Distribution of Lymph Node Metastases in Esophageal Carcinoma Patients Undergoing Upfront Surgery: A Systematic Review

Metastatic lymphatic mapping in esophageal cancer is important to determine the optimal extent of the radiation field in case of neoadjuvant chemoradiotherapy and lymphadenectomy when esophagectomy is indicated. The objective of this review is to identify the distribution pattern of metastatic lymphatic spread in relation to histology, tumor location, and T-stage in patients with esophageal cancer. Embase and Medline databases were searched by two independent researchers. Studies were included if published before July 2019 and if a transthoracic esophagectomy with a complete 2- or 3-field lymphadenectomy was performed without neoadjuvant therapy. The prevalence of lymph node metastases was described per histologic subtype and primary tumor location. Fourteen studies were included in this review with a total of 8952 patients. We found that both squamous cell carcinoma and adenocarcinoma metastasize to cervical, thoracic, and abdominal lymph node stations, regardless of the primary tumor location. In patients with an upper, middle, and lower thoracic squamous cell carcinoma, the lymph nodes along the right recurrent nerve are often affected (34%, 24% and 10%, respectively). Few studies describe the metastatic pattern of adenocarcinoma. The current literature is heterogeneous in the classification and reporting of lymph node metastases. This complicates evidence-based strategies in neoadjuvant and surgical treatment.


Introduction
Esophageal cancer patients often present with an advanced disease stage, encompassing metastatic lymph nodes or distant metastases. The presence and number of lymph node metastases are among the most important prognostic factors in esophageal carcinoma and are independent predictors for long-term survival [1][2][3][4][5][6].
The location of metastatic lymph nodes depends on tumor histology, primary tumor location, T-stage and neo-adjuvant therapy [7]. The vessels in the dense lymphatic network surrounding the esophagus are complexly aligned and they contribute to a multidirectional spread of lymph node metastases in the abdomen, the mediastinum, and the neck [8,9]. Additionally, 'skip metastasis', skipping the first and directly metastasizing into the second or third lymph node echelons, are frequently seen in both esophageal adenocarcinoma and squamous cell carcinoma [10,11]. This contributes to the presence of lymph node metastases at unexpected distant sites, which makes standardization of the extent of the radiation field and lymphadenectomy in the treatment of esophageal cancer difficult. Not surprisingly, the optimal extent of the lymphadenectomy in esophagectomy has been subject of a global

Data Extraction
The primary endpoint is the metastatic rate of lymph node metastases in esophageal carcinoma per lymph node station or region. When available, the following variables were extracted from the included studies: year of publication; country; study design; inclusion period; lymph node classification system used (JES, Japan Esophageal Society; AJCC, American Joint Committee on Cancer; other; none); description of how detailed lymph node regions or stations are described and reported; number of patients; patient characteristics (gender, age); tumor histology; tumor location (upper thoracic esophagus, middle thoracic esophagus, lower thoracic esophagus or gastroesophageal junction (GEJ)); c/pT-stage; c/pN-stage; type of surgical approach; lymphadenectomy (complete 2-or 3-field); use of immunohistochemistry staining; prevalence of lymph node metastases per lymph node station; number of patients with lymph node metastases; overall percentage of positive lymph nodes; number of patients per lymph node location with resected nodes and positive nodes in that station or region. Methodological quality was assessed using the Methodological Index for Non-Randomized Studies (MINORS) checklist [21].

Statistical Analysis
Descriptive statistics summarized the characteristics of included studies, patient characteristics, and the outcomes of each included study. Since not all studies used the same classification or definition for lymph node stations, and some studies only reported lymph node data for different regions instead of stations, we combined the two mostly used systems (JES and AJCC) and grouped lymph node stations into five regions: cervical, upper mediastinal, middle mediastinal, lower mediastinal, and abdominal (Table 1). Studies with a reported number of patients with metastatic lymph nodes per station were pooled separately from the studies only describing data per region. The prevalence of patients with lymph node metastases per station or region were calculated by summarizing all the patients with lymph node metastases per lymph node station or region and dividing them by the sum of all patients who had a lymph node dissection in this station or region. Results were stratified for tumor histology and primary tumor location. Finally, studies describing the lymphatic distribution pattern according to the pT-stage were pooled and described separately.

