Comparison and Agreement between Cardiovascular Computed Tomography-Derived Mid-Diastolic and End-Diastolic Ventricular Volume in Patients with Congenital Heart Disease.

Abstract

Prospective electrocardiogram (ECG)-triggered cardiovascular computed tomography (CCT) is primarily utilized for anatomical information in congenital heart disease (CHD) and has not been utilized for calculation of the end-diastolic volume (EDV); however, the mid-diastolic volume (MDV) may be measured. The objective of this study was to evaluate the feasibility and agreement between ventricular EDV and MDV. 31 retrospectively ECG-gated CCT were analyzed for the study of the 450 consecutive CCT. CCT images were processed using syngo.via with automatic contouring followed by manual adjustment of the endocardial borders of the left ventricles (LV) and right ventricles (RV) at end-diastolic and mid-diastolic phase (measured at 70% of cardiac cycle). The correlation and agreements between EDV and MDV were demonstrated using Spearman rank coefficient and intraclass correlation coefficient (ICC), respectively. Mean age ± SD was 28.8 ± 12.5 years, 19 were male (61.3%) and tetralogy of Fallot (TOF) was the most common diagnosis (58.1%), 35% (11/31) patients with a pacemaker, ICD or other such contraindication for a CMRI, 23% (7/31) with claustrophobia, and 6.5% (2/31) with developmental delay with refusal for sedation did not have a previous CMRI. The mean ± SD indexed LV EDV and LV MDV were 91.1 ± 24.5 and 84.8 ± 22.3 ml/m2, respectively. The mean ± SD indexed RV EDV and RV MDV were 136.8 ± 41 and 130.2 ± 41.5 ml/m2, respectively. EDV and MDV had a strong positive correlation and good agreement (ICC 0.92 for LV and 0.95 for RV). This agreement was preserved in a subset of patients (21) with dilated RV (indexed RV EDV z-score > 2). Intra-observer reliability (0.97 and 0.98 for LV and RV MDV, respectively) and inter-observer reliability (0.96 and 0.90 for LV and RV MDV, respectively) were excellent. In a select group of patients with CHD, measuring MDV by CCT is feasible and these values have good agreements with EDV. This may be used to derive functional data from prospectively ECG-triggered CCT studies. Further large-scale analysis is needed to determine accuracy and clinical correlation.