Current and Proposed Treatment Strategies for Siewert Type II Adenocarcinoma of the Esophagogastric Junction

therapy.


Introduction
In the 1970s, squamous cell carcinoma accounted for most esophageal cancers in both Western nations and Japan, and adenocarcinoma accounted for only a small proportion. However, in Western countries, the increased incidence of Barrett esophagus, a condition that gives rise to esophageal adenocarcinoma, as well as the rising incidence of gastroesophageal re ux disease due to changes in dietary habits and obesity, led to adenocarcinoma's accounting for more than half of all cases of esophageal cancer in the 1990s [1,2]. In Japan, the decreasing rate of Helicobacter pylori infection and the rise in obesity and re ux esophagitis have led to concerns about rising rates of esophageal cancer similar to those observed in Western countries. However, the number of cases in Japan is still insu cient to clearly identify this as a trend [3,4]. The optimal treatment strategy for Siewert type II adenocarcinoma of the esophagogastric junction (AEG) is still unclear, and no consensus exists whether to treat it as an esophageal cancer or as a gastric cancer, which would include adjuvant therapy [5]. The 10th edition of the guidelines for management of esophageal cancer in Japan provides separate guidelines for AEG [6]. The frequency of this type of adenocarcinoma is expected to increase in Japan, and advanced disease has a poor prognosis [7]. Procedures for the extent of esophageal resection and mediastinal or abdominal lymph node dissection remain controversial. A thoracic approach can successfully remove the mediastinal lymph node but may markedly reduce quality of life after surgery [8]. The 5-year survival rate in patients with super cially invasive adenocarcinoma con ned to the mucosa and submucosa is unsatisfactory, and there is no consensus on whether to treat surgically in such cases [9,10]. We perform total gastrectomy for advanced AEG, but we think that minimally invasive surgery could be more appropriate for early-stage cases. However, there are few published studies of patients with early (T1) AEG.
A prospective nationwide multicenter study in Japan revealed that complete nodal clearance along the distal portion of the stomach offers marginal survival bene ts for patients with AEG less than 4 cm in diameter. The optimal extent of esophageal resection and the bene ts of mediastinal node dissection remain issues to be addressed in managing patients with esophagus-predominant AEG [11]. The purpose of this study was to determine the optimal range of resection and lymph node dissection for super cial and advanced Siewert type II AEG, considering lymph node metastases from AEG, and to develop a strategy for treatment that includes adjuvant therapy to improve the survival rate.

Patients
The patients were selected from a retrospective study by the Department of Surgery, Institute of Gastroenterological Surgery, Tokyo Women's Medical University, that was conducted from 1990 to 2019. We compared 88 cases of advanced AEG in patients who underwent surgery with lymph node dissection with 52 cases of super cial AEG, 23 of whom underwent endoscopic treatment (endoscopic mucosal resection [EMR] or endoscopic submucosal dissection [ESD]), and 29 of whom underwent surgery with lymph node dissection. All tumors ful lled the criteria of histologically proven adenocarcinoma with the center located within 1 cm of the junction on the esophageal side or 2 cm on the gastric side (Siewert type II classi cation). We retrospectively investigated the patient characteristics, pathological ndings, surgical procedures, optimum extent of lymph node dissection, location of metastatic lymph nodes, recurrence pattern, and survival curves. This study was approved by the ethics committee of Tokyo Women's Medical University in Tokyo Japan.

Classi cation of the Clinical Evaluation
The progression of the tumors was evaluated in accordance with the Union for Inter-national Cancer Control (UICC) and the American Joint Committee on Cancer (AJCC) guidelines for gastric carcinoma, 8th edition [12]. Lymph node station numbers and pathological responses after preoperative treatment were de ned in accordance with the Japanese Classi cation of Gastric Carcinoma (14th edition) and the Japanese Classi cation of Esophageal Cancer (11th edition) [13]. The depth of tumor invasion was measured as T1a (m1, carcinoma in situ; m2, invasion to the lamina propria mucosa; or m3, muscularis mucosa) or T1b (invasion to but not beyond the submucosa; divided into three equal parts: sm1, sm2, and sm3). The macroscopic types were de ned as type 0-Ip (pedunculated), 0-Is (sessile), 0-IIa (slightly elevated), 0-IIb ( at), 0-IIc (slightly depressed), or 0-III (super cial and excavated). Lymphatic vessel invasion was expressed according to the degree of invasion: ly0 (no lymphatic invasion), ly1 (focal lymphatic invasion), ly2 (moderate lymphatic invasion), or ly3 (prominent lymphatic invasion). Blood Association [14], with well and moderately differentiated tubular adenocarcinoma and papillary adenocarcinoma classi ed as differentiated-type carcinomas, and poorly differentiated adenocarcinoma, signet ring cell carcinoma, and mucinous carcinoma classi ed as undifferentiated-type carcinomas. The Japanese Esophageal Society subclassi cation of Nishi's proposal has ve types, depending on the relative extent of esophageal (E) or gastric (G) involvement (E, EG, E=G, GE, and G)

