Robert A. Close

X-ray videodensitometric methods for blood flow and velocity measurement: a critical review of literature.

Blood flow rate and velocity are important parameters for the study of vascular systems, and for the diagnosis, monitoring and evaluation of treatment of cerebro- and cardiovascular disease. For rapid imaging of cerebral and cardiac blood vessels, digital x-ray subtraction angiography has numerous advantages over other modalities. Roentgen-videodensitometric techniques measure blood flow and velocity from changes of contrast material density in x-ray angiograms. Many roentgen-videodensitometric flow measurement methods can also be applied to CT, MR and rotational angiography images. Hence, roentgen-videodensitometric blood flow and velocity measurement from digital x-ray angiograms represents an important research topic. This work contains a critical review and bibliography surveying current and old developments in the field. We present an extensive survey of English-language publications on the subject and a classification of published algorithms. We also present descriptions and critical reviews of these algorithms. The algorithms are reviewed with requirements imposed by neuro- and cardiovascular clinical environments in mind.

COMPARISON OF NEURAL NETWORK AND K-NN CLASSIFICATION METHODS IN VOWEL AND PATELLAR SUBLUXATION IMAGE RECOGNITIONS

We make a comparision of classification ability between BPN (BackPropagation Neural Network) and k-NN (k-Nearest Neighbor) classification methods. Voice data and patellar subluxation images are used. The result was that the average recognition rate of BPN was 9.2 percent higher than that of the k-NN classification method. Although k-NN classification is simple in theory, classification time was fairly long. Therefore, it seems that real time recognition is difficult. On the other hand, the BPN method has a long learning time but a very short recognition time. Especially if the number of dimensions of the samples is large, it can be said that BPN is better than k-NN in classification ability.

Regularization method for scatter-glare correction in fluoroscopic images

Fluoroscopic images are degraded by scattering of x-rays from within the patient and by veiling glare in the image intensifier. Both of these degradations are well described by a response function applied to either the scatter-free or primary intensity. The response function is variable, with dependence on such factors as patient thickness and imaging geometry. We describe an automated regularization technique for obtaining response function parameters with a minimal loss of signal. This method requires a high-transmission structured reference object to be interposed between the x-ray source and the subject. We estimate the parameters by minimizing residual correlations between the reference object and the computed subject density after a scatter-glare correction. We use simulated images to evaluate our method for both ideal and clinically realistic conditions. We find that the residual root-mean-square (rms) error ideally decreases with an increasing number of independent pixels (N) as (1/N)1/2. In simulated 256x256 angiograms mean normalized rms errors were reduced from 40% to 11% in noise-free images, and from 41% to 17% in noisy images, with a similar improvement in densitometric vessel cross-section measurements. These results demonstrate the validity of the method for simulated images and characterize its expected performance on clinical images.

Simulating coronary arteries in x-ray angiograms.

Clinical validation of quantitative coronary angiography (QCA) algorithms is difficult due to the lack of a simple alternative method for accurately measuring in vivo vessel dimensions. We address this problem by embedding simulated coronary artery segments with known geometry in clinical angiograms. Our vessel model accounts for the profile of the vessel, x-ray attenuation in the original background, and noise in the imaging system. We have compared diameter measurements of our computer simulated arteries with measurements of an x-ray Telescopic-Shaped Phantom (XTSP) with the same diameters. The results show that for both uniform and anthropomorphic backgrounds there is good agreement in the measured diameters of XTSP compared to the simulated arteries (Pearsons correlation coefficient 0.99). In addition, the difference in accuracy and precision of the true diameter measures compared to the XTSP and simulated artery diameters was small (mean absolute error across all diameters was < or = 0.11 mm +/- 0.09 mm).

Comparison of methods for instantaneous angiographic blood flow measurement.

Several different algorithms have been reported for measurement of blood flow rates and velocities from digital x-ray angiograms. We compare four videodensitometric methods: (1) distance-density curve matching (DDCM), (2) distance-density curve matching with curve-fitting (DDCM-F), (3) bolus mass tracking with curve-fitting (BMT-F) and (4) fluid continuity method (FCM). We tested the flow algorithms with simulated angiograms and with images obtained from a programmable flow phantom under clinically realistic flow and contrast injection conditions including imperfect mixing. All methods perform well for simulated angiograms. On phantom angiograms with constant flow, all methods tended to underestimate flow velocities by at least 7% and demonstrate high variability between consecutive measurements. The FCM demonstrated relatively low variability, but a large negative bias. The DDCM method was moderately biased and had the highest variability. The BMT-F method demonstrated the lowest bias (-7.1%) and the lowest variability both within (27%) and between (27%) studies. No method yields reliable measurements near the peak contrast opacification, when little or no gradient of contrast is present. The extrapolating version of the BMT-F method was also the most robust for estimation of interframe displacements longer than the field of view.

