Nodal skip metastasis in thoracic esophageal squamous cell carcinoma: a cohort study

Esophageal cancer is a very aggressive malignancy, with poor prognosis in resected patients. Although the incidence of esophageal adenocarcinoma (EAC) has been increasing in Western Countries [1], esophageal squamous cell carcinoma (ESCC) is still the predominant esophageal malignancy in Eastern Countries, like Japan and China [2, 3].

The most important factor affecting the prognosis of ESCC patients is lymph node (LN) involvement, which is included in the American Joint Committee of Cancer (AJCC) TNM classification [4]. The role of lymph node metastases (LNMs) on prognosis has been widely investigated, from the simple involvement to the number of involved nodes [5‐7]. The ratio of LNMs on harvested nodes has also been evaluated to take into account the variability of lymph node dissection, but its role is still unclear [8]. The 7th edition of the AJCC staging system took into account all these results on LN involvement and it increased the classes of N-stage according to the number of LNMs [4]. Apart from the number, the localization of LNMs could also play a role in ESCC prognosis. Since the esophagus is an organ that passes through three main anatomic regions (neck, chest, abdomen), the prognosis of patients with the same number of positive LNs might be different due to localization of LNMs in one or more anatomic regions [9].

Nodal skip metastasis (NSM) is a particular pattern of LNMs, which involves the LNs distant from the tumor site but not the peritumoral LNs. This pattern has been evaluated to be relevant in other kinds of malignancies [10, 11] but its role in esophageal cancer patients is still controversial [12]. The aim of this study was to investigate the factors predicting NSM and to assess its effect on survival and recurrence in thoracic ESCC.

NSM is not an uncommon form of metastatic spread to lymph nodes and has been found to be of clinical importance in some sites of malignancies, i.e. non-small cell lung cancer or thyroid cancer [10, 11]. The presence of NSM has also been evaluated in esophageal cancer but its prognostic role is still unclear and requires further investigation [12, 14‐18]. Recent studies on this topic are summarized in Table 3. The rate of NSM in esophageal cancer varied widely across the studies (from 20.3% to 76.3%, Table 3), increasing the uncertainty about this type of metastatic spread. Such variability could be explained by the different criteria of lymph node classification criteria that were used by the authors. Briefly, NSM has been defined as a) metastatic involvement of distant LNs with peritumoral LNs free of tumor infiltration [12, 16, 18,], b) cervical and/or abdominal involvement but no mediastinal metastasis [14, 15, 19], or c) metastatic infiltration of N2 through N4 lymph nodes but not of N1 nodes according to the Japanese staging system of the Japanese Society for Esophageal Disease [15, 17]. The first definition (metastatic involvement of distant LNs with peritumoral LNs free of tumor infiltration) was used in the present study because we believed it would afford a stricter definition of skip lymph nodes. It is noteworthy that Wu et al. [15] used different lymph node classification criteria to evaluate the same sample and reported very different rates of NSM (24.2% and 69.7%). In addition, the different node dissection (both 2-field and 3-field node dissections have been reported) [19] and the inclusion of both EAC and ESCC patients in 2 studies could have contributed to the variability of the reported rate of NSM (Table 3).

Predictors of NSM in esophageal cancer has been investigated in three previous studies [12, 17, 18] that reported tumor location as the main factor associated to NSM (Table 3). In our series, the number of NSM patients did not allow any meaningful multivariable analysis of risk factors for NSM. However, univariate analysis suggested a higher NSM occurrence in upper and middle thoracic esophagus. In literature, tumors located in the middle thoracic esophagus were identified as risk factor for NSM, despite slight differences among the studies regarding the comparison of tumor sites (Table 3). The association between NSM and middle thoracic esophageal tumors could be explained by its anatomic location (which allows both upper and lower spread directions) and by the definition itself of NSM (cervical and abdominal lymph nodes are possible NSMs), which increase the likelihood of finding NSMs in tumors located in middle thoracic esophagus [20‐22].

In our series, patients with NSM showed similar overall survival and disease free survival compared to those with peritumoral LNM, adjusting for the number of LNMs. Similar findings on overall survival in ESCC patients were shown by 3 previous Chinese studies [12, 15, 16] while Prenzel et al. [17] reported a favorable prognosis associated with the presence of NSM in a heterogeneous group of both ESCC and EAC patients. It was surprising that NSM (distant metastases in absence of local metastases) did not affect overall survival. Our hypothesis is that patient immune response to cancer cells cleared the peritumoral metastases while some surviving clones proceeded to further nodal stations. Therefore, NSM might represent a final escape step of ESCC progression [23]. In addition, previous data on favorable prognosis of NSM in ESCC might have been biased by the lower number of LNMs in NSM patients. We think that these considerations might be extended to disease free survival, but the literature did not provide useful data on the association between NSM and disease free survival in ESCC.

Our series included over 50% clinical N0 stages, thus the role of radical lymphadenectomy is further enhanced, even in case of apparent N0 stage at clinical evaluation. Moreover, NSM was less likely identified than local LN involvement during preoperative evaluation (cN0 rate 81.2% vs. 51.4%). These findings support the potential usefulness of innovative techniques such as sentinel node assessment in order to intraoperatively identify the pattern of LN involvement [24].

The present study has some limitations. First, it is a retrospective study on a single-institution series. Second, the number of patients may have prevented some findings from being extrapolated. Third, the quality of lymphadenectomy could have affected the results. However, all esophagectomies were performed by the same surgical team; − thus the warranted homogeneity of the surgical approach - and the number of harvested nodes were acceptable. In fact, 2-field lymphadenectomy failed to identify LNM in only 3 patients (2.1%) who had cervical nodal metastasis and had been staged N0 at pathologic evaluation upon final specimen.

NSM is a common pattern of metastatic LN involvement in thoracic ESCC, but it does not affect the prognosis. The heterogeneity of the studies on NSM in literature requires further evaluation in order to investigate this lymph node metastatic spread.