Results
Details of the literature search and study selection are shown in Figure 1. Fourteen studies met the inclusion criteria and were included in this review [7,[22][23][24][25][26][27][28][29][30][31][32][33][34]. Characteristics of the included studies are presented in Table 2. A total of 8952 patients were evaluated, including 409 (5%) with an adenocarcinoma and 8543 (95%) with a squamous cell carcinoma. Among all patients with a squamous cell carcinoma, 726 (9%) patients had a tumor located in the upper thoracic esophagus, 5130 (60%) patients had a tumor in the middle thoracic esophagus and 2687 (31%) had a tumor in the lower thoracic esophagus. None of these studies described patients with a cervical or GEJ squamous cell carcinoma. For adenocarcinoma, 32 (8%) tumors were located in the distal esophagus and 377 (92%) were located at the GEJ. The c/pT stage varied among studies, but most of the patients had a c/pT3 tumor. Details of the study populations can be found in Tables 2 and 3. adenocarcinoma and 8543 (95%) with a squamous cell carcinoma. Among all patients with a squamous cell carcinoma, 726 (9%) patients had a tumor located in the upper thoracic esophagus, 5130 (60%) patients had a tumor in the middle thoracic esophagus and 2687 (31%) had a tumor in the lower thoracic esophagus. None of these studies described patients with a cervical or GEJ squamous cell carcinoma. For adenocarcinoma, 32 (8%) tumors were located in the distal esophagus and 377 (92%) were located at the GEJ. The c/pT stage varied among studies, but most of the patients had a c/pT3 tumor. Details of the study populations can be found in Tables 2 and 3.

Reporting Standard
Standard reported lymphadenectomy differed among studies. In 1 study, patients underwent a 2-field lymphadenectomy, in 9 studies, patients underwent a 3-field lymphadenectomy, and in 4 studies, both procedures were included, resulting in a different nodal yield per study. In addition, the definition of anatomical locations of lymph node stations differed amongst studies; 1 study used AJCC, 5 used JES, and 8 did not use a standard classification system. Moreover, 6 studies described the prevalence of lymph node metastases per lymph node station, 5 only described the prevalence of lymph node metastases per region, and 3 reported a combination of both. The reported regions and stations also varied among studies (Table 2). Some studies described for some stations both sides together, and other studies separated left and right. One study combined tumor locations when reporting the number of patients per lymph node station and could therefore not be pooled with the others studies [7].

Distribution Pattern for Esophageal Squamous Cell Carcinoma
Eleven studies [7,22,24,[26][27][28][29][30][31][32]34] described the location of lymph node metastases in patients with a squamous cell carcinoma (n = 8543). Table 4 shows the prevalence of lymph node metastases per lymph node station among the seven studies [22,26,[28][29][30][31]34] that reported data per lymph node station per tumor location. For patients with an upper thoracic tumor, lymph node metastases are most frequently seen along the right recurrent nerve (60%) and cervical paraesophageal lymph nodes (right 34% and left 22%). For patients with a middle thoracic tumor, the prevalence of lymph node metastases was highest along the right recurrent nerve (23%), right cervical paraesophageal lymph nodes (24%), and middle thoracic paraesophageal lymph nodes (23%). The lymph nodes along the left gastric artery (28%) and lower thoracic esophagus (23%) had the highest prevalence of lymph node metastases in patients with a tumor in the lower thoracic esophagus. Six studies [24,[26][27][28]30,32] described the location of the lymph node metastases per region. The results of these studies are shown in Figure 2.
Two studies [23,25] described the location of lymph node metastases in regions. Pooled numbers show that for patients with a GEJ tumor, 20% (4 out of 10) had lymph node metastases in the cervical region and 25% (31 out of 124) had metastases in the abdominal lymph node stations (other regions were not reported). For patients with an adenocarcinoma of the lower thoracic esophagus, 35% (6 out of 17) had lymph node metastases in the cervical region, 71% (12 out of 17) had lymph node metastases in the lower mediastinal region, and 71% (12 out of 17) had lymph node metastases in the abdominal region.
One study [7] combined patients with a tumor of the distal esophagus and GEJ. In this study, a prevalence of 30% in the periesophageal lymph nodes, 37% in the paracardial lymph nodes, 35% in the perigastric lymph nodes, and 14% in the celiac axis was reported.

