Surgery is the primary and most effective treatment of breast cancer, but minimal residual disease is probably unavoidable. Whether residual disease results in clinical metastases depends on numerous factors, including anti-tumor cell mediated immunity and angiogenic and growth signals in sites of residual disease. At least three perioperative factors adversely affect these: 1) the neuroendocrine stress response to surgery, 2) volatile anesthetics, and 3) opioids. Animal studies indicate that regional anesthesia and optimum postoperative analgesia independently reduce the metastatic burden in animals inoculated with breast adenocarcinoma cells following surgery. Retrospective studies in humans also suggest that regional analgesia may reduce recurrence risk after cancer surgery. We will test the hypothesis that local or metastatic recurrence after breast cancer surgery is lower in patients randomized to paravertebral or high-thoracic epidural analgesia combined with sedation or light anesthesia than in patients given intraoperative volatile anesthesia and postoperative opioid analgesia. In a Phase III, multi-center trial, Stage 1-3 patients having mastectomies for cancer will be randomly assigned to thoracic epidural or paravertebral anesthesia/analgesia, or to sevoflurane anesthesia and morphine analgesia. The primary outcome will be cancer recurrence. Enrolling 1100 patients over 5 years will provide 85% power for detecting a 30% treatment effect at an alpha of 0.05. We plan four equally spaced interim analyses, each evaluating efficacy and futility. Confirming our hypothesis will indicate that a small modification to anesthetic management, one that can be implemented with little risk or cost, will reduce the risk of cancer recurrence - a complication that is often ultimately lethal.