Reliability of MDCT in characterizing pulmonary venous drainage, diameter and distance to first bifurcation: an...

To evaluate the interobserver agreement of readers in evaluating pulmonary venous anatomy and in measuring pulmonary vein ostial diameters and distance to first bifurcation.

This study was approved by our institutional review board. Thin-section contrast material-enhanced multidetector computed tomography examinations of the thorax were retrospectively reviewed in 200 consecutive patients (38 females and 162 males), age 24-79 years (mean, 52.8) referred for imaging before radiofrequency ablation therapy for atrial fibrillation. For each patient, pulmonary venous anatomy and drainage patterns including the number of venous ostia was assessed independently by experienced cardiothoracic radiologists. Pulmonary vein ostial diameter and distance to the first bifurcation of the four major pulmonary veins (right inferior and superior, left inferior and superior), the middle lobe pulmonary vein, and any anomalous pulmonary veins (common trunks and accessory veins) were measured independently at a workstation. Interreader assessment of pulmonary venous anatomy was evaluated using the Kappa statistic. Interreader variation in measurements of venous diameter and distant to first bifurcation were estimated by Bland-Altman plots and Pitman's test of difference in variance.

Very good to excellent interreader agreement in detection of anomalous pulmonary venous anatomy, middle lobe pulmonary venous drainage, and other thoracic venous anomalies. No significant variation between readers in pulmonary vein ostial diameter measurements for the four major and middle lobe pulmonary veins, or the anomalous pulmonary veins. Significant interreader variability was noted in measurements of the pulmonary vein distance to first bifurcation for the right inferior (P = .017), middle lobe (P = .005), and left inferior (P = .015) pulmonary veins.

There is excellent interobserver agreement when evaluating normal and anomalous pulmonary venous drainage patterns, and when measuring normal or anomalous pulmonary vein diameters. However, measurements of distances to first bifurcation were less reliable across readers.