X-Ray Projection Imaging: Detector Technology

Question 1

What are the dominant photon interactions within the diagnostic energy range? Which of these is used for x-ray photon detection and why?

Answer 1

• Rayleigh scattering. Elastic scattering of photon from a bound electron => not useful for x-ray photon detection.
• Compton scattering. Inelastic scattering of photon from an outer shell electron (ejecting the electron in the process). However, the ejected electron typically has low energy and a short path length making it difficult to detect before absorption. Resultant radiative transitions are also difficult to detect and localise. => not useful for x-ray photon detection.
• Photoelectric effect. Inner shell electron is ejected after absorbing energy of incident photon. Photoelectron energy depends on incident photon energy and inner shell electron binding energy. Radiative transitions involve K-shell and are typically in the visible spectrum. => useful for x-ray photon detection.

Question 2

How do the dominant photon interactions within the diagnostic energy range approximately vary with E and Z.

Answer 2

• Photoelectric effect proportional to Z^3/E^3.
• Compton scattering proportional to 1/E.
• Rayleigh scattering proportional to Z^2/E^2.

Question 3

What does the photon interaction cross-section describe and how can it be measured?

Answer 3

The cross-section is the probability of an interaction. For the photoelectric effect, this can be measured by looking at the mass-attenuation coefficient.

Question 4

What properties of a material are required for efficient detection of photons in the diagnostic energy range?

Answer 4

• Want a material with a high probability of photoelectric interaction within the diagnostic energy range. As the probability of photon interaction is proportional to Z^3, we want a material with a high Z.
• We also want a material in which transition energies match that of the incoming photon. This results in a K-edge, ‘boosting’ the photoelectric effect cross-section in the diagnostic energy range.

Question 5

How does direct x-ray photon detection work in general?

Answer 5

Direct detection utilises the photoelectrons produced in the photoelectric interaction between incident x-ray photons and the material (x-ray -> charge -> signal).

Question 6

How do direct semiconductor detectors work?

Answer 6

Photoelectrons (resulting from photoelectric interaction between incident x-ray photons and the semiconductor material) generate subsequent ionisation events within the material. This creates electrons/holes and a measurable charge for detection. A large electric field is applied across the material to prevent lateral spread of charge carriers so the point of interaction can be localised for good spatial resolution. a-Se is the most common.

Question 7

What are the limitations of direct a-Se detectors?

Answer 7

• K-edge of a-Se is at low energy (useful for mammography) => thicker layer required for efficient detection. This means a larger electric field is required to maintain spatial resolution => increased dark current and reduced SNR at low doses. Electron mobility/trapped electron issues increase. Cost also increases.
• Cooling is required to prevent a-Se crystallisation.
• Resultant pixel value depends on energy deposited at that pixel - ideally want equal efficiency across the whole energy range.

Question 8

How does indirect x-ray photon detection work in general?

Answer 8

Indirect detection utilises the light photons produced in radiative transitions subsequent to the photoelectric interaction between incident x-ray photons and the material (x-ray -> light -> charge -> signal). This could be in the form of fluorescence (photon produced in rapid decay of electron back to ground state) or phosphorescence (photon produced in delayed decay of trapped electron to ground state).

Question 9

How do indirect flat-panel detectors work?

Answer 9

Photons produced via fluorescence in a ‘scintillation layer’ are converted to charge via a light detection system (e.g. photodiode a-Si array or CCD - former is most common). Doped CsI is the most common scintillation material. Gadolinium-based materials also used but not as common.

Question 10

How do indirect CR detectors work?

Answer 10

Incident x-ray photons excite electrons within a photostimulable phosphor material (BaFBr:Eu is the most common) into trapped states. Red light exposure releases these trapped electrons producing blue light photons in a process known was photostimulated luminescence. Photons are detected by an array of PMTs. CR plate erased by exposure to intense white light.

Question 11

Why are indirect CsI-based detectors more efficient that gadolinium-based indirect detectors?

Answer 11

• CsI can be deposited in a needle-like crystal structure. The improves spatial resolution where the needle structures act as a ‘light guide’ for photons => CsI layer thickness can be increased, thus increasing efficiency.
• Gadolinium-based detectors are formed in a powder structure. This means a trade-off is required as increasing thickness (to improve efficiency) will reduce spatial resolution due the increased scatter probability for photons.

Question 12

What are the limitations of indirect CsI-based detectors?

Answer 12

• Fill factor associated with the readout electronics in a photodiode pixel array means pixel size can only be made so small, limiting the spatial resolution achievable. This means they are not as useful as direct detectors for mammography.
• Relatively fragile.
• Expensive.

Question 13

What are the limitations of indirect CR detectors?

