CT Imaging and Image Reconstruction

• An image is reconstructed from individual picture elements, known as pixels
• Image pixels correspond to a volume element or voxel
• Slice thickness corresponds to the depth of a voxel

Computed Tomography (CT) Imaging

• Early CT imaging uses a narrow X-ray beam scanned across a patient, synchronized with a radiation detector on the opposite side
• For a monoenergetic or nearly monoenergetic beam, the transmission of X-rays through the patient is defined as I = Ioe−μx
• I represents the intensity of the attenuated X-ray beam
• I0 is the unattenuated X-ray beam
• x is the thickness of the material
• When X-rays pass through multiple regions (n) with varying linear attenuation coefficients, the transmission is I = I0e−Σ(i=1 to n)μixi
• Σ(i=1 to n)μixi = (μ1x1 + μ2x2 + ... + μnxn)

Image Reconstruction and Processing

• CT image reconstruction converts X-ray penetration measurements into a digital image of a body section
• The process is a mathematical one, using image reconstruction methods

Reconstruction Techniques

• Back projection
• Algebraic reconstruction
• Iterative reconstruction
• Filtered back projection

Backprojection

• Summed attenuation values are averaged out over the row
• With several projections, the result gets closer to the actual image
• Too few projections can cause artifacts
• Blurred images can be solved by filtered backprojection
• Multislice scanners use filter interpolation, where all projections within an axial slice get summed and averaged

Iterative Reconstruction

• Filtered back projection assigns a number value to all pixels in the matrix
• A computer calculates expected detector values from the generated image
• The computer compares expected values to actual detector measurements and adjusts image values accordingly to bring all closer to the true values
• It is used almost exclusively now
• It can reduce CT dose
• Weakness: Complex and lengthy calculations

Image Composition

• A CT image comprises pixels along a greyscale
• The level of grey indicates the material density, expressed as the linear attenuation coefficient
• It is represented numerically by Hounsfield Units, also known as the CT number
• Hounsfield Units are set relative to water, which measures 0

Hounsfield Unit Formula

• HU = 1000 x (μt - μw) / μw
• μt = attenuation coefficient of tissue
• μw = attenuation coefficient of water

Hounsfield Unit Values in CT

• Bone: +1000 HU
• Liver: 40 to 60 HU
• White Matter: 20 to 30 HU
• Grey Matter: 37 to 45 HU
• Intravascular Blood: 30 to 45 HU
• Fresh Clotted Blood: 70 to 80 HU
• Muscle: 10 to 40 HU
• Kidney: 30 HU
• CSF: 15 HU
• Water: 0 HU
• Fat: -50 to -100 HU
• Lung: -600 to -950 HU
• Air: -1000 HU

CT Number Window

• Hounsfield units are visualized in an 8-bit greyscale, offering 128 grey values
• Parameters that define display include:
• Window Level determines the CT number of mid-grey
• Window Width is the range of Hounsfield Units from black to white
• WW and WL settings are dictated by clinical need
• Selecting the appropriate window width and level leads to optimum visualization of tissues
• Different WW and WL settings are used to visualize soft tissue, lung tissue, or bone