Cubic-Spline Interpolation for Sparse-View CT Image Reconstruction With Filtered Backprojection in Dynamic Myocardial Perfusion Imaging.

We investigated a projection interpolation method for reconstructing dynamic contrast-enhanced (DCE) heart images from undersampled x-ray projections with filtered backprojecton (FBP). This method may facilitate the application of sparse-view dynamic acquisition for ultralow-dose quantitative computed tomography (CT) myocardial perfusion (MP) imaging. We conducted CT perfusion studies on 5 pigs with a standard full-view acquisition protocol (984 projections). We reconstructed DCE heart images with FBP from all and a quarter of the measured projections evenly distributed over 360°. We interpolated the sparse-view (quarter) projections to a full-view setting using a cubic-spline interpolation method before applying FBP to reconstruct the DCE heart images (synthesized full-view). To generate MP maps, we used 3 sets of DCE heart images, and compared mean MP values and biases among the 3 protocols. Compared with synthesized full-view DCE images, sparse-view DCE images were more affected by streak artifacts arising from projection undersampling. Relative to the full-view protocol, mean bias in MP measurement associated with the sparse-view protocol was 10.0 mL/min/100 g (95%CI: -8.9 to 28.9), which was >3 times higher than that associated with the synthesized full-view protocol (3.3 mL/min/100 g, 95% CI: -6.7 to 13.2). The cubic-spline-view interpolation method improved MP measurement from DCE heart images reconstructed from only a quarter of the full projection set. This method can be used with the industry-standard FBP algorithm to reconstruct DCE images of the heart, and it can reduce the radiation dose of a whole-heart quantitative CT MP study to <2 mSv (at 8-cm coverage).

Ultra-low dose quantitative CT myocardial perfusion imaging with sparse-view dynamic acquisition and image reconstruction: A feasibility study.

PURPOSE: We implemented and validated a compressed sensing (CS) based algorithm for reconstructing dynamic contrast-enhanced (DCE) CT images of the heart from sparsely sampled X-ray projections. METHODS: DCE CT imaging of the heart was performed on five normal and ischemic pigs after contrast injection. DCE images were reconstructed with filtered backprojection (FBP) and CS from all projections (984-view) and 1/3 of all projections (328-view), and with CS from 1/4 of all projections (246-view). Myocardial perfusion (MP) measurements with each protocol were compared to those with the reference 984-view FBP protocol. RESULTS: Both the 984-view CS and 328-view CS protocols were in good agreements with the reference protocol. The Pearson correlation coefficients of 984-view CS and 328-view CS determined from linear regression analyses were 0.98 and 0.99 respectively. The corresponding mean biases of MP measurement determined from Bland-Altman analyses were 2.7 and 1.2ml/min/100g. When only 328 projections were used for image reconstruction, CS was more accurate than FBP for MP measurement with respect to 984-view FBP. However, CS failed to generate MP maps comparable to those with 984-view FBP when only 246 projections were used for image reconstruction. CONCLUSION: DCE heart images reconstructed from one-third of a full projection set with CS were minimally affected by aliasing artifacts, leading to accurate MP measurements with the effective dose reduced to just 33% of conventional full-view FBP method. The proposed CS sparse-view image reconstruction method could facilitate the implementation of sparse-view dynamic acquisition for ultra-low dose CT MP imaging.

Ultra-Low-Dose Sparse-View Quantitative CT Liver Perfusion Imaging.