Priority of Nodal Dissection
To evaluate which nodes, have priority for dissection, we adopted a method using the therapeutic value index introduced by Sasako et al [15]. The index is calculated by multi-plying the frequency of metastasis to the station by the 3-or 5-year survival rate of patients with metastasis to that station. Nodes were a priority for dissection if the value of the index in the station had a metastatic incidence exceeding 5%.

Statistical Analyses
All data are presented as the medians and total ranges unless otherwise stated. The survival was calculated by the Kaplan-Meier method and then compared between groups with the log-rank test. The hazard ratio of the survival was calculated with Cox proportional hazard analyses. p < 0.05 was statistically signi cant.

Patient Characteristics
Of the 140 total patients, 52 had super cial AEG and 88 had advanced AEG ( was sm2 in 4 patients and sm3 in 1 patient who had upper thoracic and lower thoracic lymph node metastases. Five patients in whom the tumor invaded to the submucosa received chemoradiotherapy or surgery as additional treatment. In super cial AEG, lymphatic vessel invasion was more common than vascular invasion. The proportion with undifferentiated histological type was higher in patients with advanced cancer than in those with super cial cancer. Most patients (90% with super cial AEG and 70% with advanced AEG) underwent lower esophagectomy + proximal gastrectomy. The surgical approach was transhiatal in 69% of those with super cial AEG and left thoracotomy in 56% of those with advanced AEG. Re-construction was performed with gastric tube in 55% of patients with super cial AEG and in 73% of patients with advanced AEG. Double-tract reconstruction was performed in 45% of cases of super cial AEG (Table2). Roux-Y 0 (0%) 6 (7%) Colon 0 (0%) 4 (4%) Abbreviations: EMR-endoscopic mucosal resection, ESD-endoscopic submucosal dissection.

Lymph Node Metastasis
The calculated index for each nodal station in patients with advanced AEG was in stations 1, 2, 3, 7, 11 and 110 for 3-year survival and in stations 1, 2, 3 and 7 for 5-year survival (Table 3). We did not calculate the index for cervical, paraaortic lymph nodes and paraesophageal nodes at the esophageal hiatus of the diaphragm because only a few patients underwent dissection of these stations. The location of affected lymph nodes varied with the extent of esophageal involvement as well as with cancer stage (Table 4).

Survival
The median follow-up was 1476 days (range 125-4549). The 5-year overall survival rates were 88% for patients treated with ESD, 78% for those with super cial AEG who under-went surgery, and 24% for those with advanced AEG (p = 0.011; Fig. 1). Only two patients with super cial AEG experienced lymph node metastasis after surgery, but 13 patients with advanced AEG had metastases to thoracic lymph nodes and 12 had metastases to abdominal lymph nodes after surgery. Despite of lymph node dissection, twenty-ve patients experienced lymph node metastasis after operation in advanced AEG and there were many disseminations in advanced AEG (Table5).

