The purpose of this study was to compare the repeatability of selected transthoracic measurements acquired from two-dimensional echocardiography (2DE), 2D guided M-mode, and three-dimensional echocardiography (3DE) in normal conscious dogs and determine the optimal number of repetitions required to maximize precision for each modality.
Five conscious dogs (18-25 kg) underwent transthoracic 2DE and 3DE examinations by four investigators (SG - cardiologist, AS - cardiologist, CH - cardiologist, RF – cardiology resident). Each dog was imaged five times by each investigator. Each investigator was limited to three studies on any single day. Measurements were performed by the investigator that acquired the images off line at a later time. One additional investigator (MM - cardiologist) completed five 2DE and 3DE examinations on one dog over two days and three different investigators (SG, AS, RF) then performed measurements off line and were blinded to each other’s results.
Repeatability, intra-observer, and inter-observer image acquisition variability and measurement variability were quantified by average coefficient of variation (CV). Lower and upper limit bounds on the coefficient of variation were calculated using exact theory and logistic regression of the upper limit against repetitions of 5, 10, 15, and 25 was used to determine the highest precision for each modality.
Evaluation of the left ventricular size by M-mode had the highest repeatability (CV = 8.8%) and lowest intra- and inter-observer variability for both image acquisition (CV = 9.3%) and measurement (CV = 4.3%). Repeatability of 3DE image acquisition (CV = 21.1%) and measurement (CV = 8.6%) was slightly better than 2DE image acquisition (CV = 21.5%) and measurement (CV = 8.9%). The intra- and inter-observer variability of 3DE (CV = 18.9%) was also slightly better than then 2DE (CV = 22.3%). 2DE left atrial to aortic ratio measurements had higher repeatability (CV = 9.0%) and less variability (CV = 9.5%) than volume assessed by 3DE image acquisition (CV = 15.1 %) and measurement (CV = 12.0%). M-mode assessment of left ventricular size achieved acceptable precision after 5 repetitions while left ventricular size assessed by 3DE and 2DE required 10 and 15 repetitions respectively. These data may be useful in the design of study protocols that plan to use echocardiographic assessment as an outcome variable.
Clinical Assistant Professor
University of Illinois at Urbana-Champaign
Ryan Fries, DVM, DACVIM (Cardiology)
Dr. Ryan Fries graduated from Ross University College of Veterinary Medicine in 2007. After a year in general practice, he completed in a small animal rotating internship at The Ohio State University College of Veterinary Medicine in 2009. He then completed a 3-year residency program in cardiology at Texas A&M University and became board certified by the American College of Veterinary Internal Medicine in 2012. Dr. Fries completed a cardiovascular MRI course through Northwestern University in 2015 and has obtained level one certification in cardiovascular MRI. Dr. Fries is a member of the Society for Cardiovascular MRI and is actively performing veterinary clinical research in this field. In addition to cardiac MRI, Dr. Fries’ research includes: echocardiography and advanced imaging, feline and canine cardiomyopathies, and congenital heart disease. He is currently a clinical assistant professor and the head of cardiology at the University of Illinois at Urbana-Champaign.
Thursday, June 14
4:10 PM – 5:00 PM
Thursday, June 14
5:45 PM – 6:00 PM
The asset you are trying to access is locked. Please enter your access key to unlock.