Aim: Noninvasive electrophysiological mapping (NEM) can be performed using computed or magnatic resonance tomography (CT or MRI). However, the question arises as to whether the accuracy of the electrical processes reconstruction on three-dimensional heart models is the same when different types of tomography are used. Purpose: To compare the left ventricle (LV) late activation zones determined using NEM with sequential CT and MRI. Methods: group included 11 patients aged: median (min; max) – 67 (57; 78). All patients had sinus rhythm and complete left bundle branch block (LBBB) activation pattern (n=10) or a nonspecific intraventricular block of the QRS complex (n=1) with median (min; max) QRS duration - 195 (141; 235) ms. The NEM was performed using Amycard 01C EP LAB system (EP Solutions SA, Switzerland) with 3 steps. At the first stage a 240-multichannel body surface ECG was recorded. At the second stage a torso and cardiac CT was performed. At the third stage torso and cardiac MRI was performed. The activation of the LV epicardium with LBBB was evaluated using NEM. Reconstructed polygonal epicardial models based on CT and MRI were imported into custom written software with Python (Python Software). The two models were fused using the iterative closest point algorithm and both quantitative and qualitative comparisons were performed. All fused models were visualized using Paraview (Kitware Inc). Results: When the reconstructed polygonal epicardial models based on CT and MRI were combined using the Python program, a segmental coincidence of the LV late activation zones was revealed. The quality of fusion was sufficient, the error was minimal, maps were not completely identical but main activation patters were similar. Conclusion: When CT or MRI is used for non-invasive electrophysiological mapping, similar results are obtained when three-dimensional models are constructed and the late zones of LV activation are evaluated.