Accuracy assessment of layer decomposition using simulated angiographic image sequences

Layer decomposition is a promising method for obtaining accurate densitometric profiles of diseased coronary artery segments. This method decomposes coronary angiographic image sequences into moving densitometric layers undergoing translation, rotation, and scaling. In order to evaluate the accuracy of this technique, the authors have developed a technique for embedding realistic simulated moving stenotic arteries in real clinical coronary angiograms. They evaluate the accuracy of layer decomposition in two ways. First, they compute tracking errors as the distance between the true and estimated motion of a reference point in the arterial lesion. The authors find that noise-weighted phase correlation and layered background subtraction are superior to cross correlation and fixed mask subtraction, respectively. Second, they compute the correlation coefficient between the true vessel profile and the raw and processed images in the region of the stenosis. They find that layer decomposition significantly improves the correlation coefficient.

Improved image guidance of coronary stent deployment

Accurate placement and expansion of coronary stents is hindered by the fact that most stents are only slightly radiopaque, and hence difficult to see in a typical coronary x-rays. We propose a new technique for improved image guidance of multiple coronary stents deployment using layer decomposition of cine x-ray images of stented coronary arteries. Layer decomposition models the cone-beam x-ray projections through the chest as a set of superposed layers moving with translation, rotation, and scaling. Radiopaque markers affixed to the guidewire or delivery balloon provide a trackable feature so that the correct vessel motion can be measured for layer decomposition. In addition to the time- averaged layer image, we also derive a background-subtracted image sequence which removes moving background structures. Layer decomposition of contrast-free vessels can be used to guide placement of multiple stents and to assess uniformity of stent expansion. Layer decomposition of contrast-filled vessels can be used to measure residual stenosis to determine the adequacy of stent expansion. We demonstrate that layer decomposition of a clinical cine x-ray image sequence greatly improves the visibility of a previously deployed stent. We show that layer decomposition of contrast-filled vessels removes background structures and reduces noise.

Fluid equations applied to blood flow measurement using digital videodensitometry.

RATIONALE AND OBJECTIVES.Time–dependent fluid flow is computed from projection radiographs without bolus tracking by applying the fluid equations of continuity and incompressibility. METHODS.The fluid equations are combined and integrated to yield an equation that describes instantaneous mass conservation within a vessel segment. The technique is demonstrated using phantom images and patient data obtained using a digital subtraction angiography (DSA) system. RESULTS.Instantaneous and mean flow rates are successfully computed with this algorithm, but the uncertainties are overestimated. In a 1.0–cm diameter tube, instantaneous and mean velocities corresponding to 7.3 cm per frame are computed within 13% uncertainty using a 4.0–cm segment length. Mean flow rates computed from standard diagnostic angiograms taken from three different projections agree within 16%. CONCLUSIONS.This technique can successfully compute timedependent flow rates from DSA image sequences with large fluid displacements between frames. The accuracy is strongly dependent on the magnitude of the contrast density gradient.

Automatic correction of biplane projection imaging geometry.

A novel method is presented for correcting errors in measurements of biplane projection imaging geometry without prior identification of corresponding points in the two images. For imaged objects that project onto both images, a constraint equation is obtained that relates weighted integrals along corresponding epipolar lines. The integrals are computed to first order in the angular beamwidth, which is assumed to be small. Starting from measured or estimated values, geometrical parameters are computed iteratively in order to maximize the correlation between epipolar line integrals in the two images. Improvement in the computation of corresponding epipolar lines is demonstrated on images of a wire phantom. The root mean square distance of the epipolar lines from the corresponding reference points is improved from 15 pixel widths to less than 4 pixel widths (1.3 mm). Convergence is demonstrated on phantom images for individual parameter variations up to 70% in relative magnification, a relative shift of the imaging planes by 50 pixels, or a relative rotation of at least 35 degrees around either of two axes. Applicability to clinical images is demonstrated by using a biplane angiogram of a pig to align corresponding points determined from images of a Perspex cube acquired with the same geometry.

Motion-compensated signal and background estimation from coronary angiograms

Conventional vessel tracking and segmentation techniques identify the positions and two- dimensional structure of arteries in each frame of the angiographic sequence, but cannot distinguish the artery and background contributions to the intensity. We report a new technique for motion-compensated estimation of artery and background structures in coronary angiograms. The image within a region of interest is modeled as consisting of a sum of two independently moving layers, one of which contains the artery and one consisting of only background structures. The density of each of these layers is solved under two assumptions: (1) within each layer, the density varies from frame to frame only by rigid translation, and (2) the sum of the densities of the two layers equals the actual image density. This technique can be used to enhance image sequences by subtracting the component of the background whose temporal variation is entirely due to rigid translation. The feasibility of this technique is demonstrated on synthetic and clinical image sequences.