PhD Scientific Days 2026

Cardiovascular Medicine and Research 2.

A Predictive Model Based on Clinical and Echocardiographic Parameters to Assess LVOT-related Bias in Aortic Stenosis Severity Among TAVR Candidates

Name of the presenter

Mohácsi, Fanni

Institute/workplace of the presenter

Gottsegen National Cardiovascular Center

Authors

Fanni Mohácsi MD¹, Franciska Benyó MD¹, Zsolt Szedlacsek MD¹, Anita Káposzta MD², Mercédesz Matus MD², Abdelkrim Ahres MD, Phd², Andréka Péter MD, Phd², Márton Kolossváry MD, Phd², Dénes Mónika MD, Phd²
1: Gottsegen National Cardiovascular Center; Semmelweis University Doctoral College
2: Gottsegen National Cardiovascular Center

Text of the abstract

Introduction:Echocardiography systematically underestimates the left ventricular outflow tract (LVOT) area, which consequently leads to an underestimation of the aortic valve area (AVA).
Aims:This study aimed to identify predictors of measurement discrepancy between echocardiographic and computed tomography (CT)-derived LVOT area and to develop a model capable of estimating CT-based AVA values (hybrid AVA, defined using CT-derived LVOT measurements) based solely on clinical and echocardiographic parameters.
Methods:We retrospectively analyzed consecutive patients with severe aortic stenosis between January 1, 2020, and September 1, 2024, to identify predictors of relative measurement inaccuracy using linear regression analysis. Based on these findings, we developed a predictive model utilizing exclusively clinical and echocardiographic variables to estimate CT-derived hybrid AVA. The model was subsequently validated in an independent cohort of consecutive patients enrolled between September 2, 2024, and September 1, 2025.
Results:A total of 1,367 patients were included (mean age 79.3 ± 6.3 years; 725 women).Echocardiographic assessment underestimated LVOT area compared to CT by a mean of −162 ± 98.2 mm², resulting in a mean AVA difference of −0.29 ± 0.18 cm². This discrepancy led to the reclassification of 31% of patients when applying a hybrid AVA threshold of ≥1.0 cm². Factors associated with reduced relative measurement error included older age, female sex, larger LVOT diameter, higher aortic valve Vmax, and higher Doppler index. In contrast, greater height and increased end-systolic and end-diastolic volumes were associated with increased measurement error. The predictive model demonstrated strong correlation with CT-derived hybrid AVA values (r = 0.82, p < 0.001) and maintained its performance in the validation cohort (n = 448; r = 0.85, p < 0.001).
Conclusion:Echocardiography significantly underestimates LVOT area relative to CT, leading to the reclassification of nearly one-third of patients with severe aortic stenosis. The proposed model, based solely on clinical and echocardiographic parameters, provides a reliable approach for estimating CT-derived hybrid AVA and may facilitate CT-independent assessment in selected patients.
Márton Kolossváry was supported by the János Bolyai Scholarship and NRDI‑funded grants PD 147269 and Excellence 151118.