Prognostic Impact of Pathological Adverse Features in Surgically Resected Adenocarcinoma of Esophagogastric Junction

Objective The prognostic value of lymphovascular invasion (LVI), perineural invasion (PNI), and poor differentiation (PD) has been widely studied in different solid tumors. However, it was still controversial in adenocarcinoma of esophagogastric junction (AEG). We investigated the prognostic impact of combining LVI, PNI and PD for predicting the survival in patients with AEG. Methods We retrospectively investigated the data of patients who performed surgical resection of AEG on Guangdong Provincial Hospital and Guangdong Provincial Hospital of Chinese Medicine from Jan. 2004 to Dec. 2018. According to the status of LVI, PNI and differentiation, pathological adverse features were divided into three groups: 0, 1 or 2 and 3 adverse features, their impact on prognosis was evaluated. Results Univariate analysis indicated pT, pN, LVI , PNI , PD and pathological adverse features were risk factors for both overall survival (OS) and disease-specic survival (DSS), and multivariate analysis indicated that pathological adverse feature was independent risk factor for both OS and DSS. In subgroup analyses, adverse features were independent risk factor for DSS of stage II AEG but not for stage I or III. Conclusions The pathological adverse features were independent prognostic factors for AEG patients and they can help for further risk stratication in stage II patients. research. features can help for better risk stratication across different subtypes of AEG.


Introduction
With the gradual increase in incidence, adenocarcinoma of esophagogastric junction (AEG) has attracted attention worldwide and gradually become a separate entity for treatment and research. Accurately staging not only help for predicting patients' survival and is of great signi cance for individualized and precise treatment. Up to now, staging system for AEG is still under debate. This study aimed to explore the impact of the pathological factors, including lymphovascular invasion (LVI), perineurial invasion (PNI), and poor differentiation (PD) on the prognosis of surgically resected AEG, investigating whether they could help for better risk strati cation.

Patients
We retrospectively collected data of AEG patients who underwent surgery at Guangdong Provincial Hospital and Guangdong Provincial Hospital of Chinese Medicine from January 2004 to December 2018. Included criteria: (1) The tumor invaded the dentate line, and tumor center was located at the area 5 cm above or below the dentate line; (2)

Data
The clinicopathological data were extracted from the electronic patient management system, including age, sex, Siewert subtype, LVI, PNI, degree of differentiation, pathological T stage, pathological N stage. Tissue slides of each patient were rstly examined by the junior pathologist and the reports were recon rmed by at least one senior pathologist. The tumors were classi ed by the Siewert classi cation, based on the patient's gastroscopy, computed tomography, contrast radiography and postoperative specimens. Pathological T staging and N staging were based on the TNM staging standards of the AJCC and UICC 8th edition.
Siewert type I and type II refer to stages of esophageal cancer, and Siewert type III refers to stages of gastric cancer. LVI, PNI, PD were named as pathological adverse features, each patient was assigned to 0, 1 or 2, or 3 adverse features group based on the pathology reports.  Abbreviations: PNI = Perineural invasion. LVI = Lymphovascular invasion. PD = Poor differentiation; NPD = Non poor differentiation.

Pathological Data
The relation between tumor invasion (pT), lymphatic node metastasis (pN), LVI, PNI, and degree of differentiation were shown in Table 2. LVI, PNI and PD were all signi cantly correlated with higher pT stage and pN stage. Patients with LVI signi cantly had higher percentage of PNI or poor differentiation, vice versa.

Survival Analyses
Univariate analysis indicated pT, pN, LVI, PNI, PD, pathological adverse features were risk factors for both overall survival (OS) and disease-speci c survival (DSS), while multivariate analysis indicated pathological adverse features were independent risk factor for both OS and DSS (  (Fig. 2). In subgroup analysis, pathological adverse features were independent risk factor of DSS in stage II AEG. The median DSS age were not reach in 0 and 1-2 adverse features, and 19(13.0-25.0) months in 3 adverse features group. The 5-year DSS rate was 81.3%, 65%, 16.7%, respectively (Fig. 3).

Discussion
Due to the special anatomical site, controversies exist in many aspects of adenocarcinoma of esophagogastric junction (AEG), such as the pathogenesis, the origin of tumor cells, and surgical strategy [1][2][3]. Siewert et al. [4] de ned AEG as adenocarcinoma invading the dentate line within 5cm crossing the gastroesophageal junction, which played an important role guiding clinical treatment and research.
Con icts still exist on how to stage AEG more accurately. The 8th edition AJCC/UICC considered Siewert III AEG as gastric cancer and Siewert I and II remain to esophageal cancer entity, changing from previous edition that all three subtypes staging as esophageal cancer. Better risk strati cation and more accurate assessment of prognosis would be of great importance for individualized and precise management of AEG. Our research aims to investigate the impact of pathological risk factors on AEG prognosis and whether they could help for better risk strati cation.
LVI, PNI, and PD were priori of malignant biological behavior of many solid tumors, indicating worse prognosis [5][6][7][8][9]. NCCN guidelines had also pointed out that patients with early stage esophageal or gastric cancer receiving endoscopic treatment, when pathological risk factors exist, more comprehensive systemic treatment and careful follow-up should be considered [10,11]. In terms of AEG, con icts still existed on whether the pathological adverse features would independently impact prognosis [12][13][14] like esophageal cancer or gastric cancer [15][16][17]. As preoperative chemoradiotherapy or chemotherapy has become the standard procedure for advanced AEG [24,25], recent study showed that when AEG has more pathological adverse features, they had worse response to neoadjuvant therapy but bene t more in postoperative treatment [26,27]. However, as D2 gastrectomy followed by adjuvant chemotherapy was the standard procedure for gastric cancer in Asia, most of the patients in earlier year did not receive preoperative treatment in our study.
Whether pathological adverse features could predict the tumor response to neoadjuvant therapy in patients needed further investigation.
Our study had some limitations. It contained patients only from two institute. Since neural invasion was not standardize including in the pathology report until 2008 in our center, we exclude the patients without neural invasion record. Thus this retrospective study with few cases and a selection bias. Because of the uneven distribution of cases, subgroup analysis for different stage patients may have been affected by the small number of cases in some subgroups.
In summary, pathological adverse features were independent prognostic risk factor for AEG. In stage II AEG, it helped to further stratify the disease-speci c survival rate. Pathological adverse features can help for better risk strati cation across different subtypes of AEG.

Declarations
Funding This work was not supported by any funding sources.
Competing interests The authors declare that they have no competing interests.
Availability of data and materials The datasets used or analysed during the current study are available from the corresponding author on reasonable request.