Accuracy of three-dimensional computational modeling in prediction of the dynamic neo left ventricular outflow tract with transcatheter mitral valve replacement

Background: Transcatheter mitral valve replacement (TMVR) offers a valuable treatment option for inoperable patients suffering from a degenerated mitral valve after previous ring annuloplasty. Dynamic obstruction of the left ventricular outflow tract(LVOT) is a procedural risk with detrimental consequences and can be estimated upfront using a multi-slice computed tomography(MSCT) derived 3D computational model(3DCM). This study explored the accuracy of pre-procedural neo-LVOT prediction in TMVR using 3DCMs of multiple cardiac phases. Methods: We obtained both pre- and post-procedural MSCT scans of a patient who underwent uncomplicated TMVR and derived 3DCMs from each cardiac phase. Virtual implantations of the deployed valve were performed and neo-LVOT dimensions were semi-automatically calculated in the pre-procedural models and matched with the post-procedural models. Predicted and post-procedural neo-LVOTs were compared between 3DCMs. Results: From cardiac phases 20–70%, 11 matched 3DCM pairs were generated. The mean difference between predicted and post-TMVR neo-LVOT area was 3 ± 23 mm². The intra-class correlation coefficient for absolute agreement between predicted and post-procedural neo-LVOT area was 0.86 (95%CI 0.56–0.96, p < 0.001). Conclusion: 3DCMs could accurately predict post-TMVR neo-LVOT dimensions in a patient with a pre-existing mitral annular ring. Prospective research is warranted to demonstrate the accuracy of these models in larger samples and different mitral annular phenotypes.