AIM: To test the hypothesis that magnetic resonance (MR)-guided hepatic tumour ablation is (i) safe and feasible, (ii) is associated with favourable patient survival, and (iii) decreases viable tumour. MATERIALS AND METHODS: One hundred and twenty-five MR-guided laser thermal ablations (LTA) were performed on 35 patients with hepatocellular carcinoma (HCC, n = 19), hepatic metastases (n = 11, mainly colorectal) and carcinoid liver tumours (n = 5). RESULTS: Mean overall survival was 14.8 months (HCCs 14.6 months, metastases 15.2 months). Near real-time T1-weighted colourized thermal maps correlated moderately with follow-up gadolinium-enhanced MR imaging in predicting ablated tumour area (Pearson correlation coefficient = 0.5). There was a significant difference in percentage enhancing pre- and post-LTA (Wilcoxon signed ranks test = 0.0001). An average of 50.7% of tumour was ablated by each treatment. In patients with multiple liver tumours ablated tumours grew significantly less than untreated tumours (108% compared with 196% growth, follow-up period 5.8 months, WSRT p = 0.07). CONCLUSION: MR-guided LTA of primary and secondary liver tumours is safe, feasible, and significantly decreased amount of enhancing or viable tumour. MR-guided LTA produces a better survival in patients with HCC than would be expected in untreated patients, and has a mean survival in patients with metastases at least equal to the longest median survival in untreated patients. (C) 2003 The Royal College of Radiologists Published by Elsevier Science Ltd. All rights reserved.