Aim. To compare endocardial and epicardial left ventricular (LV) pacing using real-time electrocardiography (ECG)-synchronized three-dimensional echocardiography (3DE).Methods. Experimental intraoperative study included 88 points obtained from 12 patients with compensated heart failure of II-IV functional class NYHA (LV ejection fraction < 35%) and cardiac resynchronization therapy indications - ECG pattern of complete left bundle branch block (LBBB) and QRS complex duration > 150 ms. During isolated LV pacing as part of cardiac resynchronization therapy implantation procedure endocardial and epicardial stimulation points matched under fluoroscopic control using quadripolar coronary sinus leads and endocardial electrodes for temporary pacing were obtained. The overall number of corresponding pacing sites included 44 endocardial and 44 epicardial stimulation positions. The mean age of patients was 68.5 [63; 73.5] years, 83% males (n=10). Before study enrollment, 12-channel ECG, echocardiography, and a six-minute walk test were performed for all participants along with cardiac magnetic resonance imaging and control coronary angiography if indicated. The prevalence of coronary heart disease was 50% (n=6) while dilated cardiomyopathy was the most common etiology of chronic heart failure in other cases. Intraoperative ECG with estimation of paced QRS complex morphology at each point was registered via LabSystem Pro Electrophysiological Recording System (Bard Electrophysiology, USA). 3DE was performed using TomTec and Philips Qlab 3DQ Advanced software (Philips Medical Systems, USA).Results. Three-dimensional parametric imaging of LV regional segmental excursion and myocardial contractility using 3DE revealed statistically significant difference in semi-quantative parameters such as ExcAvg (p<0.001), ExcMax (p=0.001), ExcMin (p<0.001) and LV ejection fraction based on 3D modelling (p=0.003) while endocardial pacing was more beneficial. During the course of endocardial stimulation, the 3DE dyssynchrony index estimated at the 2nd stimulation site was also significantly lower (p=0.03). Identical dyssynchrony parameters valid for the 16 and 12-segment 3D models (SDI-16, Tmsv-12SD) (at p=0.06) demonstrated only a tendency for significant difference. The duration of QRS complex at the time of endocardial pacing was significantly shorter (<190 [179;215] ms) (p=0.0008). Semi-quantitative and quantitative 3DE parameters showed the benefit of endocardial pacing resulting in cardiac contractility improvement with less dyssynchrony and LV volume reducing during intraoperative period.Conclusion. Endocardial pacing has potential benefit over the epicardial pacing represented by intraoperative dynamics of LV global and local contractility, intraventricular dyssynchrony estimated by 3DE and also ECG criteria. 3DE is helpful in more precise and reproducibile determing of late activation zone for target LV lead placement that is more manoeuvrable in case of endocardial stimulation.