Ability to predict response to cardiac resynchronization therapy (CRT) has been one of the challenges in the treatment of heart failure (HF). Although CRT has been associated with significant clinical improvement in large clinical trials (1-3), approximately 30% of the patients do not experience clinical improvement (4). Several factors of CRT nonresponse have been implicated, such as inappropriate left ventricular (LV) lead positioning, the extent and location of See page 1128 scar tissue, lack of LV mechanical dyssynchrony, and inadequate atria-to-ventricle (AV) or ventricle-to-ventricle (VV) optimization. Recent trials demonstrating benefit in patients with New York Heart Association (NYHA) functional class I and II HF (5,6) will likely increase the number of eligible patients for CRT. However, despite advances in treatment, technology, and experience, the ratio of nonresponders still remains at approximately one-third (6). With the expansion of CRT-eligible patients, the absolute number of nonresponders will likely increase, underlining the importance of improving strategies in predicting nonresponse and therefore improving selection for CRT. Prediction of response. Several strategies have been developed to improve response to CRT, including better selection of patients, optimal LV lead placement, and optimization of device programming. The characteristics of patients likely to respond to CRT are already well-known (Fig. 1), such as nonischemic dilated cardiomyopathy, wide QRS (Ն150 ms), left bundle branch block, low burden of myocardial scar, preserved posterolateral wall function, and placement of LV lead in posterior or lateral position (7). In this issue of the Journal, Duckett et al. (8) observed that optimizing LV lead placement with the rise in maximum rate of left ventricular pressure (LV-dP/dt max ) was associated with reduction in left ventricular end-systolic volume (LVESV), improvements in NYHA functional class, and quality of life at 6 months. Acute hemodynamic improvement in dP/dt max with optimal LV lead position has been described before (9 -11). Duckett et al. (8) for the first time extend these findings to demonstrate the association of optimal coronary vein selection guided by dP/dt max with reversal of remodeling and clinical response.The patients enrolled in the study by Duckett et al. (8) fulfilled the standard guideline criteria for CRT but also had left bundle branch block, and their mean QRS duration was 160 Ϯ 23 ms. Seventy percent demonstrated an improvement in dP/dt max during LV pacing; and 77% of these patients demonstrated a reduction in LVESV. A Ն10% rise in dP/dt max during LV pacing at the optimal site predicted reduction in LVESV with a 94% sensitivity and 64% specificity. This approach is quite sensitive in predicting the responders but does not seem to provide incremental sensitivity over conventional criteria such as QRS duration. In the study by Duckett et al. (8), QRS duration of Ͼ146 ms had a higher (100%), and the echocardiographic measurement of interventricular mech...