of 2 cm-5 cm from the oesophagogastric junction with V4.0were classi ed to the proximal gastric tumors in the V5.0. Moreover, the tumors of 2 cm-5 cm was same as the Siewert III. A multivariate analysis showed only lymph node metastases predicted the survival of gastric carcinoma [25]. The majority of patients were found to have more lymph node involvement in the Siewert III [26], which was consistent with our results. Based on these ndings, the Siewert III tumors were no longer included in oesophagogastric junction adenocarcinoma and the treatment for oesophagogastric junction tumors needs update.
Total gastrectomy or more extensive distal gastric lymph node dissection may not be considered. On the other hand, the scope of proximal gastric tumors also changed, with the update of the de nition of oesophagogastric junction adenocarcinoma. The proximal gastric adenocarcinoma statistically exhibited larger tumor, advanced T-classi cation and N-classi cation than distal gastric adenocarcinoma, which was not identi ed in the V4.0. Notably, proximal gastric tumors, rather than oesophagogastric junction adenocarcinoma, may have the worse survival and need aggressive treatment. Besides, there was no signi cant difference in histologic types between gastric and oesophagogastric junction adenocarcinoma a, indicating that the tumor morphology could not be regarded as the criteria to distinguish oesophagogastric junction from gastric adenocarcinoma.
Furthermore, we also compared the clinicopathological features between the V4.0 (group 1) and V5.0 (group 2) of oesophagogastric junction adenocarcinoma. The data showed the group 2 had smaller tumor, earlier T-classi cation and N-classi cation than group 1. It seemed the group 1was less invasive. Notably, the extent of lymph node dissection in the mediastinum and the choice of distal esophagectomy made great importance in the treatment for oesophagogastric junction adenocarcinoma. A multicenter retrospective study indicated only the distance from the oesophagogastric junction was signi cantly related to metastasis. The longer the distance is, the higher rate of lymph node metastasis is [27]. In newer version, oesophagogastric junction adenocarcinoma was smaller, which supported that oesophagogastric junction adenocarcinoma was less likely to have lymph node involvement. Besides, some studies suggested lower mediastinal lymphadenectomy should be recommended for oesophageal invasion of 3 cm or less, while the extent of upper or middle mediastinal lymphadenectomy was for oesophageal invasion of ≥ 3 cm [28,29]. Although shorter in diameter, the group 2 had a higher rate than older ones in terms of the extent of oesophageal invasion of ≥ 3 cm, which showed oesophagogastric junction adenocarcinoma need upper or middle mediastinal lymphadenectomy. Therefore, mediastinal lymph node dissection and surgical resection were not completely uni ed in oesophagogastric junction adenocarcinoma, even though the scope of oesophagogastric junction adenocarcinoma became narrowed under the newer version.
With that, we compared the difference of mediastinal lymph node involvement between the two groups. The proportion of group2 (57%) were slightly higher than group 1, but with no statistical difference. This may be due to the lack of an accurate assessment of the extent of oesophageal invasion before operation and the incomplete extent of lymph node dissection.
HER2 test is another signi cant point which was formally recommended in the V5.0. Its expression and relevant clinicopathological features has beenwell studied in previous researches with Siewert's classi cation [30,31]. We re-evaluated the clinicopathological features of total 566 cases with their HER2 status, using V5.0. The HER2-expression in oesophagogastric junction and proximal tumors were statistically higher than that in body and distal tumors, which was consistent with other studies [30,32]. On the contrary, a Japanese study indicated that the HER2overexpression was not associated with tumor location with Siewert's classi cation [16]. We assessed the HER2 status with both immunohistochemistry and FISH, and classi ed the tumor's location with V5.0. Our results demonstrated oesophagogastric junction tumors had a higher expression of HER2 than body and distal tumors even if the scope of oesophagogastric junction adenocarcinoma was narrowed with the new criterion. This should be critical to emphasize the HER2 test in oesophagogastric junction adenocarcinoma. In addition, our analyses showed a statistically signi cant association between HER2-expression and pathological grade, tumor diameter and M-classi cation for gastric tumors. HER2-expression tumors had poor differentiation, larger diameter and more metastasis than HER2-negative ones, that indicated HER2expression tumors were more aggressive. These results were consistent with previous studies [33]. Therefore, the relevant clinicopathological features of HER2-expressingoesophagogastric junction adenocarcinoma remained unchanged in the V5.0. MSI was another signi cant molecular in gastric carcinoma, which was related to the contraction or expansion or of microsatellite sequences owing to the replication errors caused by mutations in the mismatch repair (MMR) in most cases [34]. More than 30% patients with MSI-H were likely to develop Lynch syndrome. Even though no statistically signi cant in difference between microsatellite status and HER2-expression, there was merely one of the 19 MSI-H cases showing positive for HER2, while other 18 cases were all negative for HER2. This demonstrated that the HER2-expression cases probably did not suffer from MSI-H, but further veri cation is required.
Our study comprehensively compared the clinicopathological features among oesophagogastric junction, proximal and distal gastric adenocarcinoma. The analyses showed that the difference was mainly between oesophagogastric junction and distal adenocarcinoma in the V4.0, while some were identi ed between proximal and distal adenocarcinoma in the V5.0. The proximal gastric tumors seems more invasive than oesophagogastric junction and distal gastric tumors in newer version. The clinical management of proximal gastric tumors may need more attention. In addition, the Siewert III tumors usually showed the worst prognosisinV4.0, which were no longer included in the category of oesophagogastric junction adenocarcinoma in the newer version. The treatment of the oesophagogastric junction tumors would not be too aggressive. However, our data showed that tumors with extent of oesophageal invasion of ≥ 3 cm were the majority in candidates with V5.0. It should be noted that these tumors require additional middle even upper mediastinal lymph node dissection and longer scope of esophagectomy. Besides, although the scope of oesophagogastric junction adenocarcinoma became narrowed after the revision, the expression of HER2 in oesophagogastric junction and proximal gastric adenocarcinoma was still higher than that in gastric body and distal site, which was basically consistent with the conclusion of the V4.0.

Conclusions
The clinicopathological parameters of the oesophagogastric junction adenocarcinoma changed in the updated WHO classi cation. Although the treatment tends to be more uni ed and standardized, the oesophagogastric junction tumors with extent of oesophageal invasion more than 3 cm required additional management. The HER2-expression of oesophagogastric junction adenocarcinoma is still higher than that of other sites of gastric in the updated version. Therefore, the emphasis on the detection of HER2 in oesophagogastric junction tumors is of great signi cance in clinical practice. The overall analyses showed it's reasonable to recommend the updated V5.0 in pathological diagnosis, as well as clinical practice.