Transcatheter placement of left atrial closure device is an attractive therapy for patients with atrial fibrillation (AF), to avoid anticoagulation and reduce cerebrovascular events; however peri-device leaks occur. The geometry of the left atrial appendage (LAA) is not well understood, largely owing to limitations of 2-dimensional imaging techniques. We sought to better define the LAA orifice geometry, by performing 3-dimensional multi-detector computed tomography measurements. We prospectively recruited 105 consecutive patients referred for pulmonary vein ablation (PVA) and age-matched controls. Area, short and long-axis measurements were performed. Eccentricity was calculated as 1-(short axis/long axis). Multiple clinical variables were tested for their ability to predict appendage orifice eccentricity using univariate linear regression models. The PVA cohort demographics included; 25 (24%) females, mean age 59 years (SD = 10), median height (1.55-2.03), weight 89 (56-139) kg and body surface area 2.1 (1.61-2.58). In the PVA cohort, there was a significant difference between the long and short-axis; median short-axis dimension was 20.5 (12.9-35.4) mm, versus long-axis median 30.4 (17.7-43.8) (p < 0.001). Mean eccentricity score was 0.4. When compared with controls, there was a significant difference in the short and long-axis measurements (p < 0.001) as well as eccentricity (p = 0.04). All clinical variables tested showed limited ability to predict appendage eccentricity (p = NS). LAA ostium is an elliptical structure in the setting of AF with a high eccentricity index and uniformly significant differences between short and long-axis. There were significant differences between these parameters when compared with controls. A deeper appreciation of LAA geometry and eccentricity may allow for reduction in peri-closure leaks.