Transcript Lecture 5 - Digital, Tomography and dual

FRCR: Physics Lectures
Diagnostic Radiology
Lecture 5
Digital detectors, tomography and dual
energy imaging
Dr Tim Wood
Clinical Scientist
Overview
• What is a digital image?
– The advantages of digital
– The disadvantages
• Computed Radiography (CR)
– The theory of CR
– The advantages and disadvantages
• Digital Radiography (DR)
– The theory of DR
– The advantages and disadvantages
• DR vs CR
• Tomography
• Dual energy imaging
The story so far…
• We know how X-rays are made in the X-ray tube and
how they interact with the patient
• We know how we control the quality and intensity of the
X-ray beam, and hence patient dose, with kVp, mAs,
filtration and distance
• We discussed the main descriptors of image quality
– Contrast
– Spatial Resolution
– Noise
• Discussed ways to improve contrast by minimising
scatter and using contrast agents
• Remember, there is always a balance between patient
dose and image quality – fit for the clinical task!
• Film is a dying medium for X-ray imaging…
Digital imaging
What is a digital image?
• A digital image can be thought of as an array of
pixels (or voxels in 3D imaging) that each take a
discrete value
• The value assigned is dependent on the X-ray
intensity striking it
• Depending on its value, each pixel is assigned a
shade of grey
• Pixel size may determine the limiting spatial
resolution of the system
What is a digital image?
Why bother with digital?
• Film has been used since the beginning, so why
are we changing to digital techniques?
– Increased latitude and dynamic range
– Images can be accessed simultaneously at multiple
workstations
– Viewing stations can be set up in any location
– Uses digital archives rather than film libraries
– Images quicker to retrieve and less likely to be lost
– Post processing
– Softcopy reporting – lower cost if do not print
– No need for dangerous processing chemicals
Dynamic Range - Film
• With conventional film,
too low a dose will
results in a ‘thin’ film
• Too high a dose results
in a very dark film
• Fixed and limited
dynamic range – must
match exposure
parameters to the film
being used
– Gives a measure of
control over patient dose!
Dynamic Range - Digital
• With digital, too low a dose will
still produce a recognisable
image (just a bit noisy!)
• Similarly, too high a dose will
produce a recognisable image
(but with very little noise!)
• Consequences:
– Less retakes = GOOD
– Dose creep = BAD – must pay
special attention to digital imaging
to ensure doses are optimised
Detector Dose Indicators (DDI)
• The restricted latitude of film gives a clear
indication of dose
– Too dark a film = overexposure
– Too light a film = underexposure
• Film has ‘in-built’ quality control of exposure
• Digital images will be presented with the
greyscale optimised no matter what dose is
given
– Will always see a recognisable image, but the noise
will vary
• Can result in dose creep
– Increased dose (lower noise) is not punished by the
detection medium, so tendency to go for slowly
increasing image quality – NOT ACCEPTABLE!
DDI
• The DDI has been introduced for digital imaging
as an indication of the level exposure on a broad
region of the detector
• Analogous to the OD of film
• The definition of DDI is manufacturer specific!
–
–
–
–
Some manufacturers have high DDI = underexposure
Some the other way round
Some are a function of the log of dose
Some are linear…
• Manufacturers will provide an indication of
acceptable range of DDI, but local departments
must validate these – DRLs and OPTIMISATION
• Operators should monitor DDI of patient
exposures to ensure doses remain acceptable
Computed Radiography
Computed Radiography (CR)
• The most common technique for producing
digital images
• Was the first digital technique available
commercially
• Exploits storage phosphors which emit light that
is proportional to the intensity of the X-rays that
hit it, when they are stimulated by a laser
beam
• Primary reason for being the most common
technique is that it is the cheapest (at least in the
short term)
– Old X-ray sets used for film-screen radiography can
be used, provided exposure factors and AECs are
adjusted for the new type of detector
CR Components
CR Stage 1: Image Capture
• Image receptor is a laser stimulable phosphor,
known as an image plate (IP)
• Capture image by irradiating an IP in the same
way as conventional film
– Does not need a new X-ray system when replacing
film-screen (just make sure automatic exposure
controls are re-calibrated)
• Typically ~40% of X ray photons are absorbed
• IPs retain majority of absorbed X-ray energy as
a pattern of electrons in meta-stable energy
states
– The spatial distribution of stored electrons is
equivalent to the pattern of absorbed x rays – latent
image
Electron Trapping
• Photon is absorbed by
an electron
• Electron can move
through conduction
band
• They can then be
trapped in Colour
Centres which forms
our latent image
CR Stage 2: Image Read Out
• Electrons are
actively stimulated
to release their
stored energy
• This is done by
scanning the IP
with an intense
laser beam
CR Stage 2: Image Read Out
• A red Laser is used as this matches the energy
gap between Colour Centre and conduction
band
• Light in the blue end of the visible spectrum is
emitted
• Hence, optical separation of input and output
light photons
– Means a colour filter can be used to prevent laser
photons contaminating the output signal
• Blue light photons are collected via a
photomultiplier tube and digital image is
produced
Read Out
• Stimulation of IPs
with laser causes
trapped electrons
to transfer to
conduction band
• These then relax to
the ground state,
emitting blue light
photons
How Does CR work?
