Introduction

Esophagogastric junction cancer (EGC) invades the esophagogastric junction and involves the anatomic border between the proximal stomach and esophagus, with an increasing incidence worldwide [1, 2]. The number and location of metastatic lymph nodes (LNs) are considered independent prognostic predictors in EGC patients [3,4,5]. However, the extent of LN dissection remains controversial. Wang et al. [6] showed that the dissection of parapyloric LN (No. 5 and 6 LNs, PLN) brought a survival benefit to Siewert type II/III EGC patients. Jun Peng et al. [7] recommended that at least 12 LNs should be examined to ensure accurate staging for Siewert type II EGC patients. In contrast, no significant survival differences were observed between patients with loco LN dissection alone and those with extended lymphadenectomy (including station 8/11) [8]. Different ethnicities might be the cause of these differences.

Due to strong invasive properties, a large fraction of EGC patients is detected at a locally advanced stage, and neoadjuvant therapy, including radiotherapy and chemotherapy, is required for most of them [9,10,11]. Accumulating evidence has shown that neoadjuvant therapy can effectively improve the clinical outcomes of EGC patients such as large clinical trials CheckMate 577 [12], PRODIGY [13], and CROSS [14]. Combined neoadjuvant immunotherapy and chemoradiotherapy have also demonstrated favorable efficacy and safety [15]. However, whether neoadjuvant therapy affects the number of dissected LNs remains unclear. A single-center, retrospective study has shown that neoadjuvant therapy decreased the number of dissected LNs in esophageal cancer patients [16]. Nevertheless, another study based on the Surveillance, Epidemiology, and End Results (SEER) database argued that there was no significant difference in the number of dissected LNs between esophageal cancer patients with preoperative radiotherapy and those without [17]. These inconsistent results imply that more studies are needed. Besides, these studies mainly focused on esophageal cancer rather than EGC. The impact of neoadjuvant therapy on the number of dissected LNs is yet to be fully evaluated in EGC patients. Therefore, in the present study, we explored whether neoadjuvant therapy impacted the number of LN examination and evaluated the optimal number of LN examination using the SEER database.

Methods

Patients

We retrieved patients from the SEER database (2006–2017). Patients between 2004 and 2005 were not included in the study because of the missing chemotherapy data in the SEER database. The inclusion criteria were as follows: (1) positive histology; (2) underwent radical surgery; (3) definite tumor stage (T-category) and nodal stage (N-category), according to the 8th edition of the American Joint Committee on Cancer (AJCC) criteria; (4) first primary tumor; (5) at least one regional lymph node dissection based on pathologic evidence; (6) treatment with/without preoperative radiotherapy and/or chemotherapy; and (7) no distant metastasis. The exclusion criteria were: (1) under 18 years old; (2) unavailable follow-up data; (3) unknown cause of death. We retrieved baseline characteristics, including the year of diagnosis, race, age, sex, histology, grade, marital status, T-category and N-category. The data from all subjects from the SEER database was obtained legally.

Statistical analysis

Differences between patients with neoadjuvant therapy and those without were analyzed by χ2 and t-tests. Univariate and multivariate analysis were performed with a Cox proportional hazards regression model to explore prognostic factors for EGC patients with neoadjuvant therapy. The optimal number of LN examination was explored using X-tile software (Yale University, Version 3.6.1). The overall survival (OS) was analyzed with a Kaplan–Meier analysis. Data were analyzed with PASW Statistics 18. A two-sided p-value < 0.05 was considered to indicate a significant difference.

