CT optimisation

Question 1

What is radiation dose, what does it include in CT?

What 4 things does image quality depend on in CT?

What are image artefacts?

Answer 1

1) The total radiation exposure delivered to the patient during a CT exam.
Includes: scout/planning image + the distance covered in the z-axis.

2) IQ in CT depends on 4 factors: image contrast, spatial resolution, image noise, artefacts

3) A system discrepancy between the CT numbers (Hounsfield units, HU) in the reconstructed image and the true attenuation coefficient of the object.

Question 2

What are the acquisition parameters in CT imaging?
[5]

Answer 2

• Tube potential - kV
• Tube current (product) - mA (not mAs in CT)
• Tube rotation time - seconds
• Detector configuration - slice thickness, number of slices, width of array
• Pitch - distance travelled per rotation compared to the slice thickness

Question 3

Tube potential (kVp)
- spectrum of…
- not all…
- what is the potential?
- what does the kVp affect?
- quality of the beam?
- how does filtration affect the quality?

Answer 3

• spectrum of photon energies
• not all photons have the same energy
• potential - highest energy the photons can be
• affects the number of photons reaching the detector.
  Increased kVp - photons have more energy - more transmission, less attenuation
• quality of the beam is a combination of average energy and range of energies
• filtration creates a good quality beam by narrowing the range.

High beam quality = low photon attenuation, low beam quality = high photon attenuation

Question 4

kVp and radiation dose
- higher kVp means what?
- However the opposite…?
- So to combat this, you would need to…

Answer 4

• higher kVp = higher radiation dose, lower kVp = lower radiation dose
• however using a lower kVp = means less photons reach the detector, decreased SNR
• so you need to increase the number of photons (mAs) which in turn increases radiation dose
• Using a higher kVp = can allow a lower mAs = reducing dose.

Question 5

kVp and image quality

In terms of noise, contrast, artefacts:

• high kVp means what?
• low kVp means what?
• how does a beam get harder?

Answer 5

• high kVp = less image noise, reduced image contrast, reduced beam hardening artefacts
• low kVp = more image noise, more image contrast, increased beam hardening artefacts
• If you reduce the number of lower energy x-rays (they get attenuated), the average energy increases - beam hardening, average energy increases as beam travels through patient.

Question 6

What is the most common image artefact related to kVp/tube potential?

What is the process?

What are the two types of beam hardening?

Answer 6

Beam hardening

• x ray beam becomes harder as it passes through tissue
• lower energy photons are removed (attenuated)
• average beam energy increases (beam becomes harder)
• more transmission with a harder beam
• more photons reach the detector than expected
• back projection of lower value CT numbers

Two types: cupping and streaking

Question 7

One type of beam hardening - cupping.
Why does it occur?
What does it look like on an image?

Answer 7

• photons pass through the edge of a uniform structure more easily
• less material at the edge
• more material at the center - beam becomes harder and has a higher average energy when it reaches the detector (less attenuation)

–> reduction in CT HU (hounsfield units) in the centre of the image
(less attenuation in centre, more darker, more attenuation at edges, whiter)

Question 8

Another type of beam hardening - streaking
Why does it occur?
What does it look like on an image?

Answer 8

• this occurs when the x-ray beam passes through an area of heterogenous density - such as the skull
• irregular and dense structures mean that the degree of beam hardening is not the same in all projections
• this causes streak artefacts
• beam hardening artefacts can also occur because of objects outside the scan field of view (such as patients having their arms down by their side)

Question 9

What artefacts can cause streaking on a reconstructed image?

What can be done to reduce this?

How does the scanner work?

What 3 ways do manufacturers use to reduce beam hardening?

Answer 9

• Metal artefacts such as dental implants, prosthetics and surgical clips

Metal artefact reduction:
- smaller reconstructed slice thickness and an increased kVp can reduce this
- beam hardening correction can be required:

The scanner software uses a process known as interpolation to predict what the values should be either side of the very dense object
- metal object that is beyond the HU range recognised by the scanner.

1) Filtration - remove lower energy photons from x-ray beam (bow-tie filters)
2) Calibration correction - phantoms of different sizes are used to calibrate detector response
3) Beam hardening correction software (MARS) - if the artefacts can’t be removed

Question 10

On an image, what appear as dark streaks and what appear white?

