C.P. Melo

Estimation of myocardial kinetic energy in gated SPECT images

Conventional methods for the analysis of left ventricle (LV) wall motion are based on 2D images with many hardly fulfilled assumptions. The methods in use today suffer from several problems as they do not take into account three-dimensional (3D) information that can be provided by image modalities such as nuclear medicine and MRI. This work describes a method to quantify 3D LV motion by a series of 3D velocity vector fields that are computed automatically for each voxel on the sequence of 16 cardiac volumes. Using the velocity it is possible to estimate the kinetic energy for each voxel. The integration of this quantity over the LV muscle gives an estimate of the left ventricle wall kinetic energy. A curve representing the myocardial kinetic energy can be plotted and used as an indicator of cardiac conditions. The proposed method was applied to the study of normal and abnormal hearts. Results have shown that the curve pattern for hearts with infarct differs substantially from a normal heart.

3-D analysis of left ventricle dynamics

Describes a method which quantifies 3D LV motion by means of the optical flow technique extended to the voxel space. Motion is represented by a 3D velocity vector field which is computed for each voxel on the series of cardiac volumes. The algorithm developed has two steps: (1) the spatial and temporal derivatives of the 3D image brightness are computed, after the cardiac volumes have been convolved with a symmetrical Gaussian function filter; (2) The optical flow is computed as the solution to a linear algebraic system of equations whose coefficients are determined by the derivatives that were computed in (1). The velocity vector field estimated from a sequence of 3D images has been computed for normal and infarcted hearts obtained at InCors Nuclear Medicine Department.<<ETX>>

Wigner-Ville distribution applied to cardiac motion estimation

Some heart abnormalities like ischemia and infarction of myocardium can produce significant changes in cardiac wall motion and decrease severely the cardiac function. For this reason the cardiac motion estimation and quantification have been the focus of many investigations. The aim of this work is to study cardiac motion by means a dense velocity vector field based on the computation of the Wigner-Ville distribution, a common transformation used in the spatiotemporal-frequency (STF) analysis. The major motivation for considering the use of STF comes from observations in mammalian vision. This paper describes a method to obtain a 3D-frequency spectrum, based on the computation of the Wigner-Ville distribution, and a procedure to determine, automatically, the velocity of pixels in a temporal series of images. The method is also applied to synthetic images and real images obtained from gated-SPECT.

Segmentation of left ventricle myocardium in MRI using endocardium layer expansion model

We present a volumetric approach for segmentation of left ventricle (LV) myocardium in MRI based on fuzzy connectedness and incorporation of voxel with minimum cost. It is a two-step technique: first, segmentation of LV cavity and then expansion of muscle layers keeping roughly the endocardium morphology. The LV cavity is extracted applying competitive fuzzy connectedness between cavity and myocardium. A small region sample of each object is provided by the user to characterize their membership function that is based on likelihood. From the endocardium surface, the connected voxels with minimum insertion cost and below a calculated threshold are incorporated in an orderly fashion. The cost function encompasses local gradient, local texture and shape.

Myocardial motion estimation in gated-MRI using optical flow refined with scale-space

Although MR imaging techniques are able to show cardiac motion qualitatively, no information is provided about motion of the tissue within the epicardial and endocardial boundaries using standard procedures. This work describes a method to quantify the myocardial motion from 3D MR images based on Optical Flow technique. To improve derivatives estimation, a fundamental step in any Optical Flow implementation, the authors applied the one parameter Gaussian derivative operator, obtained from the scale-space approach. The method quantifies the 3D LV motion by a series of 3D velocity vector fields and this vectorial information is used to estimate the scalar parameter. Myocardial kinetic energy. The evolution of the kinetic energy parameter during the cardiac cycle may represent the cardiac conditions.

The assessment of myocardial motion in gated-SPECT using a multiresolution technique

The assessment of the left ventricular (LV) motion in gated SPECT images is routinely used to identify myocardial ischemia. The precise evaluation of LV motion changes remains a largely qualitative process due to difficulties in quantifying endocardial border position and regional wall motion. In a previous work the authors described a method to estimate motion in image sequences based on the computation of the Optical Flow (OF). A common basic problem in estimating OF concerns how to compute derivative approximations from discrete data. The intrinsic ill-posedness of numerical differentiation can be transformed into a well-posed problem by using filters in the form of scaled convolution operators. Using this approach, the computing of OF can be improved and the information obtained used to measure other physical quantities in a sequence of images. In this work the authors propose and apply to a set of volunteers the measurement of the kinetic energy index, based on OF estimation.

Cardiac motion estimation using a spatiotemporal-frequency approach

The complex deformation that occurs during the contraction and relaxation of the left ventricle (LV) has forced investigators to make major simplifying assumptions to interpret LV motion. It is frequent to model the LV as some particular shape or assume symmetry, such as the elliptical chamber, in order to study ventricular changes of shape. The proper quantification of the cardiac motion still remains an open and challenging research problem. In the paper we describe a method for motion detection based on a spatiotemporal-frequency approach. This method has been applied to compute the optical flow field obtained from a series of gated-SPECT images. Optical flow is a dense velocity vector field that represents the motion of brightness patterns between successive image frames. It has been widely used in computer vision to provide important cues for image and vision analysis.

Assessment of 3D myocardium segmentation in MRI using endocardium layer expansion model

In this work the authors present an assessment of 3D Left Ventricle Myocardium (LV) segmentation based on the Layer Expansion model. The evaluation was carried out using simulated myocardium with noise and MR images of the heart. The segmentation approach consists of first determining the LV cavity using competitive fuzzy connectedness between cavity and myocardium region. Myocardium is segmented using a layer expansion model after the determination of endocardium surface. Connected voxels with insertion cost below a threshold are incorporated in an orderly fashion, emulating a balloon expansion. The cost function encompasses local texture, local gradient and global shape information. The segmentation error (percentage of misclassified voxels) for the phantoms were below 11.1% for cavity and 19.7% for myocardium. The proposed approach can be used for higher dimensional data, such as dynamic 3D scenes.

Quantification of myocardial kinematics in perfusion images from gated SPECT

Left ventricle contractile abnormalities can be an important manifestation of coronary artery disease. The abnormalities in wall motion may represent ischemia or infarction of myocardium. This kind of disease is often regional. It is a fundamental goal of many cardiac imaging modalities and image analysis methods to measure the regional function of the left ventricle (LV). In this work we describe a method to quantify the 3D myocardial kinetic energy (ke) via polar map or ke bulls eye obtained from gated SPECT images. For each slice a maximum-energy profile is generated by dividing the LV into sectors and taking the highest voxel value within a sector. The map profiles are plotted as concentric circles onto a polar map. Specific sectors show the regional variations of the ke on the LV wall. The current research includes a comparison of ke bulls eye patterns for normal and abnormal hearts as well as the discrimination capability for the proposed method.

A microcomputer system for quantitative analysis of neural activity in the regulation of the cardiovascular system

A microcomputer-based system designed to verify the aspects involved in the regulation of the cardiovascular function is described. Neural discharge measurements are associated with changes in heart rate and arterial blood pressure on a beat-by-beat basis. The system uses an IBM-PC with analog-to-digital converter and a dedicated device to study a point process. The software has been applied in research protocols using rats to clarify different aspects of neural control of the cardiovascular system. The relationship of pulse pressure and renal sympathetic activity in conscious rats has been studied. Beat-by-beat analysis of renal sympathetic activity in sinoartic denervated rats has shown that in neurogenic hypertension there is not only an increase of neural discharges but also a change in the pattern of renal sympathetic activity correlated with arterial pressure.<<ETX>>