BACKGROUND Obtaining diagnostic coronary CT angiography with low radiation exposure in patients with irregular heart rhythms such as atrial fibrillation (AF) remains challenging. OBJECTIVE We evaluated image quality and inter-reader variability with the use of prospective electrocardiographic (ECG)-triggered sequential dual-source acquisition at end systole for coronary artery disease (CAD) evaluation in patients with AF. METHODS Thirty consecutive patients with AF who underwent prospective ECG-triggered sequential dual-source acquisition were evaluated. Images were reconstructed every 50 milliseconds from 250 to 400 milliseconds after the R wave. Two independent, blinded readers evaluated the coronaries for image quality on a 5-point scale (worst to best) and stenosis on 5-point semiquantitative (none to severe) and binary scales (>50% or <50%). Diagnostic image quality was graded for each reconstruction. RESULTS Eleven patients (37%) had significant (≥50% stenosis) CAD. Average heart rate was 82 ± 20 beats/min and variability range was 71 ± 22 beats/min. Mean effective radiation dose was 6.5 ± 2.4 mSv. Diagnostic image quality was noted in 97.9% of 304 coronary segments with median image quality of 3.0. The 300-millisecond reconstruction phase provided the highest image quality; 70% of patients showed diagnostic image quality. Combination of all phases (250-400 milliseconds) performed significantly better than single or other phase combinations (P < 0.0005 for all comparisons). Inter-reader variability for stenosis detection was excellent, with 98.4% concordance by using a binary scale (50% stenosis cutoff). CONCLUSIONS Prospective ECG-triggered sequential dual-source CT acquisition with the use of end-systolic acquisition provides diagnostic image quality with potentially low radiation doses for evaluation of CAD in patients with AF. Use of multiple end-systolic phases over a 150-millisecond window improves diagnostic image quality.