Multimodal therapeutic concepts for the treatment of liver metastasis are currently undergoing evaluation--prompted by the fact that few patients have lasting benefit from resection alone. Thus, for example, in practice, virtually only metastases from colorectal carcinoma or neuroendocrine tumors can be referred to surgical treatment, and, of these, only 20-25% are technically resectable. Furthermore, even after an R0 resection, recurrent disease subsequently develops in 60-75% of the cases. In primarily non-resectable colorectal liver metastases, prior systemic treatment with 5-FU/folinic acid and oxaliplatin can result in partial or complete remission in 50-60% of cases and, depending on patient selection criteria, a secondary R0 resection rendered possible in 14-38%. Theoretical oncological considerations suggest that neoadjuvant treatment should be applied in the case of resectable liver metastases too. The question of whether the prognosis is then improved compared with resection alone is currently being investigated in a prospective multicentre study conducted by the EORTC. The value of adjuvant therapy administered with the aim of lowering the risk of recurrence following "curative" resection of liver metastases is presently not considered to have been adequately demonstrated. With regard to the efficacy of regional chemotherapy, the results of two prospective randomized studies are contradictory. Nor can the multimodal approach decisively improve the outcome of non-radical resection of metastatic lesions. This means that primary or secondary resection with a margin of clearance continues to represent the gold standard for the treatment of colorectal liver metastases. Neoadjuvant or adjuvant chemotherapy--where applicable with the additional use of various methods of thermal ablation (cryotherapy, laser therapy, high-frequency thermotherapy)--should be restricted to clinical trials.