Radiation dose of computed tomography liver perfusion imaging can be reduced by collecting fewer x-ray projections in each gantry rotation, but the resulting aliasing artifacts could affect the hepatic perfusion measurement. We investigated the effect of projection undersampling on the assessment of hepatic arterial blood flow (HABF) in hepatocellular carcinoma (HCC) when dynamic contrast-enhanced (DCE) liver images were reconstructed with filtered backprojection (FBP) and compressed sensing (CS). DCE liver images of a patient with HCC acquired with a 64-row CT scanner were reconstructed from all the measured projections (984-view) with the standard FBP and from one-third (328-view) and one-fourth (246-view) of all available projections with FBP and CS. Each of the 5 sets of DCE liver images was analyzed with a model-based deconvolution algorithm from which HABF maps were generated and compared. Mean HABF in the tumor and normal tissue measured by the 328-view CS and FBP protocols was within 5% differences from that assessed by the reference full-view FBP protocol. In addition, the tumor size measured by using the 328-view CS and FBP average images was identical to that determined by using the full-view FBP average image. By contrast, both the 246-view CS and FBP protocols exhibited larger differences (>20%) in anatomical and functional assessments compared with the full-view FBP protocol. The preliminary results suggested that computed tomography perfusion imaging in HCC could be performed with 3 times less projection measurement than the current full-view protocol (67% reduction in radiation dose) when either FBP or CS was used for image reconstruction.

Refining a region-of-interest within an available CT image.

This paper describes a numerical method for refining the image of a region-of-interest (RoI) within an existing tomographic slice, provided that projection data are stored along with the image. Using the attributes of the image, projection values (ray-sums) are adjusted to compensate for the material outside the RoI. Advantage is taken of the high degree of overdetermination of common computed tomography systems to reconstruct an RoI image over smaller pixels. The smaller size of a region-of-interest enables the use of iterative methods for RoI image reconstruction, which are less prone to error propagation. Simulation results are shown for an anthropomorphic head phantom, demonstrating that the introduced approach enhances both the spatial resolution and material contrast of RoI images; without the need to acquire any additional measurements or to alter existing imaging setups and systems.

Correlation between helical surface waves and guided modes of an infinite immersed elastic cylinder.

Scattering of obliquely incident plane acoustic waves from immersed infinite solid elastic cylinders is a complex phenomenon that involves generation of various types of surface waves on the body of the cylinder. Mitri [F.G. Mitri, Acoustic backscattering enhancement resulting from the interaction of an obliquely incident plane wave with an infinite cylinder, Ultrasonics 50 (2010) 675-682] recently showed that for a solid aluminum cylinder, there exist acoustic backscattering enhancements at a normalized frequency of ka⩽0.1. The incidence angle α(c) at which these enhancements are observed lies between the first (longitudinal) and second (shear) coupling angles of the cylinder. He also confirmed the observations previously reported by the authors that there exist backscattering enhancements of the dipole mode at large angles of incidence where no wave penetration into the cylinder is expected. In this paper, physical explanations are provided for the aforementioned observations by establishing a correlation between helical surface waves generated by oblique insonification of an immersed infinite solid elastic cylinder and the longitudinal and flexural guided modes that can propagate along the cylinder. In particular, it is shown that the backscattering enhancement observed at ka⩽0.1 is due to the excitation of the first longitudinal guided mode travelling at the bar velocity along the cylinder. It is also demonstrated that the dipole resonance mode observed at incidence angles larger than the Rayleigh coupling angle is associated with the first flexural guided mode of the cylinder. The correlation established between the scattering and propagation problems can be used in both numerical and experimental studies of interaction of mechanical waves with cylinders.

An alternative method for plotting dispersion curves.

Solving the frequency equation and plotting the dispersion curves in problems of wave propagation in cylinders and plates, particularly when the material is anisotropic, are complicated tasks. The traditional numerical methods are usually based on determination of the zeros of the frequency equation by using an iterative find-root algorithm. In this paper, an alternative method is proposed which extracts the solution of the frequency equation in the form of dispersion curves from the three-dimensional illustration of the frequency equation. For this purpose, a three-dimensional representation of the real roots of the frequency equation is first plotted. The dispersion curves, which are the numerical solutions of the frequency equation, are then obtained by a suitable cut in the velocity-frequency plane. The advantages of this method include simplicity, high speed, low possibility of numerical error, and presentation of the results in a graphical form that promotes ease of interpretation. This method is not directly applicable to problems which incorporate high damping or leaky waves. However, if the damping is not very high, it could be a good estimate of the true dispersion curves.