Comparison of single-source cardiac CT and CMR quantified ventricular volumes and function in congenital heart disease

Background: Cardiac CT (CCT) is important for anatomic evaluation of congenital heart disease (CHD) prior to pulmonary valve replacement (PVR). However, volumetric and functional criteria for PVR are derived from cardiac MRI (CMR). Systematic differences between CCT and CMR volumes are underexplored in patients with CHD. Methods: Retrospective review of CHD patients with CMR and single-source CCT<180 days apart. Ventricular volumes were recontoured by blinded experts and global agreement was compared. Right ventricular regional differences in contours were assessed. Agreement of CCT with CMR-defined criteria for PVR was reported. Results: Twenty-nine patients (mean age 33 years, 48 ​% tetralogy of Fallot, 24 ​% congenital pulmonary stenosis, 83 ​% evaluated for PVR) had average CMR RVEDVi 152 ​mL/m², RVESVi 80 ​mL/m², RVEF 49 ​%, and RVEDV:LVEDV 1.9:1. CCT measured significantly higher RVEDVi (mean difference (MD) +17 ​mL/m²), RVESVi (MD +17 ​mL/m²), and RVEDV:LVEDV (MD +0.1) with no difference in stroke volume. There was a lower RVEF (MD -5 ​%). CCT had 90–100 ​% sensitivity/NPV to identify CMR-defined RV PVR thresholds, but had lower specificity and PPV. Faster heart rates had higher RVESVi CCT-CMR difference. The basal and mid-inferior RV contours contributed the most to CCT-CMR differences. Conclusions: Single-source CCT measures higher RV volumes and lower EF compared to CMR (i.e. more adversely-remodeled). Mechanisms include inferior stretch due to differences in breathing-instruction, and misidentification of end-systole. CMR-derived PVR thresholds applied to CCT would lead to more proactive intervention. “Adjusting” single-source CCT volumes by the observed difference between modalities is a reasonable approach. Single-source CCT-specific volumetric recommendations for PVR are needed.