Abdominal organ segmentation is highly desirable but difficult, due to large differences between patients and to overlapping grey-scale values of the various tissue types. The first step in automating this process is to cluster together the pixels within each organ or tissue type. We propose to form images based on second-order statistical texture transforms (Haralick transforms) of a CT or MRI scan. The original scan plus the suite of texture transforms are then input into a Hopfield neural network (HNN). The network is constructed to solve an optimization problem, where the best solution is the minima of a Lyapunov energy function. On a sample abdominal CT scan, this process successfully clustered 79-100% of the pixels of seven abdominal organs. It is envisioned that this is the first step to automate segmentation. Active contouring (e.g., SNAKE's) or a back-propagation neural network can then be used to assign names to the clusters and fill in the incorrectly clustered pixels.

A triple-detector, multipinhole SPECT system was optimally configured to perform simultaneous 201 Tl (stress)/ 99m Tc (rest) myocardial perfusion imaging (MPI) using a protocol that permitted direct diagnostic comparison of this multipinhole SPECT system with conventional rotational SPECT. Methods: Both the rotational and the multipinhole SPECT systems used the same model g-detectors. The 2 systems were applied in tandem to study 26 patients with documented coronary status. Visual image evaluation of the MPI together with quantitative analysis using circumferential profile curves (CPC) were used for interpretation of stress/rest myocardial flow differences. A dual-peak attenuation compensation (DPAC) technique was applied to the stress 201 Tl multipinhole SPECT images by weighted combination of the images from the upper and lower peaks. Results: Detection of myocardial infarction by location and extent correlated closely, and correlation of differential flow changes between stress and rest indicated similar accuracy in terms of location and extent of myocardial blood flow differences as well. In addition, the application of DPAC clarified the multipinhole stress 201 Tl images through reduced background and increased statistics and also improved the relative superposition of the normalized CPC, especially for the inferior and more basal reconstructed regions. Conclusion: The prototype 3-detector multipinhole SPECT system achieved diagnostic results comparable to those for rotational SPECT and required only a single image-acquisition session to generate stress/rest MPI and 16-segment poststress gated studies. This reduction in acquisition time significantly improves productivity without compromising diagnostic accuracy. In addition, DPAC is a useful adjunct to the multipinhole SPECT modality because it improves both the visual clarity of the stress images and the stress/rest quantitative comparability.

A relatively simple, bedside, radionuclide technique has been developed to measure right ventricular ejection fraction. This technique uses a collimated scintillation probe (5 x 5 cm sodium iodide crystal) and 113mIndium injected into the superior vena cava to record a right ventricular time-activity curve. The radionuclide method was validated in 34 men (14 normals, 20 with coronary artery disease) with biplane right ventriculography (r= 0 82). Using this radionuclide method right ventricular ejection fraction was measured in 26 men (average age 51 years) with an acute transmural myocardial infarction. Right ventricular ejection fraction was initially depressed (0 40±0 02; mean±SEM; normal 0-57±0-01) in all of 11 men with an acute inferior infarction but returned to normal by the third day in 10 of them (0-58±0.01; NS). In the 15 men with an acute anterior infarction the average right ventricular ejection fraction was normal initially (0.54+ 0*01; NS) and individually 10 of 15 had a normal ejectionfraction. Left ventricular ejectionfraction was initially depressed in allpatients and only 4 of 34 (12%) had returned to normal at the third day. These results suggest that right ventricular ejection fraction is regularly depressed in patients with an acute inferior inifarction but normal in those with an anterior infarction. Right ventricular performance rapidly improves after inferior infarction; whereas less improvement occurs in left ventricular ejection fraction.Left ventricular ejection fraction has been found to artery disease or acute myocardial infarction. be regularly depressed in acute myocardial infarcWe had previously developed a radionuclide tion both early (Kostuk et al., 1973) and late after method for estimating left ventricular ejection infarction (Rackley et al., 1970; Hamilton et al., fraction which uses a portable scintillation probe 1972; Stewart et al., 1974). Kostuk and associ-(Steele et al., 1974). The purpose of the present ates (1973) have shown that a relation exists study was to develop a radionuclide method for between left ventricular ejection fraction and myo-estimating right ventricular ejection fraction and to cardial infarct size, as measured with serial deter-apply this technique to the serial study of patients minations of serum creatine kinase activity. Though with acute myocardial infarction. a disparity between right and left ventricular filling pressures in acute myocardial infarction Subjects and methods has been appreciated (Forrester et al., 1971) and infarction of the right ventricle has been recog-Right ventricular ejection fraction was measured nised both clinically (Cohn et al., 1974; Rotman from radiocardiograms obtained using a colliet al., 1974) and pathologically (Wartman and mated scintillation probe by a modification of a Hellerstein, 1948;Wade, 1959;Erhardt, 1974), radionuclide method for estimating left ventricular relatively little attention has been paid to right ejection fraction (Steele et al., 1974). To measure ventricular performance in either chronic co...

No abstract

No abstract