ARRS 2022 Abstracts


1867. CT-Based Left and Right Ventricular Longitudinal Strain Provides Superior Prediction of Mortality Compared to EF In TAVR
Authors * Denotes Presenting Author
  1. Rock Savage *; Medical University of South Carolina
  2. Gilberto Aquino; Medical University of South Carolina
  3. Tilman Emrich; Medical University of South Carolina
  4. Logan Fitzpatrick; Medical University of South Carolina
  5. Akos Varga-Szemes; Medical University of South Carolina
  6. U. Joseph Schoepf; Medical University of South Carolina
This study aims to assess the predictive value of pre-procedural, multiphasic cardiac CT-based left (LV) and right ventricular (RV) long axis strain (LAS) for mortality in patients undergoing transcatheter aortic valve replacement (TAVR).

Materials and Methods:
One hundred and seventy-five patients (78.8 ± 8.9 years) with severe aortic stenosis who had previously undergone TAVR were retrospectively evaluated in this 24-months follow-up study. Mortality was the primary endpoint. Pre-TAVR CTAs with 10-100% phases were analyzed in 4-chamber view. For LV-LAS, the distance between the epicardial border of the LV apex and the middle of the mitral plane was measured in both endsystole and enddiastole. For RV-LAS, the distance between the LV apex and the middle of the tricuspid plane was measured. LAS was then computed by LAS= ([length end systole - length end diastole)] / length end diastole) x 100 (1). ROC curve analysis was performed to acquire the optimal cutoff for LV-LAS and RV-LAS and these were then divided into normal or abnormal LAS. Survival analysis was done to obtain hazard ratios (HR) and Harrell’s c-index was used to evaluate risk models. LV-LAS and RV-LAS were each adjusted for the Society of Thoracic Surgeons (STS) score for mortality currently used to risk stratify TAVR patients.

Median follow-up was 21 months. There were 38 deaths (22%). The optimal cutoffs for LV-LAS and RV-LAS were -9.0% and -11.4%, respectively. Abnormal LV-LAS (HR 2.8; 95% CI 1.2-6.6; p=0.022) and RV-LAS (HR 2.7; 1.3-5.6; p=0.008) were independently associated with mortality after adjustment for variables predictive on univariable analysis and LV and RV ejection fraction. When adjusted for the STS risk score, LV-LAS (HR 3.4; 1.5-7.7; p=0.004) and RV-LAS (HR 2.4; 1.2-4.8; p=0.016) remained significantly predictive. C-indices for LV-LAS and RV-LAS were 0.624 and 0.640, respectively. Adding LV-LAS and RV-LAS to the STS score significantly improved its c-index from 0.636 to 0.684 for prediction of mortality.

Pre-TAVR, LV-LAS and RV-LAS independently predict post-TAVR mortality beyond ventricular ejection fraction. CTA-based left and right ventricular longitudinal strain have the potential to improve risk stratification and post-treatment surveillance in patients undergoing TAVR.