Prognostic Signicance of the Proximal Margin for Esophagogastric Junction Adenocarcinoma With Type II and III Tumors After Surgery


 BackgroundThe incidence of esophagogastric junction adenocarcinoma (EJA) has increased in recent years, with surgical resection the main choice of treatment. The optimal length of the proximal margin for EJA is still under debate, and the impact of EJA survival and recurrence remains unclear. The aim of the present study was to investigate the influence of the optimal length of the proximal margin on EJA.MethodsFrom January 2011 to December 2015, 131 patients who had EJA with type II tumors were included and retrospectively analyzed. All patients underwent radical R0 resection. The proximal margin was measured promptly after resection, and the frozen-section pathological examination was negative for the margin.ResultsThere were 3 cases of Siewert type I EJA (2.9 %), 75 cases of Siewert type II EJA (57.9 %), and 53 cases (40.1%) of Siewert type III EJA. The median number of lymph nodes examined was 19 (range: 1–41), and the median number of positive lymph nodes was 2 (range: 0–18). Sixty-three patients underwent total gastrectomy (48.1%), and 68 underwent proximal gastrectomy (51.9%). The median follow-up time was 57.3 months: (range 1.9–174.1); 34 patients (26%) relapsed and 74 (56.5%) died. The 5-year overall survival rate of type II tumor patients was 68.2%, and that of type III tumor patients was 38.5% (P = 0.02). For patients with a proximal margin <2 cm, the median recurrence time was 41.6 months, whereas it was for 42.8 months for patients with proximal margin >2 cm (log–rank: 0.496). Our data analysis found that a proximal margin length of 2 cm was a prognostic variable for type II and type III tumors.ConclusionsThere are a number of factors associated with recurrence and overall survival at 5 years for patients who have EJA with type II and type III tumors, and a proximal margin >2 cm may indicate better prognosis.


Abstract Background
The incidence of esophagogastric junction adenocarcinoma (EJA) has increased in recent years, with surgical resection the main choice of treatment. The optimal length of the proximal margin for EJA is still under debate, and the impact of EJA survival and recurrence remains unclear. The aim of the present study was to investigate the in uence of the optimal length of the proximal margin on EJA.

Methods
From January 2011 to December 2015, 131 patients who had EJA with type II tumors were included and retrospectively analyzed. All patients underwent radical R0 resection. The proximal margin was measured promptly after resection, and the frozen-section pathological examination was negative for the margin.
For patients with a proximal margin <2 cm, the median recurrence time was 41.6 months, whereas it was for 42.8 months for patients with proximal margin >2 cm (log-rank: 0.496). Our data analysis found that a proximal margin length of 2 cm was a prognostic variable for type II and type III tumors.

Conclusions
There are a number of factors associated with recurrence and overall survival at 5 years for patients who have EJA with type II and type III tumors, and a proximal margin >2 cm may indicate better prognosis.

Background
The incidence of esophagogastric junction adenocarcinoma (EJA) has increased in recent years, particularly in Western and Asian countries. [1][2][3] According to data from Japan, the incidence of EJA has increased by 7.3% from the 1960s to the beginning of the 21 st century. 4 A single-center registration study of gastric cancer in China found that the proportion of esophageal-gastric junction (EGJ) cancer increased from 22.3% to 35.7% between 1988 and 2012. 5 According to the Siewert classi cation, there are three types of EJA: Siewert type I is de ned as tumors located 1-5 cm above the esophagogastric junction, Siewert type II tumors are located at the upper l-2 cm below the esophagogastric junction, and Siewert type III tumors are located 2-5 cm below the esophagogastric junction. 6 According to the AJCC Cancer Staging Manual, 8th edition, EJA is categorized and staged as esophageal cancer, as long as the tumor center is within 2 cm of the junction, regardless of whether it invades the esophagus. If it is not within 2 cm of the junction, the tumor is grouped and treated as stomach cancer, even if it has invaded the EGJ. 7 Currently, surgical resection is regarded as the cornerstone of curative treatment, although the introduction of neoadjuvant/adjuvant chemotherapy and radiotherapy and chemotherapy have been found to improve disease prognosis. 8 Due to the complexities of EJA tumor location, a consensus has yet to be reached on the best surgical strategy. The appropriate resection range of the esophagus and stomach, the scope and location of lymph node resection, and the best surgical method are still unclear. 9 The optimal length of the proximal margin for EJA is also still under debate, with only a limited number of studies published on this. In their study, Barbour et al. showed that the proximal margin length might be associated with patient survival in type II-IV tumors, but not in type I tumors, and they found that if the length of the proximal margin was >3.8 cm, then the prognosis of type II+ tumor patients could be signi cantly improved. 10 However, Mine et al. found that, for patients who have EJA with type II or type III tumors, the length of the proximal margin exceeds 2 cm, which seems to be satisfactory. 11 The proximal margin length is key for R0 and R1 resection status, and thus for survival outcome. 12,13 Therefore, it is crucial to determine a safe operation range when performing surgery.