Distribution of LN Metastases in Relation to pT-Stage
Three studies [24,27,30] stratified the prevalence of nodal metastases per pT-stage. All three studies described patients with squamous cell carcinoma. Table 5 shows the rate of patients with lymph node metastases per region, divided into four groups: patients with pT1-2 and pT3-4 and patients with pT1 and pT2-4. Among patients with higher T-stages, a higher prevalence of lymph node metastases is seen per region, while the distribution remained similar.

Discussion
This study describes the sites of lymph node metastases in esophageal cancer patients according to the histology and primary tumor location based on literature published before July 2019. This is the first study systematically combining available evidence on lymph node metastases pattern, contributing toward revealing the lymphatic metastatic pattern of esophageal carcinoma. This study showed that both squamous cell carcinoma and adenocarcinoma metastasize to cervical, thoracic, and abdominal lymph node stations, regardless of the location of the primary tumor.

Lymphatic Distribution Pattern for Squamous Cell Carcinoma
Multiple studies have attempted to define the lymph node metastases pattern in esophageal cancer [7,22,23,[25][26][27][30][31][32][33]. Most of these studies describe squamous cell carcinoma only. The available evidence of lymph node metastases pattern in adenocarcinoma of the esophagus is scarce, and the available literature for both tumor types is very heterogeneous. Whereas some studies report data per lymph node station, others report per region. To make this even more complex, not all studies adhere to the same boundaries of lymph node regions and not all studies use the same anatomical definition of lymph node stations. To define anatomical sites on lymph node stations, some use standardized classification systems such as the AJCC or JES, while others do not use any standardized classification system. Moreover, not all studies report the exact extent of lymphadenectomy. In addition, some studies excluded from this review combined patients with adenocarcinoma and squamous cell carcinoma and/or different tumor locations, which makes data hard to interpret, since these factors could influence the distribution pattern [35][36][37] All together, these factors make comparing available evidence on the lymph node metastases distribution in esophageal cancer difficult.
For upper, middle, as well as lower thoracic esophageal squamous cell carcinoma, the stations around the esophagus are among those with the highest prevalence of lymph node metastases. However, not only lymph node stations in the same region as the tumor have a high prevalence of nodal metastases; stations in different regions are affected as well. For example, 13% of the patients with a lower esophageal tumor have right cervical lymph node metastases.

