1348772 - Comparison of Cardiac CT and MRA Measurements for Determination of Candidacy for Transcatheter Pulmonary Valve Implantation in Native or Post-surgical Right Ventricular Outflow Tracts

Background: Although transcatheter pulmonary valve replacement (TPVR) has been available for over two decades, utilization has been limited in patients with a dilated native or surgically repaired right ventricular outflow tract (RVOT).  The Harmony Transcatheter Pulmonary Valve™ and Edwards Sapien 3 Transcatheter Pulmonary Valve with Alterra Adaptive Prestent™ are new transcatheter systems which may be implanted in patients with native or surgically repaired RVOTs and may overcome this limitation in many cases. Many candidates for pulmonary valve replacement (PVR) with native or surgically repaired RVOTs will have had CMR with magnetic resonance angiography (MRA) to determine clinical indication for PVR; yet, industry guidance requires all patients to undergo retrospectively gated CT regardless of prior imaging to inform candidacy for TPVR.  We sought to 1) assess the adequacy of MRA at our center to measure RVOT dimensions similar to those measured on CT; and, 2) evaluate agreement between CT and MRA dimensions.  We hypothesized that CT and MRA dimensions would have a high level of agreement.

Methods: Patients who underwent retrospectively gated CT and MRA within one year at our center were identified.  MRA at our center is performed using a 3D respiratory-navigated, inversion recovery prepared sequence with gradient echo readout, typically gated to a quiescent period of cardiac motion during systole to generate a high-resolution 1.2-1.4 mm isotropic dataset.  Areas within the proximal RVOT, mid pulmonary artery, and distal pulmonary artery and RVOT lengths were measured by a single user using multiplanar reformatting on systolic CT images and then on MRA images after 24 hours had elapsed (Figure 1).  Areas were used to derive radiuses.  The user was blinded to prior CT measurements but not the imaging plane.  Measurement correlation and agreement were assessed using linear regression and Bland-Altman analyses.  We defined a priori a mean measurement difference of < 5% as sufficient to consider the measurements equivalent.

Results: Six patients were identified.  All MRAs were acceptable for performance of measurements.  There was a significant correlation between all CT and MRA measurements by linear regression analysis (p < 0.05 for all, Table 1).  The mean bias for each measurement was < 5% different than the mean of the CT and MRA measures (Figure 2).

Conclusion: Performing measurements used for determination of TPVR candidacy is feasible using cardiac MRA.  As MRAs are frequently obtained in this patient population for routine clinical care, using these data to guide TPVR may spare patients time, expense, and radiation associated with additional cardiac CT.  Further efforts are needed to explore feasibility of measurement with MRA gating to diastole, evaluate measurement equivalency among larger groups of patients, and partner with industry to explore use of MRA datasets for device fit-testing.