Pattern of Locoregional Recurrence After Radical Surgery in Patients With Esophageal Squamous Cell Carcinoma and The Design of Target Volume of Postoperative Prophylactic Radiotherapy

Objective The study was undertaken to evaluate preferred sites for locoregional recurrence after radical surgery for patients with esophageal squamous cell carcinoma (ESCC) and to conrm the target volume of postoperative prophylactic radiotherapy. Methods 139 patients with locoregional recurrent ESCC after radical esophagectomy without postoperative radiotherapy were registered in this study. The sites of locoregional recurrence for these patients were collected and inuence factors of locoregional recurrence were analyzed. Results The 1, 2, and 3-year progression-free survival rates were 48.2%, 18.0% and 8.6%, respectively. Mediastinum lymph node (LN) recurrence (74.2%) was the most frequent site of recurrence, following the anastomotic site (28.1%), supraclavicular LN (19.4%) and abdominal LN (15.1%) (P=0.000). The upper mediastinum (72.7%) was one of the most common recurrence occurred in mediastinal LN. Compared with upper and middle segment of ESCC, lower segment had the highest recurrence rate of upper abdominal LN (P=0.001). The predictive factors of anastomotic recurrence were the stage of pT3 or pT4, presence of nerve or vessel invasion, removed LN NO. ≤ 17, presence of invasion or adhesion and without postoperative adjuvant chemotherapy. In addition, the risk factors of abdominal LN recurrence in patients with middle segment ESCC included the stage of pT3 or pT4, smoking history and without postoperative adjuvant chemotherapy. Conclusion For thoracic ESCC, supraclavicular, upper mediastinum, subcarinal LNs and anastomosis should be incorporated within target volume of postoperative prophylactic radiotherapy. Regarding to lower segment ESCC, target volume should include upper abdominal LNs. And it need cautious evaluated about upper abdominal LNs when presence of clinicopathologic factor in middle segment ESCC.


Introduction
Surgical resection is the primary therapy for thoracic esophageal cancers. However, the rich lymphatic capillary network in the esophageal mucosa and submucosa facilitate locoregional recurrences after surgery. It has been reported that 25.6-41.8% of patients develop locoregional recurrence [1,2,3]. The suppression of postoperative recurrences will ameliorate the survival of patients with ESCC after curative resection. Thus, it is of clinical important to understand the patterns and prognostic factors of recurrence after esophagectomy. Rates and patterns of postoperation lcoregional recurrence have been described in previous systematic studies [2,3,4]. Nevertheless, to our knowledge, in the most of reports, the classi cation standard of mediastinal LN sites in lung cancer was common used in patients with ESCC [2,5]. For esophageal cancer, it has extensive difference of the characteristics of the LN recurrence from lung cancer. In this study, we adopted the LN sites classi cation standard of ESCC according to the Japan Esophageal Society which showed more elaborate and highlighted LN metastatic features [6].
Several reports had demonstrated that postoperative prophylactic radiotherapy with chemotherapy could improve prognosis for patients with stage IIb-III ESCC after radical esophagectomy [7,8]. Thus, knowledge of locoregional recurrence sites provides useful information on the effectiveness of treatment and may allow better planning of therapeutic intervention in an effort to improve outcomes. In this study, we retrospectively analyzed the frequent position based on classi cation standard according to the handbook of the Japan Esophageal Society in patients with postoperative locoregional recurrence ESCC and tried to con rm the target volume of postoperative prophylactic radiotherapy.

Methods
From January 2015 to January 2019, 139 patients with locoregional recurrent ESCC after radical esophagectomy without postoperative radiotherapy were registered in this study. Tumor stage classi ed according to the eighth edition of the TNM staging system of the American Joint Committee on Cancer for esophageal cancer [ 9 ]. The study protocol was approved by independent ethics committees in Jinling Hospital and each patient signed a written informed consent.
The classi cation standard of mediastinal LN sites were based the handbook of Japan Esophageal Society 6 and the radiographic anatomical delineation according to the detailed explain by Huang et al. Disease-free survival (DFS) was de ned as the time between the diagnosis of original cancer of esophagus and the diagnosis of locoregional recurrence. SPSS 24.0 statistical software was used for data analysis. A chisquare test was used for statistical data comparison. The Kaplan-Meier method was adopted to calculate the survival rate. P-values lower than 0.05 were considered statistically signi cant.

Patient characteristics
The clinical characteristics of the enrolled patients are listed in Table 1

Sites of Locoregional Recurrence
The 1, 2, and 3-year DFS rates were 48.2%, 18.0% and 8.6%, respectively ( Figure 1). The positions of locoregional recurrence after esophagectomy are showed in Table 2. Mediastinum lymph node (LN) recurrence (74.2%) was the most frequent site of recurrence, following the anastomotic site (28.1%), supraclavicular LN (19.4%) and abdominal LN (15.1%), which showed statistically difference (P=0.000). The upper mediastinum (72.7%) was one of the most common recurrence occurred in mediastinal LN, followed by the middle (19.4%) and lower mediastinum (4.3%) (P=0.000). The recurrent rates of No. 106pre, 106recR, 106recL and 106tbL in the upper mediastinum were exceptionally higher than the rest stations, which were 43.2%, 30.2%, 24.5%, 19.4% and 15.8%, respectively (P=0.000). LNs recurrence of the middle mediastinum mainly occurred in No. 107 (15.8%) followed by No. 108 (3.6%) (P=0.001). The recurrent LNs in the lower mediastinum discovered rarely and the recurrent rate of No.112 was 4.3%. The recurrence rates between the left and right supraclavicular LNs showed no statistically difference for ESCC (P=0.711). Table 3, for different segments of ESCC, the sites of No.106pre, 106recR, 106recL and 106tbL lymph nodes all showed the relatively higher recurrence rates. Compared with lower and middle segments ESCC, upper segment represents the highest recurrent rate in the site of No.106tbR (P=0.029).

