[Metastasectomy--a direct therapeutic effect or an illusion due to patient selection?]

Metastasectomy is widely practiced in surgical oncology. Well established examples are the resection of liver metastases in patients with colo-rectal cancer, wedge resections of lung metastases in sarcoma, or the surgical removal of solitary brain metastases derived from solid tumours. The latter example has been tested in randomised phase III trials and found to be an effective mode of treatment compared with radiation therapy alone. In the other fields no such trials are available, and data chiefly stem from retro- and prospective phase II trials. For a variety of reasons, phase II studies are not suitable to separate confounding factors such as selection of good-risk patients, stage migration linked to the improvement of preoperative staging techniques, and improved supportive care, from the direct therapeutic effects of the surgical procedure as such. Although desirable, it is unlikely that surgeons and patients would accept randomised trials to reassess evidence levels in this field. In practice, large prospective studies of case series are likely to provide the highest evidence level achievable, and these limitations of the literature should be taken into account in an appropriate fashion when patients with metastatic cancer are advised about options for surgery.