Contrast Echocardiography in two-dimensional left ventricular measurements: comparison with 256-row multi-detector computed tomography as a reference standard in Beagles
- Author:
Jaehwan KIM
1
;
Soyoung KIM
;
Yeonhea LEE
;
Hakyoung YOON
;
Kidong EOM
Author Information
- Publication Type:Original Article
- Keywords: Contrast echocardiography; left ventricular measurements; cardiac computed tomography; multidetector computed tomography; dogs
- MeSH: Animals; Diastole; Dogs; Echocardiography; Multidetector Computed Tomography; Systole
- From:Journal of Veterinary Science 2019;20(5):e45-
- CountryRepublic of Korea
- Language:English
- Abstract: Unenhanced echocardiography (UE), commonly used in veterinary practice, is limited by left ventricular (LV) foreshortening and observer dependency. Contrast echocardiography (CE) was used to compare two-dimensional (2D) LV measurements made using UE and 256-row multi-detector computed tomography (MDCT) as a reference standard. Seven healthy beagle dogs were evaluated in this study. Measurements obtained using CE, including LV wall thickness, internal diameter, and longitudinal and transverse length, were significantly greater than those obtained using UE. Measurements of LV internal dimension in diastole (LVIDd) and systole (LVIDs) were significantly larger with CE compared UE. Regardless of the cardiac cycle, LV longitudinal (LVLd and LVLs) and transverse diameter (LVTDd and LVTDs) measurements were significantly different with CE and approximated values from MDCT. Among automatically calculated parameters, LV end-systolic volume and the relative wall thickness were significantly different between UE and CE. In CE, the correlation coefficients of 4 major parameters (r = 0.87 in LVIDd; 0.91 in LVIDs; 0.87 in LVLd; and 0.81 in LVLs) showed higher values compared to the UE (r = 0.68 in LVIDd, 0.71 in LVIDs, 0.69 in LVLd, and 0.35 in LVLs). Inter-observer agreement was highest for MDCT and higher for CE than UE. In conclusion, CE is more accurate and reproducible than UE in assessing 2D LV measurements and can overcome the limitations of UE including LV foreshortening and high observer dependency.