Results

Baseline characteristics

We enrolled 4028 EGC patients, including 2020 without neoadjuvant therapy, 498 with neoadjuvant chemotherapy, 39 with neoadjuvant radiotherapy, and 1471 with neoadjuvant chemoradiotherapy. The baseline characteristics are summarized in Table 1. The median age of the entire cohort was 64 years (range: 20–94 years), and more patients received neoadjuvant therapy during 2012–2017 than in 2006–2011. Males and adenocarcinoma accounted for the majority in the cohort. The highest proportion of T-category was T1 and T2 among patients without neoadjuvant therapy, while T2 and T3 were predominant among patients with neoadjuvant therapy. The highest proportion of N-category was N0 among patients without neoadjuvant therapy, while N1 was predominant among patients with neoadjuvant therapy. These results indicated that patients with higher tumor stages tended to receive neoadjuvant therapy. The survival analysis showed that patients with neoadjuvant therapy had a worse prognosis than those without neoadjuvant therapy (Fig. 1, P < 0.05).

Table 1 Baseline characteristics of EGC patients
Full size table
Fig. 1
figure 1

Survival curves of the entire cohort according to the treatment modality

Full size image

The mean number of dissected LNs was 17.5 among patients without neoadjuvant therapy and 17.3 among patients with neoadjuvant therapy (P > 0.05). In detail, patients with neoadjuvant chemotherapy had the highest mean number of dissected LNs (21.0, P < 0.001), while those with neoadjuvant radiotherapy had the lowest mean number of dissected LNs (12.2, P = 0.003). The mean LN number of patients with neoadjuvant chemoradiotherapy was 16.3, which was still significantly lower than 17.5 for patients without neoadjuvant therapy (P = 0.001).

Optimal number of LN examination

We determined the optimal cutoff value for the number of dissected LNs using X-tile analysis. The optimal cutoff value was 11 for patients with neoadjuvant radiotherapy. However, the survival of patients with 1–11 LNs and those with > 11 LNs did not differ (Fig. 2A, P > 0.05). This negative result might be related to the small sample size (39 cases). For patients with neoadjuvant chemotherapy, the optimal cutoff value was 19. Patients with > 19 LNs had a better prognosis than those with 1–19 LNs (Fig. 2B, P < 0.05). For patients with neoadjuvant chemoradiotherapy, the optimal cutoff value was 9. Patients with > 9 LNs had a better prognosis than those with 1–9 LNs (Fig. 2C, P < 0.05). Therefore, we divided the entire cohort into three groups according to the optimal number of LN examination: 1–9, 10–19, and > 19 LNs.

Fig. 2
figure 2

The optimal number of LN examination calculated using X-tile. a patients with neoadjuvant radiotherapy; b patients with neoadjuvant chemotherapy; c patients with neoadjuvant chemoradiotherapy

Full size image

COX regression analyses

Furthermore, we performed univariate and multivariate COX regression analyses for patients with neoadjuvant chemotherapy and chemoradiotherapy. We did not include patients with neoadjuvant radiotherapy due to the low sample size. The univariate analysis demonstrated that grade (well/moderately differentiated vs. poorly differentiated/undifferentiated), T-category, and N-category were prognostic factors for patients with neoadjuvant chemotherapy (Table 2). The multivariate analysis demonstrated that the T-category (T1 vs. T3), N-category, and the number of LN examination (1–9 vs. > 19) were independent prognostic factors for them (Table 2). The univariate analysis demonstrated that age, sex, grade (well/moderately differentiated vs. poorly differentiated/undifferentiated), marital status, T-category, N-category, and the number of LN examination were prognostic factors for patients with neoadjuvant chemoradiotherapy (Table 3). Further multivariate analysis demonstrated that age, sex, grade (well/moderately differentiated vs. poorly differentiated/undifferentiated), marital status, N-category, and the number of LN examination were independent prognostic factors for them (Table 3).

