Evaluation of cardiac valvular regurgitant lesions by cardiac MRI sequences: comparison of a four valve semi-quantitative vs. quantitative approach

Abstract

Cardiac magnetic resonance (CMR) imaging generally allows a more accurate and valid quantification of cardiac function, mass and regurgitant volumes than echocardiography. Although recent technological advancements in CMR have made the evaluation of cardiac valves more reliable, no studies have yet been conducted to compare semi-quantitative grading (SQG) using CMR steady-state free precession (SSFP) sequences with quantitative grading (QG) based on stroke volumes and phase-velocity mapping (PVM). It is proposed that the SQG of cardiac valvular regurgitations based on CMR SSFP sequences is feasible, and highly correlative with standard CMR QG methods. CMR data obtained between January 2007 and December 2011 was evaluated prospectively for valvular regurgitant lesions. Patients were included if they had right and left ventricular volumetrics based on CMR SSFP sequences and PVM across the aortic and pulmonic valves with reported regurgitant volumes and fractions. Patients were excluded if they had prosthetic valves, cardiac arrhythmias and intra-cardiac shunts. Regurgitant lesions were semi-quantitatively (visually) graded on a standard scale of 0 to 4 (trace, mild, moderate, moderate to severe, and severe) and compared with quantitative regurgitant fractions. Correlations were evaluated by Spearman's rho formula, and kappa for intra- and inter-observer variabilities were obtained on 30% of the study sample. A total of 97 patients (58 males, 39 females; average age 55 +/- 18 years) representing 134 valvular regurgitations [mitral (MR), aortic (AR), tricuspid (TR), and pulmonary (PR)] were analyzed by semiquantitative and quantitative methods. The regurgitant lesions included 44 mitral, 50 aortic, 29 tricuspid, and 11 pulmonary. The correlation between SQR versus QG yielded the following results: 0.67, p < 0.001 (MR, r = 0.66, p < 0.001; AR, r = 0.68, p < 0.002; TR, r = 0.68, p = 0.001; PR, r = 0.70, p = 0.017). The results for QG versus SQG accounting for clinically significant differences of +/- 1 grade for the group were as follows: 0.95, p < 0.001 (MR, r = 0.91, p < 0.001; AR, r = 0.96, p < 0.001; TR, r = 0.99, p < 0.001; PR, r = 0.93, p < 0.001). No discrepancy between surgical regurgitation was present (3+ versus 4+). Weighted kappa results were 0.67 and 0.61 for intra- and inter-observer variabilities. The visual assessment of cardiac regurgitant lesions is reliable, accurate and reproducible when compared to formal quantitative analysis via CMR. This confirms a robust role for CMR in assessing regurgitant lesions, particularly for surgical decision-making. These results were applicable to patients in sinus rhythm at the time of scanning.