H. Neale Cardinal

Intra- and inter-observer variability and reliability of prostate volume measurement via two-dimensional and three-dimensional ultrasound imaging

We describe the results of a study to evaluate the intra- and inter-observer variability and reliability of prostate volume measurements made from transrectal ultrasound (TRUS) images, using either the (optimal) height-width-length (HWL) method (V = pi/6 HWL) with two-dimensional (2D) TRUS images (obtained as cross-sections of three-dimensional [3D] TRUS images) or manual planimetry of 3D TRUS images (the 3D US method). In this study, eight observers measured 15 prostate images, twice via each method, and an analysis of variance (ANOVA) was performed. This analysis shows that, with the 3D US method, intra-observer prostate volume estimates have 5.1% variability and 99% reliability, and inter-observer estimates have 11.4% variability and 96% reliability. With the HWL method, intra-observer estimates have 15.5% variability and 93% reliability, and inter-observer estimates have 21.9% variability and 87% reliability. Thus, in vivo prostate volume estimates from manual planimetry of 3D TRUS images have much lower variability and higher reliability than HWL estimates from 2D TRUS images.

Semiautomatic three-dimensional segmentation of the prostate using two-dimensional ultrasound images.

In this paper, we report on two methods for semiautomatic three-dimensional (3-D) prostate boundary segmentation using 2-D ultrasound images. For each method, a 3-D ultrasound prostate image was sliced into the series of contiguous 2-D images, either in a parallel manner, with a uniform slice spacing of 1 mm, or in a rotational manner, about an axis approximately through the center of the prostate, with a uniform angular spacing of 5 degrees. The segmentation process was initiated by manually placing four points on the boundary of a selected slice, from which an initial prostate boundary was determined. This initial boundary was refined using the Discrete Dynamic Contour until it fit the actual prostate boundary. The remaining slices were then segmented by iteratively propagating this result to an adjacent slice and repeating the refinement, pausing the process when necessary to manually edit the boundary. The two methods were tested with six 3-D prostate images. The results showed that the parallel and rotational methods had mean editing rates of 20% and 14%, and mean (mean absolute) volume errors of -5.4% (6.5%) and -1.7% (3.1%), respectively. Based on these results, as well as the relative difficulty in editing, we conclude that the rotational segmentation method is superior.

Automatic needle segmentation in three-dimensional ultrasound images using two orthogonal two-dimensional image projections

In this paper, we describe an algorithm to segment a needle from a three-dimensional (3D) ultrasound image by using two orthogonal two-dimensional (2D) image projections. Not only is the needle more conspicuous in a projected (volume-rendered) image, but its direction in 3D lies in the plane defined by the projection direction and the needle direction in the projected 2D image. Hence, using two such projections, the 3D vector describing the needle direction lies along the intersection of the two corresponding planes. Thus, the task of 3D needle segmentation is reduced to two 2D needle segmentations. For improved accuracy and robustness, we use orthogonal projection directions (both orthogonal to a given a priori estimate of the needle direction), and use volume cropping and Gaussian transfer functions to remove complex background from the 2D projection images. To evaluate our algorithm, we tested it with 3D ultrasound images of agar and turkey breast phantoms. Using a 500 MHz personal computer equipped with a commercial volume-rendering card, we found that our 3D needle segmentation algorithm performed in near real time (about 10 fps) with a root-mean-square accuracy in needle length and endpoint coordinates of better than 0.8 mm, and about 0.5 mm on average, for needles lengths in the 3D image from 4.0 mm to 36.7 mm.

