BACKGROUND: Adenocarcinoma of the esophago-gastric junction (EGJ) potentially spreads to abdominal and mediastinal lymph nodes. METHODS: Eighty-five patients with type I and II EGJ cancer underwent curative esophagectomy or esophago-gastrectomy and radical abdominal and mediastinal lymphadenectomy. Microinvolvement was detected with the mAb Ber-Ep4 in all histopathologically free lymph nodes. RESULTS: In type I tumors (n = 40), lower mediastinal lymph nodes were positive in 24% and among type II tumors (n = 45) in 10% of patients. Ber-Ep4+ cells in apparently free lymph nodes were found in 49% of patients. On inclusion of Ber-Ep4+ nodes, positive mediastinal lymph node staging was rising to 40 and 33% in type I and II patients, respectively. After a median observation time of 27.1 months, 37 of 85 patients (43.5%) had died of tumor disease. Kaplan-Meier analysis revealed a significant impact of nodal microinvolvement on disease-specific survival for type I and type II tumors (P = 0.016 and P <0.001, respectively). Cox regression analysis revealed a 2.77 higher independent risk (P = 0.002) for nodal microinvolvement. CONCLUSIONS: Lymphatic microinvolvement shows a high incidence in curatively resected EGJ cancer. Spread to mediastinal lymph nodes seems to necessitate lymphadenectomy of the thoracic cavity in either type.
BACKGROUND: Adenocarcinoma of the esophago-gastric junction (EGJ) potentially spreads to abdominal and mediastinal lymph nodes. METHODS: Eighty-five patients with type I and II EGJ cancer underwent curative esophagectomy or esophago-gastrectomy and radical abdominal and mediastinal lymphadenectomy. Microinvolvement was detected with the mAb Ber-Ep4 in all histopathologically free lymph nodes. RESULTS: In type I tumors (n = 40), lower mediastinal lymph nodes were positive in 24% and among type II tumors (n = 45) in 10% of patients. Ber-Ep4+ cells in apparently free lymph nodes were found in 49% of patients. On inclusion of Ber-Ep4+ nodes, positive mediastinal lymph node staging was rising to 40 and 33% in type I and II patients, respectively. After a median observation time of 27.1 months, 37 of 85 patients (43.5%) had died of tumor disease. Kaplan-Meier analysis revealed a significant impact of nodal microinvolvement on disease-specific survival for type I and type II tumors (P = 0.016 and P <0.001, respectively). Cox regression analysis revealed a 2.77 higher independent risk (P = 0.002) for nodal microinvolvement. CONCLUSIONS: Lymphatic microinvolvement shows a high incidence in curatively resected EGJ cancer. Spread to mediastinal lymph nodes seems to necessitate lymphadenectomy of the thoracic cavity in either type.