Conduction band
Electron
traps
X Rays are
absorbed
Valence band
Photo stimulated luminescence
Conduction band
Red laser light
Blue
light
Valence band
The read/erasure cycle
• The image plate is removed from the cassette
inside the CR reader
• Scanning or laser achieved with a rotating mirror
• The light guide (with optical filter) directs the
emitted blue light to a photomultiplier tube,
which measures the intensity of the light
(proportional to the number of X-rays absorbed)
• Whilst repeatedly scanning the plate, it is moved
through the laser beam
• Once scanned, the residual signal is removed by
exposing the plate to a very bright light source
(erasure cycle)
• Takes about 30-45 s to read and erase an image
plate
CR Reader cycle
Analogue
output
Image quality
• Pixel size limits spatial resolution in CR
– For small plates (detail required) ~5.5 lp/mm
– Large plates (detail not essential) ~3.5 lp/mm
– (Film-screen ~8-12 lp/mm)
• Other limits to resolution in CR;
– Scattering of laser light in the phosphor layer results
in detected light from a larger area than expected
– Divergence of light emitted before detection
• Increases with thickness of phosphor
• Some phosphors have needle-like structure to guide light
(like an optical fibre), but quite brittle so not for general use
CR image quality
Image quality
• Image processing (e.g. edge enhancement) may
improve visibility of fine detail
• Contrast is determined by the image processing
and LUT that is applied (and the window/level
the user decides upon)
Digital Radiography
Digital Radiography
• Directly acquire the data in digital format (no
separate read-out phase like with CR)
– Improves throughput of X-ray systems – could be
important in chest clinic, mammo, etc
• Most expensive method, as it requires complete
dedicated X-ray system
• Main technologies:
– Phosphor coupled to a read out device – Indirect
conversion
– a-Se/TFT array – Direct conversion flat panels
Detective Quantum Efficiency
• DQE reflects the efficiency of photon detection
and the noise added
• Every photon detected and no noise added,
DQE = 100%
• DQE for DR ~65%
• DQE for CR and film-screen ~30%
• In principle, DR could be used with lower patient
doses as it is more efficient at using what is
available!
Indirect conversion
• Indirect conversion involves converting the X-rays
into visible light (in a phosphor), and detecting the
resulting light photons (akin to film-screen
radiography!)
• Either amorphous silicon (a-Si) photodiode TFT
array, or CCD for readout
• Sharpness limited by both pixel pitch of readout
array, and spread of light in phosphor
– Usually CsI(Tl) needle phosphors to focus light down to
the detector (like mini-fibre optics to minimise spread)
– Needle phosphors can be thicker (more efficient)
Indirect flat panels
• Phosphor => X-ray to light photons
• Light photons detected in photodiode array =>
light photons to electrical charge
• Read out by the amorphous silicon TFT array
(discussed after direct conversion)
• Can be manufactured as a single panel up 45 x
45 cm2, but in practice tend to be made up of
four smaller detectors ‘stitched’ together
– Tiled detectors
– Requires image processing and interpolation to cover
the join between panels
CCD detectors
• CCD light detectors (like in a camera) can only
be manufactured in relatively small sizes
– Usually need multiple CCDs to cover image area
(‘tiled detector’), or slot scanning technique
– Also thicker than flat panels due to the optics between
the phosphor and detector
Direct conversion flat panels
• Amorphous Selenium (aSe) is a photoconductor
– Converts X-rays directly to
electrons
• Deposited directly onto
amorphous silicon TFT
array
• No phosphor, hence no
light spread
• Resolution governed by
effective pixel pitch
Flat Panel Physics
• X ray photons reach the panel
• These are converted to an electrical charge
– Either in the phosphor/photodiode arrangement, or
photoconductor layer
• Charge read-out by the TFT array
• An image is instantaneously produced on the
computer screen
• Grey levels depend on the charge from each
sensor and the LUT/window/level settings
The TFT array
• Amorphous Silicon thin-film transistor array
• Transistors amplify electrical signals
• Electrical charge is stored in the TFT array until
release by applying a high potential
• Each row of detectors is connected to the same
activating potential (gate-line control), and each
column to a charge measuring device (read-out
electronics)
• The activating potential is applied row-by-row, so
the timing of the detected signal determines the
position of the pixel from which it originates
• Each pixel ~100 μm
The TFT array
Modern flat panels
And now for something a bit different…
Tomography and Dual Energy
Imaging
Tomography
• Conventional radiography superimposes
structures on the film/detector
• Results in;
– Inability to determine the depth of structures
– Limited ability to resolve the shape of structures
– Reduction in contrast
• Can resolve depth by using orthogonal
projections
• In other situations conventional or mechanical
tomography may be useful (or going full 3D with
CT, MRI, etc)
Tomography
• In tomography, only structures in a selected
plane of the patient, parallel to the film, are
imaged sharply – everything else above and
below appears blurred to the point they become
unrecognisable
• Blurring is produced by simultaneous movement
of at least two of;
– The tube,
– The film and/or
– The patient
• Actively exploit motion unsharpness that we
normally wish to minimise!