Experimental and theoretical x‐ray imaging performance comparison of iodine and lanthanide contrast agents

Contrast agents based on the lanthanide elements gadolinium and holmium have recently been developed for magnetic resonance imaging (MRI). Because of the increased atomic number of these elements relative to iodine, these new compounds, used as x-ray contrast agents, may yield higher radiographic contrast, and hence improved x-ray image quality, relative to conventional iodinated compounds, for clinically useful x-ray spectra. This possibility has been investigated, in independent experimental and theoretical studies, for two x-ray imaging systems: a digital radiographic system, using an x-ray image intensifier (XRII) and charge-coupled device (CCD) detector; and a conventional screen/film system, using a Lanex Regular screen. Iodine, gadolinium, and holmium contrast agents were investigated over a wide range of concentration-thickness products (0.1-0.6 M cm) and diagnostic x-ray spectra (60-120 kVp). A simple theoretical model of x-ray detector response predicts the experimental radiographic contrast measurements with a mean absolute error of 8.0% for the XRII/CCD system and 5.9% for the screen/film system, and shows that the radiographic contrast for these two systems is representative of all XRII and screen/film systems. An index of image quality is defined, and its dependence on radiographic contrast, x-ray fluence per unit dose, and detective quantum efficiency (DQE) is shown. Theoretical values of the index, predicted by our model, are then used to compare the performance of the three contrast agents for the two systems investigated. In general, iodine performance decreases steadily with increasing kVp, gadolinium performance has a broad maximum near 85 kVp, and gadolinium outperforms holmium. Gadolinium outperforms iodine for spectra above (and vice versa below) about 72 kVp, depending slightly on spectrum filtration, object thickness, and detector type. Thus, raising the kVp to shorten exposure times or reduce x-ray tube heat loading results in a loss of image quality with iodine, but not with gadolinium. Similarly, beam-hardening artifacts in performing video densitometry with iodine would be reduced with gadolinium. Gadolinium-based contrast agents are thus shown to offer several practical advantages over conventional iodinated contrast agents.

A real vessel phantom for flow imaging: 3-d doppler ultrasound of steady flow

Vascular phantoms are used to assess the capabilities of various imaging techniques, such as x-ray CT and angiography, and B-mode, power Doppler, and colour Doppler ultrasound (US). They should, therefore, accurately mimic the vasculature, blood, and surrounding tissue, in regard to both imaging properties and vessel geometry. In the past, a variety of walled and wall-less vessel models have been used. However, these models only approximate the true vessel geometry, and generally lack pathologic features such as plaques or calcifications. To amend these deficiencies, we have developed a real vessel phantom for US and x-ray studies, which comprises a fixed human vessel specimen, cannulated onto two acrylic tubes, and embedded in agar in an acrylic box. Earlier, we demonstrated a good overall correlation between x-ray angiography, CT, and 3-D B-mode US images of this phantom. Here, we extend its use to flow imaging with 3-D power and 3-D colour Doppler US.

Analysis of Linear, Area and Volume Distortion in 3D Ultrasound Imaging

We have developed a three-dimensional (3D) ultrasound imaging system that uses a side-firing probe, axially rotated under computer control, to acquire a series of 2D images, from which the 3D image is reconstructed. For an undistorted reconstruction, the inner radius R0 of the 2D images and the total scanning angle theta must be known accurately. Here, we describe (a) a theoretical analysis of the relative distortion in image shape, length, area, and volume due to an error delta R in R0 or delta theta in theta; (b) measurements of these in simulated and real 3D images; and (c) a method to calibrate R0, theta, and image scale accurately. Theoretically, all four relative distortions vary as P delta R/R + Q delta theta/theta, where magnitude of P < or = 1, magnitude of Q < or = 1, and R is the average distance of the object from the axis. In every case, the simple theoretical formulas for P and Q agree with image measurements to within the measurement uncertainty.

Theoretical optimization of a split septaless xenon ionization detector for dual‐energy chest radiography

It is proposed that digital scanned projection radiography of the chest be performed by using an energy-sensitive septaless xenon ionization detector (SXID) to obtain dual-energy images. The proposed detector is composed of a front region, sensitive to low-energy x rays, and a rear region, sensitive to high-energy x rays, separated by a suitable filter layer. We have developed a simple, precise theoretical formulation for dual-energy optimization, and applied it to the split SXID. We describe the variation of optimum detector performance with source kilovoltage and filtration (material and thickness), and hence heat loading, under conditions of constant exposure and constant dose. We estimate dose as the average absorbed dose to an equivalent water layer of suitable thickness, assuming slab geometry, so that the calculation is as simple as that for exposure.

