PET Quantification Flashcards
Why is quantification theoretically possible in PET and not in SPECT?
The attenuation along a known line of response is independent of the unknown position of the emission along that line.
What is the purpose of quantification of PET?
- Lesion characterisation
- Response assessment
- Data reduction in trials & statistical analysis
- Dose optimisation/response assessment
- Testing drug targeting
- Radiotherapy target identification – functional target volumes with
reduced inter-observer variability
What are the types of corrections and calibrations required to calculate the activity concentration? (6)
- Randoms Correction
- Normalisation
- Dead Time Correction
- Scatter Correction
- Attenuation Correction
- PET Scanner Calibration
What are ‘Randoms’?
Randoms arise when two uncorrelated single detection events happen with sufficient temporal proximity to fall within the trues detection window. They can be estimated from singles count rates.
What is the singles count rate equation?
C_ij = 2 x τ x r_i x r_j
Where Cij is the rate (randoms per second) of random coincidences along line of response Lij, which connects detectors at channel i and j. ri and rj are the singles rates at these channels respectively and τ is the timing window for coincidences.
What are the different types of randoms corrections? (3)
- Estimation from singles rates - simply integrate the expression for randoms rate over the acquisition time T.
- Delayed coincidence channel - A duplicate window is set up at a delay from the coincidence window, but is open for the same time and exposed to the same dead time conditions etc as the coincidence window. This removes the correlated events (true coincidences), but gives an accurate rate of the randoms component
- Tailfitting - looking at distribution of random coincidences at the edges of the sinogram
What is the purpose of normalisation?
To compensate for variable sensitivities.
An individual correction factor is required for each line of response in the scanner.
What are the causes of variable sensitivity? (6)
Think of the Normalisation Equation
- Axial data summing and “mashing”
- Detector Efficiency
- Geometric & Solid angle effects
- Time window alignment
- Structural alignment
- Septa
What are the reasosn for detector efficiency variation?
- Position within the block; this is systematic and predictable and is called the
“block profile”. - Physical differences in the crystal and light guide
- PM tube gains varying
The latter two should be familiar from sensitivity map corrections of gamma cameras.
What are the geometric and solid angle effects?
These effects occur as lines of response are narrower closer to the edge of the field of view. The position and angle of incidence of the incoming photon affects the depth of mateiral along its path and hence the probability of interaction causing varying detection efficiencies. Reducing acceptance angles which causes a reduction in sensitivity.
How do you mitigate geometric and solid angle effects?
The effects are calculated analytically and incorporated into the recontruction algorithms, termed the ‘arc correction’.
What is the purpose of time window alignment?
Important for coincidence timing to be accurately synchronised.
Asychronicity offsets the coincidence window which reduces sensitivity to true coincidences, but not the
number or random events detected
What is the effect of minor structural misalignment of the modules in the ring system?
They affect the sensitivity of the lines of response.
Describe the method of direct normalisation?
- Illuminating all possible lines of response using a positron source
- An analytical correction is made for non-unifrom radial illumination
- Normalisation coefficients combining all the effects are taken to be inversely proportional to the counts acquired in each LOR
Positron source typically used is 68Ge
What are the problems with direct normalisation? (3)
- Long aquisitions
- Dependant on excellent source uniformity
- Scatter isnt comparable to patient scatter which can cause artefacts and bias
What causes a finite chance of counts occurring within the minimum time window of PET systems?
The random nature of radioactive decay causes a finite chance of counts occurring within this window, even at low activity levels.
How does higher activity affect the detection accuracy in PET scanners?
At higher activity levels, the chance of overlapping counts increases, leading to greater effects of system limitations.
