The safety and feasibility of temperature-guided radiofrequency (RF) ablation of ventricular myocardium were assessed in an open-chest animal model. RF ablation was performed in the ventricles of 19 pigs using ablation catheters with a tip electrode of 2 or 4 mm length. The energy was delivered in a bipolar (2 mm tip electrode only) or unipolar configuration. Set temperature ranged from 60 to 90 degrees C. Pulse duration was 30 s. Histological examination showed coagulation necrosis with a haemorrhagic zone. However, lesions with a deep cleavage were found after five unipolar (2 mm tip electrode) energy applications. No such lesions were found after unipolar energy applications with a 4 mm tip electrode. During or shortly after ablation, premature ventricular beats and non-sustained VT were frequently observed and in some instances ventricular fibrillation. In the bipolar mode we found a positive correlation between lesion area/volume and peak temperature with a correlation coefficient of 0.48 and 0.56, respectively (P < 0.05). However no correlation was found between lesion size and applied energy. In the unipolar configuration there was no correlation between lesion size and these parameters. Median depth and area were: bipolar: 4.0 mm/23.5 mm2, unipolar (2 mm tip electrode): 3.5 mm/12.2 mm2, unipolar (4 mm tip electrode): 4.0 mm/15.7 mm2. We conclude that in a beating heart it is difficult to predict lesion size from temperature or energy.