Abstract

Theory and previous studies showed that KLT (an application of principal component transform for imaging) can be use for analysis of cardiac function. This paper presents the results of our studies concerning the applications of KLT for images smoothing, quantification of myocardial contraction, and improvement of inter-observer reproducibility in cardiac imaging. The paper also describes the use of 4D cardiac phantom to quantify Karhunen-Loeve images.

No abstract

No abstract

IntroductionAcquisition of gated single-photon emission tomography of the myocardium (gSPECT) is routinely used in clinical practice to provide information on myocardial perfusion and contractile function [1]. After reconstruction, a rigorous left ventricular contouring on each cycle image and axis of the heart is required. We developed a program for automatic contour detection; it was applied without knowledge on counts statistics, chosen axis, tracer distribution or artefacts. 2.Material and methods Patients and image acquisitionThis study was performed on 110 explorations at stress (8 images per cardiac cycle) and concerned the central scans of the 3 heart axes and the summed image. 99mTc-gSPECT studies were performed in supine patients. The exercise test consisted of physical exercise combined with pharmacological vasodilatation. After exercise stress test, a dose of 15 MBq/kg of 99mTc was administered intravenously. A rotating dual-head gamma camera was used, equipped with low-energy high-resolution collimators; the energy window was centered on 140 keV ± 15%. The system detectors were at 90°. The gamma camera was rotated through a 180° arc in a contour orbit around the patient's chest from 45° RAO at 6° increments for 40-60 s each (32 views), 8 time bins [2]. 1.33 zoom was applied. Data were stored in a 64x64 matrix; the raw scintigraphic data were preprocessed with a backprojection algorithm and a Butterworth filter (frequency 0.25, range 5) without attenuation correction was applied. Then, the series was smoothed and denoised using Karhunen-Loeve transform; we used the 3 first KLT images for reconstruction [3]. The processing was done using MATLAB version 5.3 (The MathWorks, inc.) on standard PC (PIII-733, 256Mb RAM) PropertiesThe following topographical properties of the myocardial reconstructed images are involved: -The heart takes more than a half of the image surface; with many pixels, a histogram analysis can be used. -The resulting pictures are denoised, and the "useful" intensities are 5 to 10 times higher than the noise, both in rest and stress studies. This assumption justifies the thresholding.-Polar representation of the contours is applicable because of the elliptical form of the left ventricle and its central position on the images. -The artefacts correspond to some tissues with abnormal fixation of the pharmaceutical tracer, and they are outside of the cardiac zone. Morphological operations (erosion, dilation, skeleton) are therefore possible.Skeletonization is a process for reducing foreground regions in a binary image to a skeletal remnant that largely preserves the extent and connectivity of the original region while eliminating most of the original foreground pixels.The skeleton can be produced in two ways. The first is to use some kind of morphological thinning that sequentially erodes pixels from the boundary (while preserving the end points of line segments) until no more thinning is possible; at this point what is left approximates the skeleton. The alternative method is to...