Background: The current ECG criteria for diagnosing left bundle branch block (LBBB) still cannot fully differentiate between true and false blocks. The absence or presence of an LBBB is key in improving the response rate of clinical cardiac resynchronization therapy. Methods: We hypothesized that the notch width of the QRS complex in the lateral leads (I, avL, V5, V6) on the LBBB-like ECG could further confirm the diagnosis of true LBBB. We performed high-density, three-dimensional electroanatomical mapping in the cardiac chambers of 37 patients scheduled to undergo CRT and whose preoperative electrocardiograms met the ACC/AHA/HRS guidelines for the diagnosis of complete LBBB. We compared the clinical characteristics, the real-time correspondence between the spread of ventricular electrical excitation and the QRS wave, the QRS notch width of the lateral leads, and the notch width/left ventricular end-diastolic diameter (Nw/LVd)ratio between the two groups. Results: Twenty-five patients were recruited to the t-LBBB group, and 12 to the f-LBBB group. In the t-LBBB group, the first peak of the QRS notch corresponded to the depolarization of the right ventricle and septum, the trough corresponded to the depolarization of the left ventricle across the left ventricle,and the second peak corresponded to the depolarization of the left ventricular free wall. In the f-LBBB group,the first peak corresponded to the depolarization of the right ventricle and most of the left ventricle,the second peak corresponded to the depolarization of the latest, locally-activated myocardium of the left ventricle,and the trough was caused by the off-peak delayed activation of the left ventricle. The QRS Nw (45.2 ± 12.3 ms vs.52.5 ±9.2 ms, P<0.05) and the Nw/LVd (0.65 ± 0.19 ms/mm vs. 0.81 ± 0.17 ms/mm, P<0.05) were compared between the two groups.Through ROC correlation analysis, the sensitivity (88%),specificity (58%),and cut-off value (0.56) for Nw/LVd diagnosis of t-LBBB was obtained.