Catheter ablation (CA) of idiopathic ventricular arrhythmias (VAs) from the epicardial left ventricular summit is challenging. The endocardial approach targets two sites: the endocardial closest site (ECS) to the epicardial earliest activation site (epi-EAS) and the endocardial EAS (endo-EAS). We aimed to differentiate between cases where CA at the ECS was effective and where CA at the endocardial EAS yielded success. Fifty-eight patients (47 men, 60±13 years) were analyzed with VAs in which the EAS was observed in the coronary venous system (CVS). Overall, VAs were successfully eliminated in 42 (72%) patients; 8 in the CVS, 8 where the ECS matched with the endo-EAS, 11 at the ECS, and 15 at the endo-EAS. Successful ECS ablation was associated with a shorter epi-EAS-ECS distance (10.2±4.7 vs. 18.8±5.3 mm; P<0.001) and shorter epi-EAS-left main coronary trunk (LMT) ostial distance (20.3±7.6 vs. 30.3±8.4 mm; P=0.005), with optimal cut-off values of ≤12.6 mm and ≤24.0 mm, respectively. Successful endo-EAS ablation was associated with an earlier electrogram at the endo-EAS (23 [18, 36] vs. 15 [0, 19] ms preceding the QRS; P<0.001) and shorter epi-EAS-endo-EAS interval (6 [1, 8] vs. 22 [12, 25] ms; P<0.001), with optimal cut-off values of ≥18 ms and ≤9 ms, respectively. Shorter anatomical distances between the epi-EAS and ECS, and between the epi-EAS and LMT ostium, predict a successful ECS ablation. The prematurity of the endo-EAS electrogram and a shorter interval between the epi-EAS and endo-EAS predicted a successful endo-EAS ablation.