Answer 13

• Laser spot size limits resolution.
• Light spread in phosphor material limits resolution.
• Dirt etc. can cause dropouts.
• Mechanical problems with reader.
• Cracks in plate.
• Images fade over time.
• Ghost images due to erasure problems.
• Careful storage required to ensure no unwanted exposure.

Question 14

Describe the signal transfer property (STP) for a direct digital detector absolute raw image.

Answer 14

Pixel value increases linearly with dose over a large dynamic range. It will be limited at the lower end by noise. At the higher end, it will be limited by charge saturation.

Question 15

Give some examples of how raw x-ray images are processed?

Answer 15

• Flat-fielding: A correction map is created by exposing the detector to a uniform x-ray dose. This is used to
  corrects for spatial variation in detector response.
• Defective pixels: A correction map is created to correct for defective pixels (typically by averaging the surrounding pixel values).
• Pre-processing to map the dynamic range of diagnostic information to an image (typically performed via histogram analysis).
• Post-processing to reduce the dynamic range to a level that can be displayed on a monitor (i.e. window/level).

Question 16

What is pixel value linearisation and why is it required?

Answer 16

Typically, a non-linear STP will be apparent after pre-processing. The STP is measured in routine QA. Pixel values will then be linearised to allow for meaningful quantitative analysis to take place (i.e. MTF/NPS).

Question 17

What is the detector dose index (DDI) or exposure index (EI)?

Answer 17

In digital systems, under- and overexposure can not be easily distinguished due to the re-scaling of diagnostic information from the large dynamic range into a viewable image. This has the potential to lead to dose creep. DDI/EI gives an indication of the detector exposure level to monitor doses. AAPM recommends that DDI/EI vary linearly with dose.

Question 18

What is DQE and how is it defined?

Answer 18

Detective quantum efficiency (DQE) = SNR_out^2/SNR_in^2. SNR_in depends on the Poisson statistics of the incident x-ray beam. SNR_out depends on this quantum noise and the additional noise components associated with the detector (i.e. electronic). It gives a measure of how efficient a detector is and how much noise it adds to an image in the process.

Question 19

Which detector type has the highest DQE? Why must care be taken when comparing DQEs?

Answer 19

Needle structured CsI indirect flat panel detectors typically have the highest DQE.

Care must be taken when comparing DQEs as they are dependent on beam energy (i.e. kV and filtration) and they vary with spatial frequency.

Question 20

Describe how an image intensifier works.

Answer 20

• X-ray photons are converted to light photons at an input phosphor (CsI).
• Light photons are converted to electrons at a photocathode via the photoelectric effect.
• Electrodes are accelerated and focussed by electrodes held at a high potential difference (thus increasing the electron flux).
• Electrons are converted back to light photons at an output phosphor (ZnCdS).
• Light photons are incident on a camera via an optical coupling.

Question 21

What are important properties of the photocathode in the image intensifier?

Answer 21

• Held at negative bias to repel electrons and initiate an electron beam for acceleration and focussing towards the output phosphor.
• Small work function to minimise any photon losses.
• Optimised for the light spectrum emitted by the input phosphor (CsI).

Question 22

Explain flux gain, magnification gain, brightness gain and conversion factor for an image intensifier.

Answer 22

• Flux gain: Number of light photons at output compared to those produced at the input. As electrons are accelerated, they will gain energy and produce more photons at the output phosphor.
• Minification gain: Input area/output area. There are more photons per unit area at the output phosphor.
• Brightness gain = flux gain x minification gain.
• Conversion factor = luminance out/doserate in. This accounts for photon losses.

Question 23

Explain field size/magnification alteration for image intensifiers.

Answer 23

• The voltage applied across the focussing electrodes can be used to map a smaller area of the input phosphor (field size) to the output phosphor.
• This will decease the field size, magnifying the image and enhancing spatial resolution.
• This will reduce brightness gain. Minification gain will decrease as the input area will be smaller. This means a higher dose is required to maintain SNR.

Question 24

Explain collimation alteration for image intensifiers.

Answer 24

• Reducing collimation reduces the input x-ray field size.
• The anatomy irradiated and, therefore, displayed on the screen will be less (despite the field/image size being the same). This will reduce patient dose.
• Minification gain and spatial resolution will not be affected.

Question 25

What are the limitations of image intensifiers?

Answer 25

• Resolution is limited by the output phosphor and camera.
• Veiling glare: Back shine of light photons from the output phosphor and x-rays penetrating through to the output phosphor can cause a loss of contrast at low spatial frequencies.
• Vignetting: Darkening of the edges of images resulting from poor optics/camera.
• Space-charge effects resulting from repulsion of electrons at the output). There is a trade-off between resolution and minification gain regarding the size of the output phosphor.
• Dynamic range is less than a flat panel detector.
• Distortion effects caused by the electron optics.