Linear tomography
• The tube and film/detector are linked by an
extensible rod, hinged about a pivot
• During exposure, the film/detector moves in a
straight line (e.g. right-to-left) along rails, whilst
the tube moves the other way
• Whilst moving, the tube is rotated so that the
central ray always points toward the pivot point
• Structures that are in the same plane as pivot
point (focal plane) will not move when projected
onto the film, producing a sharp image
• The projections of structures outside the focal
plane will move around on the film, and hence
are blurred in the final image
Linear tomography
Structures in focal plane
maintain same position on
the film = sharp image
Linear tomography
Structures above and
below the focal plane will
move from one end of the
film to the other = blurred
image
Linear tomography
• The further the object is from the focal plane, the
greater the blurring
– Structures lying within a plane of thickness t will be
sufficiently sharp to recognise
– Everything outside this will be too blurred and low
contrast
• Cut-height (the focal plane position) is adjusted
by lowering or raising the pivot
• t is controlled by the tomographic angle and
height of pivot (higher = thinner thickness of cut)
– Narrower angles reduces the degree of blurring, and
hence increase t (and vice-versa)
– 0° = conventional projection radiography = everything
in focus!
Linear tomography
• Typical angle = 40°, equivalent to t~3 mm
• However, thin slices and the spread of off-focus
anatomy over the whole film reduces contrast
– Use low kV (consistent with penetrating the patient)
– Most useful for high contrast structures e.g. bony
structures in the ear
• Patient dose is higher compared with projection
radiography
• Linear tomography less effective for linear
structures lying in the plane of movement
– Use alternative, more complex movements (e.g. circular,
elliptical, etc)
• Tomography equipment becoming less common
with the availability of CT
Tomography
Panoramic Dental
Radiography
Dual energy imaging
• A subtraction technique where images are taken
at high and low kV in rapid succession
– Used in general radiography (e.g. chests), CT, fluoro
– Low kV = high contrast between bone and soft tissue
(photoelectric effect)
– High kV = image contrast determined by tissue
density rather than atomic number (Compton scatter)
• Subtracting low kV image from high kV
minimises visibility of bone and improves softtissue contrast
– Remove ribs in chest radiography!
• Conversely, subtract high kV from low kV image
displays bony anatomy in greater detail
Dual energy imaging
Images (shown here) and good video of how it works can be found at:
http://www.upstate.edu/radiology/education/rsna/radiography/dual/
Dual energy imaging
Dual energy technology
• CR and DR based solutions available for
general radiography
• CR;
– Use two image plates with a Cu plate in between
– Acquire two images at the same time (no artefacts
due to motion)
– Cu plate filters the beam that leaves the first plate
(‘low energy’) to give ‘high energy’ image
– Little energy separation using this technique – results
in relatively low SNR
Dual energy technology
Dual energy technology
• DR;
– Low energy image (60 kVp), read-out, highenergy image, read-out
– Large energy separation = higher SNR
– Relatively long read-out times mean motion
artefacts are common (involuntary and
voluntary)
Dual energy technology
Dual energy spectra
Dual energy imaging
Motion artefacts
Lesion conspicuity
Dual Energy CT
• Technological challenges
– Constant data free from motion and contrast changes
– Need largest practical energy separation and detector
optimisation
• Commercial solutions
–
–
–
–
–
Multiple rotations at different kVp
Single spiral with alternating kVp
Dual x-ray source
Rapid switching kVp
Energy sensitive detectors (not double exposure)
Dual Energy CT
Dual Energy CT
Dual Energy CT
Dual Energy CT
Dual Energy CT
• Applications
–
–
–
–
–
Virtual non-contrast imaging
Bone segmentation and removal
Tissue typing?
Mono-energetic imaging – PET/SPECT apps?
Specialist breast CT?
• Not particularly proven yet – a technology looking for an
application?
Dual Energy CT – Manufacturers
• Siemens
– Dual source
– Now also provide single source solution (high and low
kVp scan)
• GE
– Rapid kVp switching
• Toshiba
– Single spiral with alternating kVp
• Philips
– Acquire two lots of data at high and low kVp, register
and use dual energy software
– BUT, they are developing dual-energy detectors (dual
layer and photon-counting)