Assessment of the left atrial appendage (LAA) for thrombus and anatomy is important prior to atrial fibrillation (AF) ablation and LAA exclusion. The use of cardiovascular CT (CCT) to detect LAA thrombus has been limited by the high incidence of pseudothrombus on single-pass studies. We evaluated the diagnostic accuracy of a two-phase protocol incorporating a limited low-dose delayed contrast-enhanced examination of the LAA, compared with a single-pass study for LAA morphological assessment, and transesophageal echocardiography (TEE) for the exclusion of thrombus. Consecutive patients (n = 122) undergoing left atrial interventions for AF were assessed. All had a two-phase CCT protocol (first-past scan plus a limited, 60-s delayed scan of the LAA) and TEE. Sensitivity, specificity, diagnostic accuracy, positive (PPV) and negative predictive values (NPV) were calculated for the detection of true thrombus on first-pass and delayed scans, using TEE as the gold standard. Overall, 20/122 (16.4 %) patients had filling defects on the first-pass study. All affected the full delineation of the LAA morphology; 17/20 (85 %) were confirmed as pseudo-filling defects. Three (15 %) were seen on late-pass and confirmed as true thrombi on TEE; a significant improvement in diagnostic performance relative to a single-pass scan (McNemar Chi-square 17, p < 0.001). The sensitivity, specificity, diagnostic accuracy, PPV and NPV was 100, 85.7, 86.1, 15.0 and 100 % respectively for first-pass scans, and 100 % for all parameters for the delayed scans. The median (range) additional radiation dose for the delayed scan was 0.4 (0.2–0.6) mSv. A low-dose delayed scan significantly improves the identification of true LAA anatomy and thrombus in patients undergoing LA intervention.