Neoadjuvant Chemoradiotherapy
There were no differences in survival between patients who received postoperative adjuvant therapy with S-1 for advanced AEG and those who received surgery alone (p = 0.5192; Fig.2). We selected new neoadjuvant therapy for the following reasons. 1. The effect of the S-1 as a postoperative adjuvant therapy was insu cient (Figure 2). 2. It seems that radiation therapy was more effective because there was dissemination after operation as a pattern of recurrence (Table 5). 3. The nab-paclitaxel was reported in a gastric cancer preclinical model with subcutaneous and peritoneal xenografts, comparing with paclitaxel [17][18][19]. We have started a Phase I trial of nab-paclitaxel combined with radiotherapy for advanced AEG (registration number UMIN000024088) to compare the e cacy of low-dose nab-paclitaxel with the current standard dose (260 mg/m2) of nab-paclitaxel as rst-or second-line chemotherapy for advanced AEG. The radiotherapy targets lymph node stations 110, 1, 2, 3, 7 and 11. Enrollment of patients was started at level 1, and the three patients at this starting dose level experienced no dose limiting toxicity (DLT). Enrollment at level 2 was started in accordance with the speci ed procedure for progression to the next level. At level 2, one of three enrolled patients experienced treatment-related adverse events corresponding to DLT. These treatment-related adverse events were grade 1 peripheral neuropathy. Enrollment at level 3 was started in accordance with the speci ed procedure for progression to the next level. These treatment-related adverse events were grade 3 peripheral neuropathy and grade 4 Febrile neutropenia. An additional three patients were enrolled at level 2, but none experienced treatmentrelated adverse events corresponding to DLT, thus leading to the recommended dose was level 2 (Fig.3).
In this study 14 patients have performed nab-paclitaxel combined with radiotherapy as neoadjuvant therapy for advanced AEG. Tumor grade after treatment for evaluation of therapeutic effectiveness would be achieving a grade> 1b was considered evidence of effectiveness. When therapeutic e cacy was evaluated, 13 patients in 14 were at Grade 1b or higher (table.6), but there were no differences in OS between the group that received neoadjuvant chemoradiotherapy and the group that received preoperative chemotherapy with S-1 (p = 0.5908; Fig.4), because the patients who received neoadjuvant chemoradiotherapy was too much small numbers. Because of the nature of a phase I trial and the small number of patients enrolled, it is not appropriate to draw meaningful conclusions concerning overall response rate (ORR) or survival.

