Rüdiger Brennecke

Automated Video-Angiocardiographic Image Analysis

This paper describes a hardware-software system for handling the two main groups of cardiological data: a) physiological variables such as voltages, pressures, etc., and b) morphological data, derived from x-ray images or angiocardiograms, such as dimensions, areas, or volumes. We will concentrate in particular on some aspects of automated image processing–i. e., the analysis of the size, shape, and contraction pattern of the ventricles from video-angiocardiograms.

Computerized video-image preprocessing with applications to cardio-angiographic roentgen-image series

We report on the enhancement of video-angiocardiographic image-series by digital preprocessing methods including a newly developed technique of interframe subtraction recording as well as computerized image subtraction, integration, and nonlinear representation techniques. Background suppression and noise reduction obtained through these processes applied to roentgen images from animal experiments are demonstrated. Image-series handling and storage are simplified by combining a new method of digitally formatted videotape recording with conventional digital storage of selected image data in the periphery of a minicomputer system.

Intravenous Angiocardiography Using Digital Image Processing: 1. Experience with Axial Projections in Normal Pigs

Computerized digitization and processing of roentgen video images was performed in four-chamber-view intravenous angiocardiography in normal pigs. Significant contrast enhancement was obtained through electrocardiogram-gated background subtraction and rescaling after integration of multiple background and contrast images. Histogram equalization and time parameter extraction or functional imaging was also used. The left and right heart were well visualized after intravenous injection of 0.5-1 ml Urografin-76% per kg of body weight. Image processing of left and right ventricular end diastolic and end systolic frames was performed as well as subtraction of right and left end systolic from end diastolic frames and subtraction of right from left heart frames. If left ventricular end systolic images were subtracted from end diastolic images, the left ventricle was seen without continuity with the left atrium. Through time interval difference processing, left ventricular wall motion per time unit could be studied. The results were the basis for subsequent experiments concerning detection of septal defects in pigs as well as clinical studies.

Assessment of cardiovascular function by digital angiocardiography

A methodology for computerized digital videoangiocardiography is briefly described. Single or biplane projection image series from the cardiovascular system are combined with the corresponding physiologic (electrocardiogram, blood pressure, etc.) reference data, digitized and stored as a block of simultaneously available information representing anatomic and functional aspects of the cardiovascular system. Simple mask mode and more complex modes of digital subtraction, image combination and manipulation techniques, as developed during the last decade, are mentioned. These techniques are primarily useful to separate the contrast bolus from the background, thereby allowing contrast enhancement with less contrast medium injected selectively, or so-called noninvasive intravenous angiocardiography. Ventricular function can be assessed by these simple digital image processing techniques. This has been proved for determining right ventricular volumes and ejection fraction with respect to reproducibility and accuracy using conventional biplane angiocardiography as reference. More complex techniques for the assessment of function, in particular blood flow distributions in the systemic circulation, are described using information from the whole digitized angiocardiographic image series by extracting time and volume parameters from the complete matrix of pixel densograms. Various modes of extraction and display of time parameters allow a generation of parametric images that display heretofore unavailable flow patterns reflecting the progress of the contrast bolus within the arterial tree. Based on an adequate temporal segmentation (for example, time segments of one cardiac cycle) and simultaneous volume determination of the circulatory structure from the area of the densogram, relative and absolute flow as well as regional flow distribution in a branching arterial system can be determined.(ABSTRACT TRUNCATED AT 250 WORDS)

Intravenous Angiocardiography Using Digital Image Processing II. Detection of Left-to-Right Shunts in an Animal Model

Digital image processing was applied in an experimental model to study the feasibility of intravenous angiocardiography for the detection of cardiac lesions with left-to-right shunting. Methods were designed to produce ventricular and atrial septal defects (VSD, ASD) as well as patent ductus arteriosus (PDA) equivalent in pigs. After intravenous (IV) injection of 0.5-1 ml of Urografin 76% per kg body weight, digitization and computerized processing of roentgen video images recorded at a rate of 50 fields per second was performed. The radiographic images were recorded in the four-chamber view obtained by 30-35 degrees of caudocranial angulation and 50-60 degrees of left anterior oblique positioning of the animal. The processing of images included electrocardiogramgated background subtraction, rescaling, and sometimes histogram equalization. Integration of multiple background and multiple contrast images was performed in order to increase the signal-to-noise ratio. Ventricular septal defects and patent ductus arteriosus could be detected per se, while atrial septal defect only could be indirectly detected. Respiratory motion artifacts could be avoided by cross-correlation respiratory gating or by subtracting left ventricular end systole from end diastole during the same cardiac cycle of left ventricular opacification. The results of the experiments in pigs are the basis for continued clinical use of digitized IV angiocardiography.