Lymphatic Distribution Pattern for Adenocarcinoma
For adenocarcinoma, although data are more limited, similar results are seen. One-quarter (25%) of patients with a GEJ adenocarcinoma had middle thoracic paraesopahgeal lymph node metastases, and when looking at zones, 20% of the patients with a GEJ adenocarcinoma had lymph node metastases in the cervical zone. An explanation could be the presence of an extensive lymphatic network in the submucosa and even in the lamina propria of the esophagus, with both intramural and longitudinal lymphatic drainage. The longitudinal nature of this network explains the variation in anatomic sites of lymph node metastases [38][39][40]. Another result of the complexity of the lymphatic vessel system is the phenomenon of skip metastases [37]. Skip metastases are distant lymph nodes with metastatic involvement, without tumor infiltration in the regional lymph nodes, and they are more often seen in early tumors [10]. It is unclear what the exact clinical value is since the literature is conflicting on the prognostic relevance [41][42][43].
Three studies [24,27,30] described lymph node metastases per lymph node station in relation to pT-stage in patients with squamous cell carcinoma. An increased prevalence of lymph node metastases is seen per region in patients with a higher pT-stage, while the distribution remained similar. These are small numbers; nevertheless, the literature points out that a higher T-stage is associated with more lymph node metastases [44].
It should be pointed out that this study defines the lymph node metastases pattern based on patients without neoadjuvant treatment because lymph node involvement may differ after neoadjuvant chemo(radio)therapy [45]. Currently, neoadjuvant chemoradiation or perioperative or neoadjuvant chemotherapy is the standard of care in most countries. This makes our results less applicable to current surgical patients, and one of the main questions for the future is whether the lymphadenectomy strategy should be based on the pattern of lymph node metastases before neoadjuvant treatment or after neoadjuvant treatment. However, the location of lymph node metastases in untreated esophageal cancer patients tells us more about the behavior of the disease, and this is fundamental, as this allows for accurately defining neoadjuvant treatment strategies by targeting high-risk regions for lymph node metastases in patients with specific characteristics. A recent study showed that after neoadjuvant chemoradiotherapy, almost half of the patients in that cohort had lymph node metastases outside the radiation field, indicating that the current radiation fields are not sufficient [46]. Although, it should be noted that radiotherapy to an elective nodal area (both metastatic and non-metastatic) does not guarantee a better outcome [47].
There are some limitations of the present study. Firstly, as previously mentioned, studies were very heterogeneous in lymph node dissection and the reporting of anatomical sites of nodal metastases. This heterogeneity might have made our pooled results less reliable. Moreover, not all studies could be pooled per station (since they only described lymph node regions) and vice versa. The inclusion of different studies in Tables 2 and 4 makes displayed percentages slightly different. In addition, few of the studies subdivided patients for T-stage and location of the primary tumor, whilst it has been proven that these factors influence lymph node metastases [26,48,49].
If we want to determine the exact distribution pattern of esophageal cancer, large well-designed prospective studies are needed. One initiative of such a study is the multinational prospective TIGER study (ClinicalTrials.gov Identifier: NCT03222895) [50].

Conclusions
Both esophageal squamous cell carcinoma and adenocarcinoma are aggressive diseases that can metastasize to cervical, thoracic, as well as abdominal lymph node stations, regardless of the location of the primary tumor. The prevalence of patients with metastatic lymph nodes per station and region could be determined for squamous cell carcinoma. However, few studies described the distribution of lymph node metastases for esophageal adenocarcinoma, and the data for both tumor types was very heterogeneous. This complicates evidence-based treatment strategies in both neoadjuvant (radiation field) and surgical (lymphadenectomy) treatment. Well-designed prospective studies are needed to determine the exact lymphatic distribution pattern of esophageal cancer.
Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.

# Searches Results
1 exp esophagus tumor/or esophagus resection/or esophagectom*.ti,ab,kw. or ((esophagus or esophageal or esophagogastric or oesophagus or oesophageal or oesophagogastric or gastroesophag* or gastrooesophag*) and (neoplas* or cancer* or carcino* or adenocarcino* or tumor or tumors or tumour or tumours or malig*)).ti,ab,kw. 127708 2 (exp lymph node/or exp lymph node metastasis/or ((lymph or lymphatic) and (node* or nodal or metastas*)).ti,ab,kw.) and (cancer staging/or (staging or TNM or number or extent or extended or scoring or score or classif* or categor* or criteria or 2-field* or two-field* or 3-field* or three-field* or node status or nodal status or D1 or D2 or N0 or N1 or N2 or N3 or pattern* or drainage or spread or pathway* or depth).ti,ab,kw.) Table A2. Cont.