Session PB7.6

Cardiac Function Estimation Using MSCT: A Comparison to Speckle Tracking Imaging

A Simon*, R Delaunay, C Leclercq, E Donal, M Garreau

Université de Rennes
Rennes, France

The goal of this work is to evaluate the capabilities of MSCT (Multi-Slice Computed Tomography) imaging to assess cardiac function and especially the most delayed ventricular segment in left ventricular dyssynchrony. This work is part of the IMOP project (Imaging for Optimisation of biventricular Pacing) which purpose is to define a CRT optimization method based on the fusion of mechanical, electrical and anatomical data. MSCT dynamic acquisition provides datasets representing the whole cardiac cycle with a high spatial resolution, enabling to describe the left ventricle both anatomically and functionally. We have previously proposed a method to estimate cardiac motion from these datasets [Simon, CinC2004]. Based on a multi-resolution surface matching, this method provides a displacements field describing the motion of the left endocardium along the whole cardiac cycle. This method enables to compute functional parameters and to display them using compact representation modes such as bull-eye representation (based on the classic 17 cardiac anatomical segments). For instance, radial and longitudinal motion components are represented and allow the extraction of specific parameters such as maximal displacement (from the first time-instant of the sequence), time-instant corresponding to this maximal displacement and total displacement (path length). In order to evaluate the resulting estimated displacements, a comparison to US Speckle Tracking Imaging (STI) has been carried out using a three steps process: (i) because the STI is in two dimensions, the manual selection of the echographic plane is carried out in the 3D CT data, followed by the extraction of the LV points in that plane and by the decomposition of the estimated displacements in CT according to the echographic axis; (ii) in order to obtain the same temporal resolution, a cubic spline interpolation is applied to the CT motion components followed by a baseline correction; (iii) a rigid spatial registration, based on an Iterative Closest Point algorithm, is used to associate, to each US point, the corresponding CT points. In order to compare the estimated displacements between the two modalities, spatio-temporal maps (classic US representation) are used. Results are presented for two patients, using apical 4- and 2-chambers long-axis views, and short-axis view. They highlight a similitude in transverse and especially in radial displacements while longitudinal displacements do not show a good coherence, which can be explained by the aperture phenomenon. These first results give very promising results for MSCT imaging to provide significant clinical parameters for dyssynchrony characterization.

(Abstract Control Number: 211)