Methods
We conducted a retrospective, observational study. Based on the classi cation of the AJCC Cancer Staging Manual, 8th edition, patients diagnosed with EJA type II and type III tumors, treated with surgery between January 2011 and December 2015, were included in the present study, except patients with gastric cancer and/or those undergoing neoadjuvant therapy. The inclusion criteria were: (i) patients undergoing radical surgery, including radical proximal or total gastrectomy; (ii) a negative con rmation of the proximal margin; and (iii) type II and type III tumors without distant metastasis. Patients with incomplete medical information, those undergoing neoadjuvant therapy, those with malignant tumors in other locations, and those who had previously had exploratory or tumor-reduction surgeries were excluded. Fresh specimens were cut longitudinally immediately after resection. The sample was then stretched to the maximum extent and xed to a plate. The surgeon then measured and recorded the length of the proximal edge. The proximal margin was sent for frozen-section pathological examination to con rm whether the proximal margin length was su cient. If insu cient, further resection was performed until there was a negative con rmation of the proximal margin. All surgical procedures and the extent of lymph node clearance conformed to the Japanese Gastric Cancer Treatment Guidelines (Japanese Gastric Cancer Association 2011). All procedures were conducted in accordance with the Declaration of Helsinki, as revised in 2013 14 .
Data on age, sex, Siewert type, extent of surgery, tumor size, proximal margin, T stage, clinical stage, lymphatic-vascular invasion, neural invasion, differentiation status, total lymph nodes, lymph node metastasis, mediastinal lymph node dissection, Lauren type, human epidermal growth factor 2 status, adjuvant therapy, and relapse or recovery were collected. All patients underwent enhanced chest and abdominal computed tomography every 6 months after discharge to evaluate tumor recurrence and distant metastasis until October 2015. Follow-up was generally conducted through outpatient visits, email, and telephone interviews, and follow-up data were updated until November 1, 2015. The follow-up rate, median follow-up time (months), and overall survival results were included in the study. The main reason that patients could not be followed up was because they declined outpatient visits or changed their telephone numbers and addresses.
All variables were analyzed using descriptive statistics. The results are presented as percentages, means, and dispersion measures. We used the unadjusted Kaplan-Meier method for visualization of the survival curves, and the log-rank test to compare survival curves using SPSS version 22.0. Logistic regression analysis was used for survival identi ed by univariate analysis were further assessed by multivariate analysis. The P-value was considered to be statistically signi cant at the 5% level. To better de ne the surgical margin, we use the receiver-operator curve (ROC). Based on the Declaration of Helsinki and the general research health law. 14 Informed consent was not required for the present study, and patient con dentiality was assured.
We present the following article in accordance with the STROBE reporting checklist.