Answer 10

• beam hardening: dark streaks on image
• photon starvation: white on image

Question 11

Tube current (mA)
- what two things combine to make mAs?
- mAs has a linear relationship with what?

Answer 11

• mA = tube current + s = rotation time -> combine to give the effective mAs.
• mAs has a linear relationship with dose.
  Dose is proportional to the number of photons produced (mA) and the time (s) for a single tube rotation
  (when changing mAs is in isolation of all other acquisition parameters)
  x-ray beam quantity. more photons = more dose, less photons = less dose.

Question 12

What is ATCM?
What is tube current adjusted on?
However what is to note about ATCM implementations?
What occurs after the scout?

Answer 12

• automatic tube current modulation
• tube current (mA) is adjusted on the basis of the density of the body area that the photons are passing through and/or size of the patient.
• not all implementations of ATCM account for patient size, some work by controlling level of noise in resultant image (GE).
• after the scout, machine knows what density is where, automatically tube will change mA (and consequently dose) as it is rotating. Some radiologists don’t like this, as it compromises image quality. If you set a noise value, it’ll try to keep the noise the same.

Question 13

What is a scout in CT imaging?

Why are they crucial?

What can they reveal?

A weakness and a strength?

Answer 13

In CT imaging, the term “scout” refers to initial images that are taken before the main CT scan.

These images are crucial for ensuring that the correct anatomy is covered and for optimising the radiation dose during the main scan.

They can reveal unexpected findings, such as pregnancy, and are used to plan the scan accurately.

While they may not always be of diagnostic quality, they can provide valuable information for setting scan parameters and managing patient care.

Question 14

Tube current product (mAs) and Image Quality

In terms of noise:
- High mAs..?
- Low mAs…?

Answer 14

• High mAs = more photons reaching the detector is associated with less image noise
• Low mAs = less photons reaching the detector is associated with more image noise

Question 15

Image quality is frequently quantified in terms of?

Halving the mAs ->

Answer 15

• Signal to noise ratio (SNR) = the ratio of true signal (anatomy) against noise (quantum mottle)
• Contrast to noise ratio (CNR) = the system’s ability to resolve increasingly small low contrast objects.

Halving the mAs - > can increase image noise by 50%

Question 16

Tube current product and artefacts

What is photon starvation?

Answer 16

• occurs in dense areas with high attenuation, dense areas such as the shoulders
• the result is that not enough photons reach the detector in certain projections
• this causes a streak on the image as there is no data to reconstruct
• as this is a 3D data set, the impact of a dense object impacts on the entire projection

Question 17

What are different ways that photon starvation can be reduced?

Answer 17

• ATCM - increase the number of photons reaching the detector in the REQUIRED projections
• Increased mAs - increase the number of photons reaching the detector in ALL projections
• adaptive filtration software - correction method to smooth the attenuation profile in areas of high attenuation prior to image reconstruction

Question 18

Insufficient photons + Artefacts

What is undersampling?
What does it lead to?

Answer 18

• related to the number of projections used, but it still occurs because of a lack of photons
• the number of projections determines the reconstructed image quality
• a large gap in projections (undersampling) can lead to a misregistration of information
• particularly applicable to sharp edges and small objects
• this leads to aliasing

Question 19

What is aliasing in CT imaging?

Answer 19

• refers to the distortion of images caused by inadequate sampling of projection data. Occurs when the sampling rate is insufficient to capture high-frequency components, leading to artefacts such as streaks, lines and Moire patterns.

Question 20

What are the 2 solutions to aliasing?

Answer 20

1) Flying focal spot
- involves imaging the patient twice
- this can be done in the rotational direction or the z-axis direction

2) Quarter detector shift
- images the patient in a slightly different plane to improve statistics
- x-ray tube and detectors are moved out of line by a 1/4 of a detector

Question 21

Tube rotation time (s)
- what is this?
- how does this relate to dose?
- how does the speed of the rotation relate to patient?