Table 2 Univariate and multivariate analyses for OS in EGC patients with neoadjuvant chemotherapy
Full size table
Table 3 Univariate and multivariate analyses for OS in EGC patients with neoadjuvant chemoradiotherapy
Full size table

Discussion

EGC is a lethal disease with an increasing incidence and a poor prognosis [18]. Considering the location and histological characteristics, EGC does not completely resemble esophageal or gastric cancer, raising debates on surgical margins [19]. Therefore, distinguishing EGC as a specific type of malignant tumor of the digestive tract has become crucial to provide insights into its clinical properties. LN metastases occur in more than 20% of T1-category patients and are considered a negative prognostic factor [20, 21]. Therefore, lymphadenectomy is a critical part of surgical treatment.

Herein, we evaluated the impact of neoadjuvant therapy on the number of dissected LNs in EGC patients using the SEER database. We found that neoadjuvant radiotherapy and chemoradiotherapy decreased the number of dissected LNs in EGC patients, while neoadjuvant chemotherapy increased it. Our results were consistent with previous studies regarding esophageal cancer in which radiotherapy/chemoradiotherapy would cause a decrease in LN number [22,23,24]. These results might be related to tumor and nodal down-staging, which also has been proved in other cancers, including rectal and non-small cell lung cancer [25, 26]. However, the impact of neoadjuvant chemotherapy alone on the number of dissected LNs in EGC or esophageal cancer has not been reported. In breast cancer, Boughey et al. [27] found that neoadjuvant chemotherapy increased the number of axillary LN dissection (21.9 vs. 20.2) although the difference did not achieve statistical significance. Additionally, other studies have shown that neoadjuvant chemotherapy is associated with a lower axillary LN count in breast cancer [28, 29]. These inconsistent results indicated that the number of LN dissection might not be strongly associated with chemotherapy. One possible explanation is that patients with neoadjuvant chemotherapy might have more severe nodal diseases. Based on our data from Table 1, N2 and N3 categories accounted for 21.3% of neoadjuvant chemotherapy while 10.6% in patients with neoadjuvant chemoradiotherapy. Moreover, the optimal cutoff value of LN dissection was 19 for patients with neoadjuvant chemotherapy, much higher than neoadjuvant chemoradiotherapy. These findings indicated that patients with neoadjuvant chemotherapy alone might undergo appropriate or even radical procedures to obtain more LNs.

Numerous data, including National Comprehensive Cancer Network (NCCN) guidelines, have indicated that at least 15 LNs should be dissected for patients without neoadjuvant therapy to guarantee accurate staging and survival benefits [23, 30,31,32]. However, the optimal number of LN dissection for patients with neoadjuvant chemoradiotherapy remains controversial. Samson et al. [23] found that removing 10–15 LNs brought additional survival benefits for esophageal cancer patients with neoadjuvant chemoradiotherapy. Another institutional analysis showed that esophageal cancer patients with > 7 LN examination had a better prognosis than those with 1–7 LNs [33]. Unfortunately, no relevant literature about the impact of neoadjuvant chemoradiotherapy on the LN count has been reported in EGC patients. In the present study, we recommended that at least 10 LNs be removed since the greatest survival difference was reached at the cut-off point (1–9 LNs vs. > 9 LNs). Different from patients with neoadjuvant chemotherapy alone, surgeons should avoid an extremely radical operation as much as possible for patients with neoadjuvant chemoradiotherapy. The removal of more LNs might increase the risk of complications without adding survival benefits.

However, our current study also has some limitations. First, this was a retrospective study with inherent limitations such as selection bias. Hence, a large prospective clinical trial is needed to validate our conclusions. Second, detailed chemoradiotherapy information was missing, including the type and dose of radiotherapy and chemotherapy, delivery methods, and duration of treatment, which might have some impact on the conclusions. Third, the SEER database has no quality-of-life information, which can not be ignored while treating malignant patients.

Conclusion

Neoadjuvant radiotherapy and chemoradiotherapy decreased the number of dissected LNs, while neoadjuvant chemotherapy increased it in EGC patients. Hence, at least 10 LNs should be dissected for neoadjuvant chemoradiotherapy and 20 for neoadjuvant chemotherapy, which could be applied in clinical practice.