Optimization of echocardiographic quantification of aortic regurgitation against CMR: novel algorithm development and...

Cardiac magnetic resonance (CMR) compliments transthoracic echocardiography (TTE) for heart valve evaluation, however TTE remains more widely available. We sought to optimize TTE quantification of significant aortic regurgitation (AR) by developing TTE-based algorithms to identify CMR-defined severe AR.

Patients with moderate-to-severe AR undergoing both TTE and CMR within 3-months were studied. A historical cohort 2006-18 (n = 193) was used to derive TTE-based decision tree regression algorithms to best identify severe AR based on holodiastolic flow reversal (HDR) using CMR, then validated in a prospective AR cohort (n = 97) during 2019-21.Mean AR regurgitant volumes, fractions, and proportions with HDR by TTE/CMR were 48/31 mL, 41/25%, and 43%/27% for the historical derivation cohort and 51/37 mL, 47/29%, and 54%/41% for the prospective validation cohort. Decision-tree analyses found regurgitant volume ≥ 45 mL and left ventricular end-diastolic volume index (LVEDVi) ≥ 93 mL/m2 by TTE to best identify CMR-derived severe AR. Areas under curves (95%CIs) of the novel algorithms (proximal isovelocity surface area and Doppler methods) compared with current guidelines criteria for detecting CMR-derived severe AR were 0.80 (0.71-0.88) and 0.74 (0.65-0.83) vs. 0.72 (0.63-0.81) in the derivation cohort, and 0.76 (0.66-0.87) and 0.71 (0.61-0.82) vs. 0.58 (0.46-0.70) in the validation cohort; and for predicting left ventricular remodelling where follow-up TTE wad available were 0.65 (0.58-0.73) and 0.62 (0.55-0.70) vs. 0.53 (0.45-0.61), respectively.

Novel TTEs algorithm increased TTE accuracy of identifying significant AR defined by CMR especially in the prospective cohort, compared to the current guidelines criteria, and was able to modestly discriminate LV remodelling.