Comparison of ECG-gated Multidetector Computed Tomography and Two-Dimensional Echocardiographic methods in the Assessment of Left Atrial Size using Left atrial-to-Aortic Root Short-Axis Ratio in Dogs

Left atrial size (LAS) is an important prognostic indicator in canine cardiac diseases. A widely used, unidimensional, echocardiographic method to estimate LAS is the left atrial to-aortic-root ratio (LA:Ao) using a short axis plane at the level of the root of the aorta. However, observer-dependant variability of this ratio has been reported, and measurements may vary considerably even among highly trained cardiologists. In humans, ECG-gated cross-sectional imaging modalities, such as multi-detector computed tomography angiography (MCTA) are used as reference methods for measuring LAS. MCTA limits cut-plane variability and should therefore provide a more accurate and reproducible evaluation of left atrial morphology and borders compared to echocardiography in dogs. We hypothesized that there is a poor correlation between LA:Ao calculated by echocardiography (LA:Aous) and by ECG-gated MCTA (LA:AoMCTA) when aligned to the same right parasternal short axis view. To investigate this hypothesis, we compared two-dimensional echocardiographic and 64-slice ECG-gated MCTA measurements of LA:Ao (LA:Aous and LA:AoMCTA respectively) at end-diastole and end-systole in dogs. Measurements of LA:Ao were performed using the Hansson method. Levels of agreement between these modalities and intra- and inter-observer variability were calculated. This single centre, prospective, cross-sectional study was approved by the local Veterinary Ethical Review Committee. Dogs of various breeds undergoing anaesthetised thoracic MCTA for various diagnostic purposes were enrolled from a previous study on left atrial volume. Eligible dogs underwent a physical examination including blood pressure measurement prior to anaesthesia and cardiac auscultation to ensure that no heart murmur was present. Echocardiography was performed immediately following ECG-gated MCTA. Measurements of LA:Ao were obtained at end-diastole and end-systole, identified as the frame at the onset of the QRS complex and the frame immediately after aortic valve closure respectively. Agreement between LA:Aous and LA:AoMCTA was evaluated using Bland-Altman analysis and Pearson correlation. Coefficients of variation (CVs) were calculated to quantify intra- and inter-observer variability. The correlation between LA:Aous and LA:AoMCTA was moderate at end-systole (r=0.52, p=0.018), whereas there was no correlation at end-diastole (p= 0.99). Echocardiography underestimated LA:Ao compared to MCTA at end-systole with a bias of -0.22 (95% limits of agreement: -0.68 to 0.24). Intra- and inter-observer variability for LA:Aous and LA:AoMCTA at end-systole and end-diastole were adequate (CVs<15%). In conclusion, our data demonstrate that the end-systole measurement of LA:Aous is a superior ratio-metric estimate of LA:Aous than the end-diastole measurement, but the ratio is underestimated compared to LA:AoMCTA.