Automated and Manual Measurements of the Aortic Annulus with ECG-Gated Cardiac CT Angiography Prior to Transcatheter Aortic Valve Replacement: Comparison with 3D-Transesophageal Echocardiography

Rationale and objectives: Multimodality evaluation of the aortic annulus is generally advocated to plan for transcatheter aortic valve replacement (TAVR). We compared aortic annular measurements by cardiac computed tomography angiography (cCTA) to three-dimensional transesophageal echocardiography (3D-TEE), and also evaluated the use of semi-automated software for cCTA annular measurements.

Materials and methods: A retrospective cohort of 74 patients underwent 3D-TEE and electrocardiogram-gated cCTA of the heart within 30 days for TAVR planning. 3D-TEE measurements were obtained during mid-systole; cCTA measurements were obtained during late-systole (40% of R-R interval) and mid-diastole (80% of R-R interval). Annular area was measured independently by manual planimetry and with semi-automated software.

Results: cCTA measurements in systole and diastole were highly correlated for short-axis diameter (r = 0.91), long-axis diameter (r = 0.92), and annular area (r = 0.96), although systolic measurements were significantly larger (P < 0.001), most notably for the short-axis diameter. Good correlation was observed between 3D-TEE and cCTA for short-axis diameter (r = 0.84-0.90), long-axis diameter (r = 0.77-0.79), and annular area (r = 0.89-0.90). As compared to 3D-TEE, annular area is overmeasured by 28 mm² on systolic phase cCTA (P < 0.008), but nearly identical with 3D-TEE on diastolic phase cCTA. Semi-automated and manual cCTA annulus measurements were highly correlated in systole (r = 0.94) and diastole (r = 0.93), although the semi-automated annular area measured 11-30 mm² greater than manual planimetry. Of note, the 95% limits of agreement in our Bland-Altman analysis suggest that the variability in annular area estimates for individual patients between cCTA and 3D-TEE (-100.9 to 99.6 mm²), as well as the variability between manual and automated measurements with cCTA (-105.9 to 45.2 mm²), may be sufficient to alter size selection for an aortic prosthesis.

Conclusions: Although all cCTA measurements are highly correlated with measurements by 3D-TEE, diastolic phase cCTA measurements tend to be closer to standard mid-systolic 3D-TEE measurements. Semi-automated measurement of the aortic annulus with cCTA is highly correlated with manual planimetry. Nonetheless, annular contours derived by semi-automated software should be visually inspected, as the variability in area estimates for individual cases between manual and automated measurements may alter the sizing of an aortic prosthesis.