SBDX system in cross-section (horizontal dimension is stretched for clarity). An electron beam is magnetically deflected across the target area. A focused multihole collimator defines an x-ray pencil beam for each focal spot position. As the pencil beam sweeps across the patient, detector data sets are acquired and reconstructed in real time. Images are reconstructed at 10, 15, 20, or , depending on the scanning technique.
Calculated x-ray fluence vs distance in the SBDX geometry, in a frame period. Since the focal spot is positioned at discrete points, some structure exists within the x-ray field. The calculation at each distance is the fluence averaged over a area.
SBDX x-ray source (left) with scale cutaway of multihole collimator (center) and details of the target, exit window, cooling, and filtration (right).
Principle of multiple-pass scanning. The electron beam visits each focal spot position several times during a frame period in order to reduce the anode temperature. (a) A hypothetical heating and cooling curve for a single dwell lasting . (b) Delivery of the same beam current in two passes, each lasting . The reduction in peak anode temperature makes higher beam currents possible without altering total dwell time.
(a) The SBDX detector consists of distinct tiles, each containing detector elements. The stepped arrangement accommodates connections between the readout chip and peripheral electronics. (b) To accommodate high count rates, each detector element has 60 binary-counting sub-elements (borders shown with dotted lines). (c) After each capture period, a detector element value is generated by summing the sub-element outputs (individually 0 or 1). A threshold is used for noise rejection and anti-coincidence logic is used to prevent double-counting due to reabsorbed fluorescence.
(a) Geometry of a single reconstruction plane, showing two overlapping backprojected detector areas. (b) Calculation of the width of a backprojected detector image and the shift between the images from consecutive focal spots, for a reconstruction plane at . (c) Example of the overlap between an image pixel and a backprojected detector element area. (: source-detector distance, : detector width, : focal spot pitch, : backprojected del pitch, : restored del width, : pixel width in reconstructed image).
(a) The projections of an object in the isocenter reconstruction plane span . (b) The backprojected object images of an in-plane object are coincident at the reconstruction plane. (c) When an object is out-of-plane the backprojected object images are spread along the reconstruction plane. The effect is modeled as convolution with a rect function with width .
Method of real-time multiplane composite reconstruction (example geometry shown on the left). Tomosynthetic images are simultaneously reconstructed at 16 planes spaced throughout the chest. Each pixel in a reconstructed image is scored to reflect the degree of local sharpness. The score images are searched at fixed pixel position to find the plane (along ) with the highest score. The plane selection map indicates, for each pixel position, which single-plane image to draw a pixel value from when forming the composite.
Measured exposure rate vs distance from source (circles), and power-law fit (line). Gantry isocenter is from the collimator exit.
(a) Multiplane reconstruction and (b) single-plane reconstruction of the anthropomorphic chest phantom. Out-of-plane blurring in the single-plane reconstruction is most evident for the arteries in the upper right of the image, and in the spine region (arrows).
Image frame from a live coronary artery injection (porcine). Multiplane image reconstruction was performed at .
Example linespread function measurement. The LSF tails, on the right, were examined by windowing down to 1% of maximum intensity and plotting out to a distance of . The ripples in the tails are caused by off-focus radiation.
Measured system MTF for an isocenter plane reconstruction and an in-plane object (circles). Modeled system MTF and its three basic components (solid lines).
Isocenter plane MTF measured with the slit coincident with the reconstructed plane (in-plane, top curve) and with the slit displaced along the axis by (middle curve) and (bottom curve).
Modeled out-of-plane blurring component of the overall MTF, for objects displaced 1, 3, 6, and above the isocenter reconstruction plane.
Isocenter plane MTF measured with various restoration kernel sizes. From top curve to bottom curve, the kernel width was 0%, 82%, 100% (the nominal value), and 123% of the backprojected del width.
SBDX prototype (lower solid line) and MTF modeled for a del and spot (upper solid line). Nominal del-to-pixel interpolation is used in both cases. Circles and squares are the MTFs measured in the horizontal and vertical directions, respectively, for a conventional angiographic system ( I.I. mode, nominal spot, CCD, ).
SBDX scanning techniques.
SBDX source exposure rate below isocenter with a scan, and measured half-value layer.
Article metrics loading...
Full text loading...