Validating a prediction modeling tool for left ventricular outflow tract (LVOT) obstruction after transcatheter mitral valve replacement (TMVR)

Objective: Demonstrate proof-of-concept validation of a computed tomography (CT) computer-aided design prediction modeling tool to identify patients at risk for left ventricular outflow tract (LVOT) obstruction in transcatheter mitral valve replacement (TMVR). Background: LVOT obstruction is a significant and even fatal consequence of TMVR. Methods: From August 2013 to August 2017, 38 patients in 5 centers underwent TMVR with compassionate use of balloon-expandable valves for severe mitral valve dysfunction because of degenerative surgical mitral ring, bioprosthesis, or severe native mitral stenosis from to severe mitral annular calcification. All patients had preprocedural CT scans performed for anatomic screening, intraprocedural TEE and invasive hemodynamics performed. Preprocedural prediction modeling was performed utilizing computer-aided design (CAD) of the neo-LVOT post-TMVR. Post-TMVR CT scans were obtained and compared to pre-TMVR LVOT modeling datasets for validation. Results: All patients underwent successful TMVR without device embolization. Seven of the 38 patients experienced LVOT obstruction, defined as an increase of ≥10 mmHg LVOT peak gradient post-TMVR. Anatomic screening using CT was validated in 20/38 patients as preprocedural predicted neo-LVOT surface area correlated well with post-TMVR measurements (R² = 0.8169, P < 0.0001). A receiver operating curve curve found a predicted neo-LVOT surface area of ≤ 189.4 mm² to have 100% sensitivity and 96.8% specificity for predicting TMVR-induced LVOT obstruction. Conclusion: CAD design and CT postprocessing are indispensable tools in predicting LVOT obstruction and necessary for anatomic screening in percutaneous TMVR.