As demonstrated in
The recurrent rates of upper abdominal LNs for upper, middle, and lower segments ESCC were 0%, 9.9%, and 29.2%, respectively, with statistically signi cant differences (P= 0.001). However, the recurrent rates of the supraclavicular LNs, mediastinal LNs and anastomosis site demonstrated no signi cantly difference in three segments of thoracic ESCC, respectively. (P= 0.964, P= 0.766 and P=0.676).

Relations between clinicopathologic factors and locoregional recurrence
The relations between clinicopathologic factors and anastomotic recurrence were displayed in Table 4. The predictive factors of anastomotic recurrence were patients with the stage of pT3 or pT4 of original neoplams during surgery, presence of nerve or vessel invasion, removed LN NO.≤ 17, presence of invasion or adhesion or without postoperative adjuvant chemotherapy after radical surgery. The collections between clinicopathologic factors and supraclvicular LNs recurrence were showed in Table 5. The results revealed that there was no relationship between esch clinicopathologic factor and supraclvicular LNs recurrence. For patients with middle segment ESCC, the high risk factors of upper abdominal LN recurrence included the stage of pT3 or pT4 of original neoplams during surgery, smoking history and without postoperative adjuvant chemotherapy (Table 6).

Disscussion
Locoregional recurrences may occur along the entire "esophageal bed" from the cervical lymph node, anastomostic site, and the mediastinum to intra-abdominal lymph nodes with different frequencies according to the primary localization of tumor [14]. The results of this study showed that the rencurrent rates of upper mediastinal LNs and subcarinal LNs, especially in site No. 106pre (43.2%), 106recR  [5] reported that in the 126 patients with locoregional LNs recurrence of ESCC, the mediastinal LN recurrent rate was signi cantly higher compared with the rate of supraclavicular and upper abdominal LNs, and the upper (73.8%) mediastinum LNs had the signi cantly recurrent tendency, which was consistent with the results in our study. The potential reason was that there were intricate anatomical structures and large blood vessels and nerves were plenty and may increase the di culty of completely lymphadenectomy in the upper mediastinum. Second, the esophagus has a well known complex submucosal lymphatic drainage system which may facilitate spread of tumor cells via sub-mucosal lymphatics.
In our study, there was no statistically difference in the rates of supraclavicular, mediastinal and anastomosis recurrence among upper, middle and lower ESCC, which was corresponding with the results noti ed in other report [5]. Furthermore, the recurrent positions of lower ESCC were also gathered at the supraclavicular and upper mediastinal lymph nodes, especially at stations No. 106pre(47.9%), 106recR(35.4%), 106recL(33.3%), 106tbL(22.9%), 104L (16.7%) and 101R(16.7%), similar results were demonstrated in other reports [5,15]. Therefore, supraclavicular, upper mediastinum LNs and subcarinal LNs should be included in the target volume of postoperative prophylactic radiotherapy of all segments thoracic ESCC.
Anastomosis recurrence was common in our sample (28.1%), which was consistent with other previous study. Mandard et al.[ 1 6] found residual tumors at autopsy in the esophageal stumps of 27% of patients after surgery for carcinoma of esophagus. Yu et al. [17] in a retrospective study indicated that anastomosis recurrence rates were 29% in post-operative esophageal cancer patients received adjuvant combined modality therapy consisting of four cycles of uorouracil-based chemotherapy and locoregional radiatherapy without coverage of the anastomotic site. The recurrent mass usually arises from an esophagogastrostomy site, and leads to luminal narrowing. However, in reported cases recurrence at the anastomotic site is not common pattern [3]. Wang et al. [ 1 8] found that the recurrence rate of anastomotic site was 9.8% for esophageal squamous cell carcinoma, and they recommend that anastomotic site should be excluded from the target volumes of post-operative radiatherapy. The possible reason was that more than half original neoplasma of esophagus locates at the upper and middle mediastinum in this study and increase the surgical operation inconvenient. The predictive factors of anastomosis recurrence were patients with the stage of pT3 or pT4 of original neoplams during surgery, presence of nerve or vessel invasion, removed LN NO.≤ 17, presence of invasion or adhesion or without postoperative adjuvant chemotherapy after radical surgery. Based on the results of our study, anastomosis site should be covered in the target volume of postoperative prophylactic radiotherapy of all segments thoracic ESCC.
According to our study, the recurrence rates of upper abdominal lymph node vary greatly based on location of ESCC and lower segment ESCC showed more frequently recurrent rates of lymph node, consistent with other previous study [ 1  There was no relationship between clinicopathologic factors and supraclvicular LNs recurrence. In our study, tumor bed recurrence was rarely discovered, probably by reason of tumor bed adjacent to thoracic vertebra and may be confused with middle (No. 108, 3.6%) and lower mediatinum LNs (No.112, 4.3%). Therefore, tumor bed should be exclude the target volume of postoperative prophylactic radiotherapy.