Cardiovascular magnetic resonance and transesophageal echocardiography in patients with prosthetic valve paravalvular leaks: towards an accurate quantification and stratification

Abstract Background Objective assessment of prosthetic paravalvular leak (PVL) is complex and challenging even in transesophageal echocardiography (TEE). Our aim was to assess the value of cardiovascular magnetic resonance (CMR) in quantifying PVL in aortic (AVR) or mitral valve (MVR) replacement. Methods Thirty-one patients (62 ± 15.1 years, 63% males) with a preliminary diagnosis of significant PVL (AVR, n-23; MVR, n = 8) were recruited. The TEE PVL grading was based on the semi-quantitative (SQ) TEE according to the VARC II PVL classification (%PVL: mild 30%). Non-contrast CMR studies were acquired at 1.5 T with a quantitative approach (phase-contrast velocity encoded imaging). The CMR PVL severity was classified according to regurgitant fraction (RF: (1) mild ≤ 20%, (2) moderate 21%–39%, or (3) severe ≥ 40%). Results All patients revealed symptoms of heart failure (71%: New York Heart Association [NYHA] II; 91%: N-terminal pro-B-type natriuretic peptide [NT-proBNP] > 150 pg/ml) and typical cardiovascular disease risk factors. The SQ-TEE results revealed several categories: (1) mild (n = 5; 16%), (2) moderate (n = 21; 67%), and (3) severe (n = 5; 16%) PVL. However, CMR PVL RF reclassified the severity of PVL: (1) mild to moderate (in 80%), (2) moderate to severe (in 47%), and (3) severe to moderate (in 40%). The receiver operating characteristic analysis showed that SQ-TEE and CMR PVL-vol and -RF predicted the upper tertile of NT-proBNP (> 2000 pg/ml) with the best sensitivity for CMR parameters. Conclusion The SQ-TEE showed moderate agreement with CMR and underestimated a considerable number of AVR or MVR-PVL.