A complete tumor removal with an adequate safety margin in all three dimensions (the oral margin, the aboral margins and the tumor bed) must be the primary aim of any surgical approach to esophageal cancer. The same goal has to be achieved in the area of the lymphatic drainage. The safety margin of lymphadenectomy can be estimated by the so-called lymph-node ratio, i.e., the ratio between the number of positive nodes and removed nodes. Several studies have shown that, for esophageal carcinoma, a lymph-node ratio below 0.2 constitutes an independent prognostic factor. Although controlled trials are still lacking, these data suggest that extensive lymphadenectomy may thus improve the prognosis in patients at an early stage of lymphatic spread, i.e., patients with only lymph-node `micro-involvement' or patients with a limited number of positive regional nodes on standard histopathologic assessment. In practice, this requires, as a minimum, a two-field lymphadenectomy. In patients with more advanced lymphatic metastases, two-field lymphadenectomy does not improve the prognosis and can only result in a reduction of local recurrences. A more extensive lymphadenectomy, i.e., three-field lymph-node dissection, increases the risk and morbidity of the surgical procedure, while a prognostic gain, if any, appears to be limited to a subgroup of patients with proximal tumors and less than five involved lymph nodes. Since, in the Western world, these patients are usually submitted to multimodal therapeutic protocols, extended three-field lymphadenectomy can currently not be recommended as standard therapy.