BACKGROUND: The Selvester QRS score is an electrocardiographic tool designed to quantify myocardial scar. It was updated in 2009 to expand its usefulness in patients with conduction abnormalities such as bundle-branch and fascicular blocks. There is need to further validate the updated score in a broader group of patients with cardiovascular disease and conduction abnormalities. We primarily hypothesized that the updated score could distinguish between presence and absence of scar by cardiac magnetic resonance imaging (CMR) with late gadolinium enhancement in 4 groups of patients with distinct conduction abnormalitites. METHODS: A total of 193 patients were retrospectively identified that had received an electrocardiogram (ECG) and a CMR scan at Duke University Medical Center between January 2011 and August 2013: 62 with left bundle-branch block, 51 with right bundle-branch block (RBBB), 43 with left anterior fascicular block (LAFB), and 37 with RBBB + LAFB. Scar sizes estimated by ECG and by CMR were compared using scatterplots, modified Bland-Altman plots, and receiver operating characteristics curves. RESULTS: Of 193 patients, 96 (50%) had no scar by CMR. The QRS score generally overestimated CMR scar. The area under the curve ranged between 0.62 and 0.65 for the different conduction types, and 95% confidence intervals included 0.5 for all conduction types. Performance was slightly improved in LAFB and RBBB + LAFB by excluding all points derived from leads V4-V6. CONCLUSIONS: The Selvester QRS score for use in conduction abnormalities needs to be improved, primarily its specificity, to enable effective clinical use in a population with a wide range of left ventricular ejection fraction and low pretest probability of myocardial scar.