Digital Videodensitometry: Some Approaches to Radiographic Image Restoration and Analysis

Videodensitometrie methods for the evaluation of blood flow and ventricular function have been developed in the past 15 years in parallel with cinedensitometry (Heintzen 1971; Heuck 1973; Heintzen and Bursch 1978). Recently, the successful application of digital image processing techniques in angiography (Brennecke et al. 1977; Kruger et al. 1978; Ovitt et al 1978) has increased the interest in video methods since they are directly compatible with digital electronic image processing systems. Digitization of the video signal brings the flexibility and reproducibility of computerized data processing to the field of densitometric analysis. Some of the degradations introduced by the imaging system can be eliminated by digital restoration techniques in order to improve the accuracy of image analysis.

Intravenous Angiocardiography Using Digital Image Processing: Experience With Axial Projections In Normal Pigs And In Pigs With Experimentally Generated Left-To-Right Shunts

Computerized digitization and processing of roentgen video images recorded at a rate of 50 per second was tested in intravenous angiocardiography in normal pigs weighing 15 to 20 kg. Roentgen video images were recorded in the 4-chamber view obtained by 30-35 degrees caudocranial angulation of the x-ray tube and 50-60 degrees LAO obliquity in the pig. Significant contrast enhancement was obtained through ECG-gated background subtraction and rescaling after integration of multiple background as well as contrast images. Occasionally, histogram equalization was used to further enhance contrast. To study temporal changes in cardiac motion, time parameter extraction or functional imaging was applied as well. The left and right heart were well visualized after intravenous injection of 1/3-1 cc. 76% Urografin per kg. bodyweight. Special purpose processing like subtraction of the end systolic phase from the end diastolic in the left and right ventricles as well as subtraction of the right ventricular phase from the left ventricular phase was also performed. If the left ventricular end systolic phase was subtracted from the end diastolic, most of the left atrium was also subtracted whereby the left ventricle was seen without continuity or superimposition of the left atrium. Experimentally generated ventricular and atrial septal defects as well as patent ductus arteriosus could be detected using the described technique. The results of the animal experiments became the basis for subsequent applications in children with congenital heart disease.

Assessment of Myocardial Perfusion Using Digital Angiocardiography

Digital processing of coronary angiocardiograms provides the capability for visualizing myocardial opacification by applying contrast enhancement techniques. Smith et al. (1978) described roentgen scanning densitometry as a new method for the analysis of the spatial distribution as well as the movement of radiopaque indicators through the coronary microcirculation. The basic procedure in these studies was the subtraction of the natural image background (without contrast material) from images of the myocardial opacification period on a picture element basis. Previous work from this laboratory has similarly demonstrated the potential enhancement of angiographic contrast by digital subtraction, integration, and rescaling techniques (Brennecke et al. 1976, 1977; Heintzen et al. 1978). It was noticed that not only coronary but likewise aortic root or left ventricular injections of contrast medium could successfully be applied for the visualization of the perfused myocardium (Radtke 1982; Radtke et al. 1981a). It was the immediate goal of this experimental study to systematically examine the needs and performance of digitized myocardiography with regard to both normal and pathologic perfusion conditions of the coronary circulation.

Three-dimensional analysis of the cardiovascular system

Angiocardiographic methods for quantitative analysis of the dynamic geometry of the heart are described. 3D spatial ‘reconstruction’ of the left or right ventricle from their traced and digitized contours after contrast injection allow the generation of simplified, but clinically useful 3D models of the ventricles which form the basis for left and right ventricular volume determination and contraction pattern analysis. These models can be improved by combining the biplane density profiles of the projection images with some a priori knowledge.

Digital Processing of 2-Dimensional Contrast Echocardiograms

The 2-dimensional ultrasonic imaging of the heart is an important diagnostic tool in cardiology. It provides information about anatomical structures and dynamic function of the heart. The peripheral injection of echo contrast material, such as glucose solution or saline,improves recognition of intracardiac defects. Using this contrast technique for determination of ventricular volume one disadvantage becomes obvious: contrast material and cardiac tissue show almost the same pattern in the ultrasonic image. Thus the borders of the cardiac chamber,which often are only poorly defined in the non-contrast echocardiogram,become unrecognizable in the contrast echocardiogram. The purpose of this study was to improve the analysis of contrast echocardiograms with special respect to border recognition by computerized processing techniques. In order to improve the border recognition in contrast echocardiograms we used modified digital image enhancement techniques primarily developed for digital video angiography.