Barrett's carcinoma, usually arising in the distal esophagus, must be considered a separate entity from squamous cell esophageal cancer. Epidemiology, etiology, patients' risk profiles, biology of metastases, and prognosis differ markedly between these two major esophageal tumor types. The preoperative work-up of patients with Barrett's cancer is primarily directed toward assessing the chances for R0 resection and estimating the risk of the patient to survive an esophagectomy. If R0 resection appears likely and the surgical risk is acceptable, the indication for an operative approach is given. From the oncologic point of view there is no difference between a radical transmediastinal approach and a transthoracic approach. A possible advantage of a transthoracic approach is the extension of lymphadenectomy to the upper mediastinum. Lymph node metastases in the upper mediastinum, however, usually indicate advanced lymphatic and subclinical systemic tumor dissemination, i.e., a poor prognosis even with extended surgery. Consequently the controversies about the surgical approach are reduced to technical and functional aspects. A better swallowing function argues for an intrathoracic anastomosis; the lower morbidity, for a cervical approach. We prefer transthoracic en bloc esophagectomy with an intrathoracic anastomosis in patients with moderate risk and early tumor stages. In all other patients radical transmediastinal esophagectomy with a cervical anastomosis is the procedure of choice. The overall 5-year survival rate of more than 40%, which is superior to most published data, supports this therapeutic strategy.