Mouse LV 3D motion and strain analysis using tagged MRI

This paper presents a framework to reconstruct mouse left ventricular motion based on tagged MRI using nonlinear Laplacian deformable models, which perform better in terms of accuracy and efficiency. Based on the deformation results, we analyze the LV 3D motion and strain of the myocardial wall to depict the contractile function of the heart. In our framework, the 2D tagging lines are extracted using active contour models. The 3D control points are then generated from the intersections of the tagging lines, and a 3D meshless model is built based on the sparse 2D contours. Finally, the initial model is driven to deform by the control points using 3D meshless deformable model with nonlinear Laplacian kernel. This method is validated using the in vivo MRI tagging data from the mouse heart. The results show that the proposed method effectively quantifies the myocardial strain distribution of the LV model in 3D, which has the potential to increase accuracy in detecting various kinds of heart diseases.