Use of non-invasive phase contrast magnetic resonance imaging for estimation of atrial septal defect size and morphology: a comparison with transesophageal echo.

BACKGROUND: Transesophageal echocardiography (TEE) is a trusted method of sizing atrial septal defect (ASD) prior to percutaneous closure but is invasive, uncomfortable, and may carry a small risk of morbidity and mortality. Magnetic resonance imaging (MRI) may be useful non-invasive alternative in such patients who refuse or are unable to tolerate TEE and may provide additional information on the shape of the A0SD.
PURPOSE: To validate the accuracy of ASD sizing by MRI compared with TEE.
METHOD: Twelve patients (mean age 30 years; range 11-60 years) scheduled for ASD closure underwent TEE, cine balanced fast field echo MRI (bFFE-MRI) in four-chamber and sagittal views and phase-contrast MRI (PC-MRI) with reconstruction using the two orthogonal planes of T2-weighted images as planning. The average of the three longest measurements for all imaging modalities was calculated for each patient.
RESULTS: Mean maximum ASD length on TEE was 18.8 +/- 4.6 mm, mean length by bFFE-MRI was 20.0 +/- 5.0 mm, and mean length by PC-MRI was 18.3 +/- 3.6 mm. The TEE measurement was significantly correlated with the bFFE-MRI and PC-MRI measurements (Pearson r = 0.69, p = 0.02 and r = 0.59, p = 0.04, respectively). The mean difference between TEE and bFFE-MRI measurements was -1.2mm (95% CI: -3.7, 1.3) and between TEE and PC-MRI was 0.5 mm (95% CI: -1.9, 2.9). Bland-Altman analysis also determined general agreement between both MRI methods and TEE. The ASDs were egg-shaped in two cases, circular in 1 patient and oval in the remaining patients.
CONCLUSION: ASD sizing by MRI using bFFE and phase-contrast protocols correlated well with TEE estimations. PC-MRI provided additional information on ASD shapes and proximity to adjacent structures.