Answer 21

• the time for 1 complete rotation of the x-ray tube around the patient
• mA x s = mAs, 90 mA x 0.5s = 45 mAs
• slow rotation - patient cannot hold their breath, a fast rotation - patient can hold their breath

Question 22

Acquired slice thickness
- what does this depend on?

Answer 22

depends on:
- detector configuration
- beam width
- beam collimation

Question 23

Acquired slice thickness, detector configuration
- what is the beam width?
- what is slice thickness?
- how does this relate to spatial resolution?
- to CNS ratio? how to improve it?

What is an example of advancing detector technology?

Answer 23

• beam width: size of the x-ray field in the z axis
• slice thickness: number of detector rows used within the beam width
• the more detector rows, the thinner the slices = better spatial resolution
• but the lower number of photons in each row = worse contrast to noise ratio
• can increase mA to improve CNR (balance between spatial and contrast resolution and dose)
• e.g. moving from 1x10mm to 16x1.25mm

Question 24

Acquired slice thickness

• radiation dose?
• image quality?
• image artefacts?
  [+ what effect links to this]

Answer 24

• you can select whatever detector configuration you like, but if you are using ATCM - this will change the mA automatically
• the same number of photons are distributed between differing number of detectors. The image can be noisier when a larger number of detector are used. mAs may need to be increased.
• if a larger slice thickness is used to acquire the data, this can limit the chance of detecting small pathology. Difficult to see a 5mm lesion with a 1x10 mm slice. [PARTIAL VOLUMING EFFECT]

Question 25

Acquired slice thickness - A smaller detector array allows....but... - A larger detector array allows...but...

Answer 25

- A smaller detector array allows better spatial resolution but decreased contrast resolution if the mA is kept constant - A larger detector array allows better contrast resolution but at the expensive of a loss of spatial resolution

Question 26

What is pitch in CT imaging? What is the difference if the pitch is <1.0, =1.0, >1.0?

Answer 26

- table movement per tube rotation divided by slice thickness - for each tube rotation, the table moves a set distance. Pitch < 1.0 = helical acquisition mode, data is overlapping, it is being sampled more than once. Pitch = 1.0 = axial acquisition mode, step-and-shoot mode Pitch > 1.0 = helical acquisition mode, data is not all covered and interpolation is required to 'fill in the gaps'

Question 27

How does the pitch relate to: - radiation dose - image quality - image artefacts

Answer 27

- radiation dose: pitch <1 = higher dose, pitch >1, lower dose - image quality: pitch <1.0 = better quality, less noise, better sampling statistics pitch >1.0 = lower quality, more noise, worse sampling statistics - image artefacts: pitch < 1.0 = less artefacts, reduced partial volume pitch > 1.0 = lower quality, increased partial volume

Question 28

Pitch and radiation dose - As you increase pitch.. - This is because.. - Pitch can be used to work out..

Answer 28

- As you increase pitch, radiation dose is decreased - This is because you are covering a greater distance in the z-axis for each rotation of the tube - pitch can be used to work out the effective mAs. effective mAs = mAs/pitch

Question 29

Pitch and Artefacts - helical artefacts in the z-axis occur when.. - what could it also be, when is it common?

Answer 29

- occur when there is a fast changing area of anatomical structure e.g. the top of the skull - there can also be interpolation errors that can lead to a windmill effect - more common with a high pitch.

Question 30

Matrix size and FOV In terms of resolution: - larger matrix = - smaller matrix = - to maintain contrast resolution..

Answer 30

- larger matrix = smaller pixel size, better resolution - smaller matrix = larger pixel size, worse resolution - higher mAs is required to maintain contrast resolution -> increased radiation dose

Question 31

Matrix size and FOV - what has a big impact on pixel size? - how do you calculate pixel size?

Answer 31

- the reconstructed field of view has a big impact on pixel size - pixel size = FOV/matrix size e.g. 600/512 = 1.7mm

Question 32

FOV and artefacts - what could still be within some of the projections? - Ideally they should be? - The CT scanner...

Answer 32

- even though the arms are outside of the reconstruction FOV, they are still within some of the projections - arms must be within the FOV or (ideally) raised so that they don't obstruct the x-ray beam - otherwise the CT scanner does not have the information to account for this.

