Using 4D Flow MRI to Study Left Ventricle Failure and Hemodynamic Biomarkers in Patients Undergoing Transcatheter Aortic Valve Replacement

What this Trial Is About

Researchers are investigating advanced MRI-based techniques to improve the automatic assessment of left ventricular LV dysfunction in patients, particularly those with severe aortic stenosis undergoing transcatheter aortic valve replacement TAVR. The study aims to develop a deep learning model to automatically segment the LV endocardial contours from cardiac MRI images, as well as to create non-invasive methods for reconstructing pressure-volume loops and evaluating hemodynamic forces in the heart. This research includes both retrospective and prospective study arms to support model training, clinical validation, and translation into clinical practice. The studys retrospective phase involves collecting and anonymizing cardiac MRI datasets to train and test a convolutional neural network for LV segmentation. The prospective phase enrolls patients scheduled for TAVR, who will undergo non-contrast 4D Flow MRI and routine invasive hemodynamic measurements during their procedure. MRI scans include standard cine sequences and 4D Flow acquisition, performed without contrast agents. Data from these imaging and hemodynamic assessments will be processed using specialized software to analyze LV blood flow and pressure dynamics, comparing non-invasive measurements with invasive catheter data. Participants will undergo MRI scans and have hemodynamic data collected during their routine clinical care without additional interventions or risks. Researchers will evaluate the accuracy of the automatic LV segmentation and compare MRI-based pressure-volume loop parameters and hemodynamic force measurements with invasive standards. The study also plans to develop a standalone, user-friendly software application for clinical use based on the findings. There are no extra follow-up visits, and participant involvement is limited to imaging, data collection, and analysis during the study period.

Phenotyping Left Ventricle Failure With Hemodynamic Biomarkers From 4D Flow Magnetic Resonance Imaging