Abstract
Background
Lymph node metastasis from esophagogastric junction (EGJ) cancer is multi-directional, and the previously reported incidences in each nodal station are heavily biased by different operative approaches. We conducted a multicenter prospective study using a standardized surgical protocol of lymphadenectomy.
Methods
This study is a single-arm, multi-institutional, prospective study. Patients with resectable cT2–4 EGJ cancer were eligible, regardless of neoadjuvant therapy. The primary endpoint was the metastasis rate of each lymph nodal station. The planned sample size was 360, and the entry period was 4 years. Protocol of surgical procedure: Patients diagnosed as having EGJ adenocarcinoma with esophageal invasion > 3 cm or with upper/middle mediastinal lymph node involvement, or those having EGJ squamous cell carcinoma, underwent an esophagectomy via right thoracotomy or thoracoscopy. The lymphadenectomy included the upper/middle/lower mediastinum, paracardiac region, around the celiac axis, and the left renal vein area (16a2lat). Patients with EGJ adenocarcinoma with esophageal invasion ≤ 3 cm underwent transhiatal lower esophagectomy and gastrectomy with lymphadenectomy of the lower mediastinum, paracardiac region, around the celiac axis, and 16a2lat.
Results
From April 2014 to September 2017, 371 patients were enrolled in this study from 42 institutions. As 8 patients were excluded due to protocol violations and later refusals, 363 patients were analyzed finally. Of 363 patients, 332 (91.4%) had adenocarcinoma and 31 (8.5%) had SCC. 86 patients (23.7%) had cT2 tumors and the other 277 (76.3%) had cT3–4 tumors. 134 patients (36.9%) had cN0 and the other 229 (63.1%) had cN + disease. Neo-adjuvant treatments (mostly chemotherapy) were given to only 121 patients (33.3%). Esophagectomies were performed in 121 patients (33.3%) and extended gastrectomies were performed in 236 (65.0%). Resection with R0 was achieved in 339 (93.4%). Based on histological examinations, the incidences of nodal involvements in each area were as follows: upper mediastinum (9/118, 7.6%); middle mediastinum (14/128, 10.9%); lower mediastinum (47/353, 13.3%); abdominal (240/358, 67.0%); the left renal vein area (16A2lat) (18/344, 5.2%).
Conclusion
For patients with EGJ cancers, the incidences of nodal involvements in the upper mediastinum and the left renal vein area (16A2lat) were lower than our expectations.
Disclosure
All authors have declared no conflicts of interest.