Results
In total, 168 patients diagnosed with EJA without neoadjuvant chemotherapy were included in the present study, according to our admission and discharge criteria. Thirty-seven patients were excluded; 4 had undergone exploration or tumor-reduction surgery, 10 were diagnosed with T1 tumors according to the nal pathological report, and 23 had incomplete information ( Figure 1) Table 1). The impact of the proximal margin length on overall survival was analyzed. A proximal margin length of 2 cm was found to be a prognostic variable for patients with type II and type III tumors in both the univariate (Kaplan-Meier method, P = 0.02) and multivariable analyses (hazard ratio: 2.00, 95% con dence interval: 0.54 -7.41, P = 0.25, Cox method) (Tables 2 and 3).
The length of the proximal edge in the surgical specimens and its relationship with recurrence and the 5- year overall survival rate, which is sorted by centimeters (from nearest to farthest), is shown in Figures 1  and 2.
In the univariate analysis, patients with poorly differentiated tumors (P = 0.005), late clinical stage (P = 0·008), and neural invasion (P = 0.044) were at higher risk of recurrence. Patients with higher pT category tumors (P = 0·028), more lymph node metastasis (P = 0·013), poorly differentiated tumors (P = 0·005), proximal margins of <2 cm (P = 0·019), late clinical stage (P = 0.012), and postsurgical recurrence had signi cantly worse survival ( In the multivariate analysis, only tumor differentiation status was found to be the prognostic factor for recurrence. The variables related to the 5-year overall survival rate were tumor in ltration (type II vs. III), proximal margin, tumor differentiation status, and recurrence. However, in the ROC analysis, we did not nd that a de nitive margin showed better tumor outcomes (Tables 3 and 4). Barbour et al. found that proximal margin length was a more signi cant prognostic factor in types II-IV tumors with N0-2 (P<0.01). However, the same cannot be said for patients with types II-IV N3 tumors (P = 0.48). A proximal length of 3.8 cm in resected specimens is considered an independent prognostic factor according to analyses limited to R0 or R1 resection. The proximal margin length was considered a prognostic factor between the esophagectomy group (5 cm) and the gastrectomy group (2 cm) and in uenced the survival of patients with Siewert type I tumors. 10  We observed a 26% (34/131) recurrence rate, and tumor differentiation status was found to be the only signi cant prognostic factor of recurrence. Most recurrences were distant (23/34, 67.6%), and localregional recurrences were relatively lower (11/34, 32.4%). According to Patrão et al., tumor differentiation status, pT stage were the strongest prognostic factors for poor outcome and relapse. 22 In their study, the relapse rate was 61% (108/177), with only 9 (8.3%) isolated cases of local-regional relapse with symptoms, whereas the majority of cases (99/108, 91.7%) presented with distant metastasis. In their study, Suh et al. had a recurrence rate of 30% after excluding type I cancers; distant metastasis was found to be more prevalent (14%), and only 4.1-0.6% of patients had local-regional recurrence 23 . This could be due to only suspicious clinical or laboratory ndings undergoing more advanced imaging examinations, such as computed tomography, magnetic resonance imaging, and gastrointestinal endoscopy, and thus many asymptomatic local recurrences are missed due to a lack of timely imaging examinations. Therefore, we may be underestimated and cannot reliably describe the overall local recurrence rate. Almost all recurrences are only diagnosed when there are symptoms, which could be at a relatively late stage, and could explain why the incidence of isolated local metastases was relatively lower. These results suggest that effective systemic treatment is important, and to a large extent, represents the unmet needs of EJA. If we consider that our study only covers type II and type III tumor patients during the 5-year follow-up period, our data are different from those reported in the literature to some extent.
Our specimens were cut longitudinally and lymph nodes were removed for pathological examination, stretched to the maximum extent, and placed on plates. The total length of the proximal edge was determined by vision and touch, and was measured and recorded by the surgeon. If insu cient and additional distal esophagus was removed, we measured the total length of the proximal edge. Because of shrinkage of the specimens, these measured lengths did not true re ect the corresponding in situ lengths before the conclusion of the operation. In 1986, Siu et al. found that esophageal specimens shrunk to approximately half their length after resection, and the upper margin was reduced to a greater extent than the lower edge (44% vs 54% of in situ length, respectively) after resection and before fxation. 18 Thus, based on their ndings, a proximal margin of 2 cm would be 4 cm, and the cited minimal proximal margin ranging between 2 and 5 cm would be between 4 and 10 cm in situ.

Limitations
The present retrospective study had several limitations. First, we did not have accurate measurements of the proximal margin length due to shrinkage of the esophagus after resection, and due to the difference between observers; therefore, the lack of a centralized examination of pathological specimens may have led to deviation in the results. Second, the sample size used in the present study was small and was limited to a single institution; thus, more prospective studies are needed to verify our ndings in the Chinese population. Finally, we did not evaluate the risk factors for the positive proximal margin, nor did we evaluate the effect of neoadjuvant therapy on the state of the proximal margin after resection.

Conclusion
There are a number of factors associated with recurrence and overall survival at 5 years for patients who have EJA with type II and type III tumors, and a proximal margin of >2 cm may indicate a better prognosis.