Panel Detector-based Cone Beam CT Research Articles

Abstract No. 454 Photon counting C-arm computed tomography for thoracoabdominal interventional imaging: an in vivo porcine feasibility study

Prior studies on conventional CT demonstrate improved contrast resolution and radiation dose efficiency using direct conversion and energy-resolving photon counting detectors (PCD). The objective of this study was to determine the feasibility of using a prototype C-arm based PCD-CT system for thoracoabdominal interventional imaging and assess the potential benefits in comparison to flat panel detector-based cone-beam CT (FPD-CBCT). A 51 cm × 0.6 cm CdTe-based PCD strip with 100 μm pixel size was mounted to the front of the FPD of a Siemens Artis Zee interventional system. The PCD has a max frame rate of 150 fps and two adjustable energy thresholds for high-speed dual energy imaging. Eight-second rotational FPD-CBCT and PCD-CT acquisitions of the upper abdomen and lower thorax were performed in three female domestic swine (49.3 ± 2.5 kg). Acquisitions in one of the animals were performed after a stent graft was placed during a TIPS procedure. To assess radiation dose reduction potential, PCD-CT scans were performed at 25% (7.5 mGy) of the standard CTDIvol level (30 mGy) used by the conventional FPD-CBCT protocol. Images were assessed for artifacts, noise, contrast and spatial resolution. PCD-CT images had lower noise due to rejection of electronic noise by the PCD. This resulted in higher spatial and contrast resolution for PCD-CT, even at a lower radiation dose (25% of the FPD-CBCT). Blooming and streak artifacts from iodine accumulated in the kidney and bone were lower on the PCD images. Increased soft tissue contrast on the PCD images improved visualization of visceral detail, especially of bowel. Blooming and shading artifacts near the TIPS stent graft were suppressed in PCD-CT images. In this preliminary study, a prototype C-arm based PCD-CT system demonstrated potential benefits in image quality and radiation dose efficiency when compared to FDP-CBCT. Further investigation of C-arm PCD-CT is warranted in order to assess the potential for applying this technology in improving thoracoabdominal interventions.

A unified scatter rejection and correction method for cone beam computed tomography.

Scattered radiation is a major cause of image quality degradation in flat panel detector-based cone beam CT (CBCT). While recently introduced 2D antiscatter grids reject the majority of scatter fluence, the small percentage of scatter fluence still transmitted to the detector remains a major challenge for implementation of quantitative imaging techniques such as dual energy imaging in CBCT. Additionally, this residual scatter is also a major source of grid-induced artifacts, which impedes implementation of 2D grids in CBCT. We therefore present a new method to achieve both robust scatter rejection and residual scatter correction using a 2D antiscatter grid; in doing so, we expand the role of 2D grids from mere scatter rejection devices to scatter measurement devices. In our method, the radiopaque septa of the 2D grid emulate a micro array of beam-stops placed on the detector which introduce spatially periodic septal shadows. By selecting sufficiently thin grid septa, the primary intensity can be reduced while preserving the uniformity of scatter intensity. This enables us to correlate the modulated pixel signal intensity in septal shadows with local scatter intensity. Our method then exploits this correlation to measure and remove residual scatter intensity from projections. No assumptions are made about the object being imaged. We refer to this as Grid-based Scatter Sampling (GSS). In this work, we evaluate the principle of signal modulation with grid septa, the accuracy of scatter estimates, and the effect of the GSS method on image quality using simulations and measurements. We also implement the GSS method experimentally using a 2D grid prototype. Our results demonstrate that the GSS method increased CT number accuracy and reduced image artifacts associated with scatter. With 2D grid and residual scatter correction, HU nonuniformity was reduced from 65HU to 30HU in pelvis sized phantoms, and HU variations due to change in phantom size were reduced from 59HU to 20HU, when compared to use of only a 2D grid. With residual scatter correction via GSS method, grid-induced ring artifacts were suppressed, leading to a 41% reduction in noise. The shape of the modulation transfer function (MTF) was preserved before and after suppression of ring artifacts. Our grid-based scatter sampling method enables utilization of a 2D grid as a scatter measurement and correction device. This method significantly improves quantitative accuracy in CBCT, further reducing the image quality gap between CBCT and multi-detector CT. By correcting residual scatter with the proposed method, grid-induced line artifacts in projections and associated ring artifacts in CBCT images were also suppressed with no compromise of spatial resolution.

FDK Half-Scan with a Heuristic Weighting Scheme on a Flat Panel Detector-Based Cone Beam CT (FDKHSCW)

A cone beam circular half-scan scheme is becoming an attractive imaging method in cone beam CT since it improves the temporal resolution. Traditionally, the redundant data in the circular half-scan range is weighted by a central scanning plane-dependent weighting function; FDK algorithm is then applied on the weighted projection data for reconstruction. However, this scheme still suffers the attenuation coefficient drop inherited with FDK when the cone angle becomes large. A new heuristic cone beam geometry-dependent weighting scheme is proposed based on the idea that there exists less redundancy for the projection data away from the central scanning plane. The performance of FDKHSCW scheme is evaluated by comparing it to the FDK full-scan (FDKFS) scheme and the traditional FDK half-scan scheme with Parker's fan beam weighting function (FDKHSFW). Computer simulation is employed and conducted on a 3D Shepp-Logan phantom. The result illustrates a correction of FDKHSCW to the attenuation coefficient drop in the off-scanning plane associated with FDKFS and FDKHSFW while maintaining the same spatial resolution.