Deformable cardiac surface tracking by adaptive estimation algorithms

Abstract

This study presents a particle filter based framework to track cardiac surface from a time sequence of single magnetic resonance imaging (MRI) slices with the future goal of utilizing the presented framework for interventional cardiovascular magnetic resonance procedures, which rely on the accurate and online tracking of the cardiac surface from MRI data. The framework exploits a low-order parametric deformable model of the cardiac surface. A stochastic dynamic system represents the cardiac surface motion. Deformable models are employed to introduce shape prior to control the degree of the deformations. Adaptive filters are used to model complex cardiac motion in the dynamic model of the system. Particle filters are utilized to recursively estimate the current state of the system over time. The proposed method is applied to recover biventricular deformations and validated with a numerical phantom and multiple real cardiac MRI datasets. The algorithm is evaluated with multiple experiments using fixed and varying image slice planes at each time step. For the real cardiac MRI datasets, the average root-mean-square tracking errors of 2.61 mm and 3.42 mm are reported respectively for the fixed and varying image slice planes. This work serves as a proof-of-concept study for modeling and tracking the cardiac surface deformations via a low-order probabilistic model with the future goal of utilizing this method for the targeted interventional cardiac procedures under MR image guidance. For the real cardiac MRI datasets, the presented method was able to track the points-of-interests located on different sections of the cardiac surface within a precision of 3 pixels. The analyses show that the use of deformable cardiac surface tracking algorithm can pave the way for performing precise targeted intracardiac ablation procedures under MRI guidance. The main contributions of this work are twofold. First, it presents a framework for the tracking of whole cardiac surface from a time sequence of single image slices. Second, it employs adaptive filters to incorporate motion information in the tracking of nonrigid cardiac surface motion for temporal coherence.

Figure 1

(a) The intersection of the biventricular cardiac model and three MR slice planes corresponding to the basal, mid-cavity, and apical sections. (b) 16-point LV model. (c) 8-point RV model. The POI locations of the LV and RV segments on the cardiac surface, identified by the indexes in the corresponding 16-point and 8-point models are as follows: Anterior: LV1, LV7, LV13, RV1, RV4, RV7; Anteroseptal: LV2, LV8; Septal: LV14; Inferoseptal: LV3, LV9; Inferior: LV4, LV10, LV15, RV3, RV6, RV8; Lateral: LV16, RV2, RV5; Inferolateral: LV5, LV11; Anterolateral: LV6, LV12.

Figure 2

Shows the tracking results for an anterior POI position for a mid-ventricular slice.

Figure 3

Shows the comparison of global torsion through the cardiac cycle estimated via the Segment software and the proposed algorithm. The mean estimation error is 0.06 rads.

Figure 4

Visualizes the deformable model framework; mapping from the material coordinate domain to the deformable model.

Figure 5

The planar shape of the RV relative to the LV for various septum aspect ratio $s_a$ and septum rotation $s_r$ values.

Figure 6

Shows the intersection of the biventricular deformable model and MRI slice plane.

Figure 7

(a) The real cardiac MRI 2D image slice. (b) The segmented binary slice measurement $z_t$. (c) The predicted binary measurement ${\widehat{z}}_t$, generated from the contours obtained via the intersection of deformable model and image slice plane.

Figure 8

(a) Shows the initial fit of biventricular model to data via nonlinear least squares optimization. (b) Shows the locations of the initial points on the model and the data selected based on 16-point LV and 8-point RV model for a mid-ventricular slice.

Figure 9

Slices from mid-ventricular section of the numerical phantom for (a) undeformed (b) deformed instances.

Figure 10

Shows the segmentations obtained via Segment software respectively for the (a) basal (b) mid-ventricular (c) apical slices.