Automatic needle segmentation in 3D ultrasound images

In this paper, we propose to use 2D image projections to automatically segment a needle in a 3D ultrasound image. This approach is motivated by the twin observations that the needle is more conspicuous in a projected image, and its projected area is a minimum when the rays are cast parallel to the needle direction. To avoid the computational burden of an exhaustive 2D search for the needle direction, a faster 1D search procedure is proposed. First, a plane which contains the needle direction is determined by the initial projection direction and the (estimated) direction of the needle in the corresponding projection image. Subsequently, an adaptive 1D search technique is used to adjust the projection direction iteratively until the projected needle area is minimized. In order to remove noise and complex background structure from the projection images, a priori information about the needle position and orientation is used to crop the 3D volume, and the cropped volume is rendered with Gaussian transfer functions. We have evaluated this approach experimentally using agar and turkey breast phantoms. The results show that it can find the 3D needle orientation within 1 degree, in about 1 to 3 seconds on a 500 MHz computer.

Analytic approximation of the log‐signal and log‐variance functions of x‐ray imaging systems, with application to dual‐energy imaging

In the analysis of x-ray system performance, the log-signal function, or negative logarithm of the relative detector signal, and the analogously defined log-variance function, are of central importance. These are smooth, monotonic functions of object thickness, which are nonlinear for nonmonoenergetic x-ray source spectra. If we assume a dual-energy decomposition of the object into two basis materials, then they can be written as analytic functions f(x,y) and f*(x,y), respectively, of the component thicknesses (x,y) of the object. In this paper, we analytically develop the Taylor series of these functions, prove that they converge everywhere, and parametrize their coefficients via suitable central spectral moments of the basis-material attenuation coefficients. We then show how the lower-order moments can be used to construct, in closed form, smooth, monotonic, second-order (conic) surface functions which closely approximate f(x,y) and f*(x,y) over the entire feasible domain. A simplified construction, based on using appropriate asymptotic values of the basis-material attenuation coefficients to match the asymptotic behavior of these functions, is also given. The inclusion of image components with K-edge absorption spectra, such as iodine, is done without effort. Extension of the results to the construction of similar (virtually exact) third-order (cubic) surface approximations is straightforward. As an illustration of the broad applicability of this approach, we extend our analysis to the construction of similar approximations to the inverse (decomposition) functions for an arbitrary dual-energy system, and investigate their numerical accuracy for a model dual-kVp system. We conclude that this extended analysis provides an accurate description of the system behavior in terms of a small number of physically meaningful parameters. This parametrization permits greater physical insight into the system behavior, while at the same time simplifying its mathematical description, and similarly facilitates the analysis of various measures of imaging performance via either analytic or numerical methods.

An in‐line optical image translator with applications in x‐ray videography

Many applications in radiography require, or would benefit from, the ability to translate, i.e. move, an optical image in the detector plane. In this paper, we describe the design and characterization of a prism-based optical image translator for insertion into existing XRII-video imaging systems. A pair of prisms rotatable about the optical axis form a very compact in-line optical image translator for installation in the parallel light path between an x-ray image intensifier and its video camera. Rotation of the prisms translates the XRII optical image on the camera target. With the addition of x-ray and light collimators to limit the image to a single video line, x-ray streak images may be acquired. By rotating an object in the x-ray beam during a streak, a complete computed tomography (CT) data set may be acquired. This image translator can translate an image anywhere in the focal plane of a 50-mm-output lens within a 40-mm-diam circle. The prisms have an aperture of 50 mm, permitting an optical speed of F/2 with a 50-mm output lens. The design is insensitive to angular alignment errors. This image translator is achromatic, since the spectral width of the output phosphorus of image intensifiers is sufficient to introduce blurring in a nonacrhomatic design. A prism-based image translator introduces image distortion, since the prisms do not operate at minimum deviation. The distortion is less than 4% over all parts of a typical detector area, and less than 1% in the central region of the image.(ABSTRACT TRUNCATED AT 250 WORDS)