Discussion
There are some differences in standard treatment for AEG between European and Japanese guidelines. In Western countries, the de nition of the esophagogastric junction (EGJ) is the oral end of the fold on the greater curvature of the stomach, according to the Prague classi cation. In Japan, the de nition is the lower end of the esophageal palisade vessels. For carcinoma of the esophagogastric junction, in Western countries the Siewert classi cation is that type I tumors (lower-esophageal adenocarcinoma) are located 1-5 cm above the EGJ, irrespective of EGJ involvement; type II tumors (cardia adenocarcinoma) are located between 1 cm above and 2 cm below the EGJ; and type III tumors (subcardial gastric adenocarcinoma) are located 2-5 cm below the EGJ with involvement of the EGJ and distal esophagus [20]. In Japan, the Japanese Classi cation of Esophageal Cancer and Gastric Cancer uses Nishi's classi cation to de ne ve types of EGJ cancer characterized by diameters of ≤40 mm and an epicenter within 2 cm proximal or distal to the EGJ, irrespective of histological type [21]. The location of the center of the tumor is about the same in Siewert type II and the Nishi classi cation. In histology, the Nishi classi cation includes squamous carcinoma, but the Siewert classi cation does not.
For Siewert type II AEG, or similar tumors classi ed as Nishi EG, E=G, or GE, the type of surgical resection and reconstruction should be individualized [22]. Two randomized controlled trials have compared surgical procedures for EGJ cancer. One was a trial in the Netherlands comparing right thoracotomy versus a transhiatal approach; the survival of patients with Siewert type II did not differ by approach, but respiratory complications were signi cantly higher in the thoracotomy group [23]. The other trial, in Japan, obtained the same results as the Dutch trial [24]. Therefore, in Japan we perform low esophagectomy and total gastrectomy through a transhiatal approach for Siewert type II AEG, as in Western countries.
The distance from the EGJ to the proximal edge of the primary tumor may be a signi cant indicator of metastasis or recurrence in the mediastinal lymph nodes in patients with Siewert type II tumors. Thorough mediastinal lymph node dissection via a transthoracic approach may provide a therapeutic bene t when the distance is >3 cm [25]. In Japan, a prospective nationwide multicenter study determined that dissecting only station 110 was su cient in cases of an EGJ tumor unless esophageal involvement , and lower mediastinal lymph nodes, regardless of the histology [11,27].
Although surgical procedures of super cial and locally advanced AEG are standardized, the role of adjuvant therapy for AEG is still controversial. In Japan, according to the results of the ACTS-GC trial in gastric cancer [28], we performed chemotherapy with S-1 as a postoperative adjuvant therapy. Recently, some institutes have begun using S-1 + docetaxcel or S-1 + oxaliplatin, based on the results of the START-2 trial [29,30]. Two trials in Europe have provided evidence for adjuvant therapy. The CROSS trial from the Netherlands showed that preoperative chemoradiotherapy improved survival among patients with potentially curable EGJ cancer. The regimen was associated with acceptable ad-verse-event rates [31].
The FLOT4 trial from Germany showed that overall survival was higher in the uorouracil, leucovorin, oxaliplatin, docetaxel (FLOT) group than in the epirubicin and cisplatin plus either 5-FU or capecitabine (ECF/ECX) group [32]. Based on these Western trials, preoperative chemoradiotherapy with carboplatin + docetaxel and as a neoadjuvant and adjuvant chemotherapy with uorouracil + leucovorin, oxaliplatin, and docetaxel is now standard therapy. We have introduced preoperative chemoradiotherapy at our institute because dissemination after surgery was a frequent and recurrent pattern. According to the report from Western population, Peritoneal or pleural recurrence rate were 5-15% [33,34]. In Japan including our report these rates were more than 30%. This data is exactly the basis for adjuvant chemoradiotherapy. In a phase 1 study, the primary objective was to determine the maximum tolerated dose (MTD) and recommended dose (RD) of nab-paclitaxel plus radiotherapy. The study was designed in accordance with a standard 3 + 3 methods. This study was approved by the ethics commit-tee of Tokyo Women's Medical University in Tokyo Japan. We select nab-paclitaxel as a drug with a radiation expecting a high response. Nab-paclitaxel show higher response than solvent-base paclitaxel in two Phase III trial. One trial in advanced non-small-cell lung cancer compared nab-paclitaxel plus carboplatin with solvent-based paclitaxel plus carboplatin. Response rate for nab-paclitaxel group is 35%, and solvent-base paclitaxel is 27% [35]. Second trial in previously treated advanced gastric cancer compared nab-paclitaxel with solvent-based paclitaxel. Response rate for nab-paclitaxel alone is 33%, and solventbase paclitaxel alone is 24% [36]. Furthermore, some data suggest that nab-paclitaxel is more effective than solvent-base paclitaxel in patients with peritoneal metastasis. In a subgroup analysis of Phase III trial in gastric cancer, overall survival, and progression-free survival in patients with peritoneal metastasis were better with nab-paclitaxel than with solvent-based paclitaxel. And a previous preclinical study showed that intravenous nab-paclitaxel is more effective than solvent-based paclitaxel in a gastric cancer model with subcutaneous and peritoneal xenografts [19]. For these rea-sons, we select nab-paclitaxel.
The RD was determined as level 2 (nab-paclitaxel 150 mg/m2 day1 and day 21 + radiotherapy 39.6 Gy). The results of this study are being sub-mitted to the journal and have not yet been published. The longterm safety and e cacy of this combination therapy will be evaluated in phase II trial. The preoperative chemoradiotherapy for AEG is slightly not popular in Japan because there were many complications after operation to receive radiation. The recommendation of combined taxane based chemotherapy and radiotherapy for advanced stage cases is not supported by the presented data.

Conclusion
In this retrospective study, we found that the optimal lymph node resection for advanced AEG consists of lymph nodes of the inferior mediastinum (No. 110), the lesser curvature (Nos. 1, 3, 7), No. 2, and No 11.
Lymphadenectomy of lymph node stations 1, 3, 7, and 11 was su cient to improve survival of patients with super cial AEG except in case of thoracic lymph node metastasis can be con rmed before surgery.
When esophageal involvement was > 40 mm, we performed esophagectomy including upper thoracic lymph nodes in advanced AEG. Because the rate of metastasis to the thoracic lymph nodes is low in super cial AEG, a surgical procedure that removes the upper edge of the tumor should be selected. We were not satis ed with the result of surgical treatment including lymph node dissection alone in advanced AEG. We will recommend nab-paclitaxel combined with radiotherapy for advanced AEG to improve the survival rate.  Figure 1 Kaplan-Meier survival curves comparing ESD or EMR, super cial AEG and advanced AEG.

Figure 2
Page 23/23 Five-year overall survival of patients receiving postoperative chemotherapy with S-1 for AEG or receiving sur-gery (OP) alone.

Figure 3
Phase I trial of Nab-Paclitaxel combined with radiotherapy for advanced AEG Figure 4 Overall survival of patients receiving neoadjuvant chemoradiotherapy or postoperative chemotherapy with S-1.