[CT coronary angiography in patients with atrial fibrillation]

Reliable visualization of the coronary arteries with multislice spiral CT angiography (MSCTA) in patients with atrial fibrillation (AF) remains a challenge despite retrospective ECG gating. A recently developed new algorithm automatically compensates dynamic changes in the heart rate during the scan, thus reducing misregistration and motion artifacts. The HeartBeat-RT algorithm combines a fixed-percent delay determined from the first R wave and the fixed offset delay based on the second R wave in the ECG cycle. The purpose of this study was to find out the optimal reconstruction window in MSCTA in patients with AF for each of the three major coronary arteries during the cardiac cycle.

20 patients with permanent AF were imaged on a 16-slice scanner (slice collimation: 16 x 0.75 mm; rotation time 0.42 s; 140 kV; 380 mAs; 120 ml Ultravist 370 (R) i.v.). The patients had not received any previous drugs for heart frequency regulation. Acquisition was started after bolus tracking of a biphasic bolus of 120 ml Ultravist 370 injected intravenously. Each coronary segment was reconstructed at 0 % - 90 % of the cardiac cycle in increments of 10 %. For image analysis we used coronary segments as defined by the American Heart Association. Two blinded independent readers assessed the image quality in terms of visibility and artifacts (five-point rating scale 1 = very poor, 2 = poor, 3 = fair, 4 = good and 5 = excellent) and the degree of stenosis (five-point rating scale 1 = 0 %, 2 = 1 % - 49 %, 3 = 50 % - 74 %, 4 = 75 % - 99 %, 5 = 100 %) on axial slices, multiplanar reconstructions and three-dimensional volume-rendered images.

The heart rate during examination ranged between 42 and 156 beats per minute, the average heart rate was 78 +/- 23. Each of the two readers evaluated 300 segments in 20 patients. Visualization of all coronary artery segments was superior at 40 % (mean score of the image quality 2.79) as compared to the standard diastolic reconstruction window at 80 % (image quality 2.33). The second best image quality (2.57) was acquired at 0 % of the cardiac cycle.

The use of a frequency adapted delay algorithm with the choice of an end-systolic reconstruction window provided diagnostically valuable images in patients with AF.