TDK conference 2020

In canine thoracic radiography the individual heart chambers are characterized as a part of cardiac interpretation (e.g. enlarged right atrium, enlarged left ventricle). This change in the size of heart cavities can be deducted from the morphology of the heart silhouette on the x-ray, but opinions can be rather subjective. Our study compares the radiographic parameters to the echocardiographic parameters as the gold standard, in terms of the individual heart cavities.

Dogs presented at the Small Animal Clinic of the University of Veterinary Medicine Budapest for both a thoracic X-ray and echocardiography were enrolled in the study. The two studies had to be conducted within a 24 hours time interval. On hundred dogs met our criteria (30 healthy, 70 ill) the sexes had a ratio of 55/45 (female/male).

The size of each heart chambers was assessed and radiographic and ultrasonographic results were compared. The interobserver variability was examined through the involvement of a beginner and expert examiner. The ultrasound results were grouped into “normal” and “enlarged” as these are the two well-defined categories, whereas we used the categories of "normal", "doubtful", "doubtful+" and "enlarged" for the radiographic findings which segmentation reflects better the subjective nature of the latter method. By taking the ultrasound as the gold standard, we calculated the specificity and the sensitivity of the radiographic studies for each heart cavity for the different observers as well as their average. The specificity for the left ventricle was 80% but was only 69.7% for the left atrium. The sensitivity for the left ventricle and atrium were 40% and 57.1% respectively. Regarding the enlargement of the right ventricle, the specificity of the radiographic study was 29.6% and 39.3% for the right atrium, while the sensitivity was 22.2% for the right ventricle and 18.2% for the right atrium.

Based on our findings, radiography seems to be the best predictor for the left ventricle and is the least reliable for the right atrium. The match between the two analyst was 63% for the left ventricle, 64% for the right atrium and 46% for both the right ventricle and right atrium.

Our results support that echocardiographic examination is fairly inaccurate in assessing the size of individual cardiac chambers and also quantify the reliability of the radiographic method.