Question 33

Image Reconstruction - what is the reconstructed slice? you can set it when..but..? - what impact on radiation dose? - what about image quality? - what about image artefacts?

Answer 33

- related to the detector configuration and is how you reconstruct the data - you can set it at the acquisition stage but it is independent - the amount of distance covered in the z-axis by each slice - radiation dose: no impact as it is a reconstruction process - image quality: thinner slices - better resolution and less artefacts but more noise / thicker slices - poorer resolution and more artefacts but less noise - artefacts: the partial volume averaging is greater with a thicker slice

Question 34

Reconstructed slice and artefacts - what is the problem with thicker slices? - what is the result of this? - when can a partial volume (cone beam) artefact occur? - What is it characterised by? - How to overcome it?

Answer 34

- partial volume averaging, different tissues within the same voxel, an averaged pixel value is the result. - can occur when dense objects are not in the center of an object and are not included in all projections - this is characterised as a shading artefact - scanning with thinner slices can overcome this type of artefact

Question 35

Explain THINNER and THICKER reconstructed slices' qualities in terms of: - dose - noise - sampling - partial volume effect - data covered - interpreting data

Answer 35

THINNER - higher dose (if mAs adjusted) - more noise - mA needs increasing to maintain SNR and CNR - reduced partial volume effect - cover all data (potentially) - more data to interpret (reporting) THICKER - lower dose - less noise - sampling statistics better (mA can be kept lower) - increased partial volume effect - you may miss some data - less data to interpret (reporting)

Question 36

What do these two terms mean? - Reconstructed slice - Reconstruction interval

Answer 36

Reconstructed slice - the image you look at - the distance in the z-axis that the slice covers - the average voxel value of this z-axis distance Reconstruction interval - the gap between the images - this determines whether you: -> look at all the data -> look at overlapping data -> miss some data

Question 37

Helical interpolation - the table moves as... - this means that the scanner.. - the same anatomy..the process is..

Answer 37

- helical reconstruction algorithms and z-axis filtering - the table moves as the x-ray tube continually rotates - this means that the scanner has to work out what the correct pixel values should be - the same anatomy will not be detected by the same row of detectors as it travels around the patient Process is called interpolation

Question 38

What are patient based and scanner based artefacts that could come up?

Answer 38

Patient based - movement - contrast media Scanner based - ring artefacts

Question 39

One artefact is motion. - what could be used to correct it? What are the two types of modes?

Answer 39

- motion correction software could be used to correct for this error - over-scan mode = an extra 10% added to rotation to increase the number of projections - under-scan mode = less than a full rotation to reduce the influence of motion (resolution may be compromised)

Question 40

Another artefact - ring. What is a ring artefact? What does this lead to? Solution?

Answer 40

- a faulty detector giving bad data throughout the rotation - can impair the diagnostic quality of the image - detector needs replacing or recalibarting - software can reduce variation in detector response

Question 41

Radiation Dose Measurement/Recording CTDI vol (mGy) - the dose.. - the average.. - what does it not indicate and why?

Answer 41

- the dose delivered for a single CT slice. - the average dose within the scan volume for a standardised (CTDI) phantom. - it doesn't indicate the total energy deposited into the scan volume because it is independent of the length of the scan - not that useful for ATCM either

Question 42

Radiation Dose Measurement/Recording DLP (mGycm-1) - what does this represent? - accounts for? - how to calculate it?

Answer 42

DOSE LENGTH PRODUCT - represents the overall energy delivered by a given scan protocol - accounts for the entire length of the scan, not just a single slice - calculated as: DLP = CTDI vol x scan length (cm)

Question 43

Radiation Dose Measurement/Recording Effective Dose (mSv) - what does this account for? - how is it done? - it's an estimate for..rather than?

Answer 43

- accounts for the biological effects from radiation, accounting for the sensitivity of the tissue or organ - done by using weighting factors for different organs - a good comparison between exams - an estimate whole-body dose for a whole exam rather than a true measure for that patient

Question 44

What must you consider when optimising protocol?

Answer 44

- consider all acquisition parameters - consider the clinical question - trade-off between radiation dose and image quality - try to minimise image artefacts - be ALARP