Topic 22: Emission Tomography Quantification and motion correction

Question 1

What are the units of the reconstructed images of CT?

Answer 1

Hounsfield Units

Question 2

What are the units of the reconstructed images of PET/SPECT?

Answer 2

Standard uptake value = tracer concentration / (injected_activity/patient_weight)

Question 3

ROI Measures

Answer 3

max, mean, volume and total uptake

Question 4

What are typical SUV clinical measures that they like to use in hospitals?

Answer 4

Region independent

SUVmax is used - which is ROI independent because choosing an ROI is subjective.

SUVpeak - average SUV within a 1cm³ spherical ROI and move that around in your image and find the maximum. (the advantage is that it is not sensitive to noise)

Region dependent

• hand-drawn or (semi)automated ROI
• Current methods
• SUV mean (=ROI mean)
• Metabolic Tumour Volume (=ROI vol)

Question 5

What are partial volume effects?

Answer 5

“When an object partially occupies the sensitive volume of an imaging instrument (in space or time) the observed signal will be reduced.”

resolution effects reduce quanitification resolution.

Question 6

What is the recovery coefficient?

Answer 6

Ratio of measured ROI mean or max over true value.

Question 7

What 2 types of Partial Volume Effects are there?

Answer 7

Spill-over

• activity from inside the ROI appears outside

Spill-in:

activity from outside the ROI appears inside

Effect:

“Recovery coefficient” depends on the activity outside the ROI

Question 8

How can we correct for PVE?

Answer 8

Deconvolution methods

• Generic
• Sensitive to noise

ROI based PV correction methods

• objects with known size
• Usually assume uniform uptake in every region

Hybrid methods

• future

Question 9

Region-based PV correction: geometric transfer matrix?

What the fuck is this boring shit.

Answer 9

1. color all tha different parts of yer image into different structures!
2. then smooth allll the regions so that they have the same emission resolution.
3. defined contribution of activities in structures to each region
4. Solve for activity in each region given measurement

only works for uniform activity in regions and if they have been divided accurately

Question 10

What about resolution modelling during reconstruction?

Answer 10

Blurring can be incorporated in the system model. “PSF modelling”

• accounts for resolution
• involves higher uncertainty in the model

Question 11

Advantages and disadvantages of tracer kinetic modelling?

Answer 11

Advantages

• Extracted parameters are independent of delivery, therefore more reliable
• Allows incorporation of biological information

Disadvantages

• longer acquisition time
• Results can depend on the accuracy of model

Question 12

How does patient motion impact our images? when they cant bloody lie still and not breath for a min ffs.

Answer 12

Puts in bloody Artifacts ofc

Diagnostic uncertainty - mate i cant diagnose you yer moving too much

radiation treatment -

Quantisation - (when converting from ac to dc the error fucks up)

Reproducibility - i cant reproduce this shit mate.

Prevents use of anatomical information

Question 13

Effect of motion in PET/SPECT

Answer 13

Blurring

Mismatch with CT/MR used for density estimation

Question 14

How to reduce effects of motion

Answer 14

1. Minimise motion
2. Gating and time frames

Split data in different “motion states”

3. Motion correction

Combine data to reduce noise

Question 15

Gating?

Answer 15

break up the breathing cycle in to expiration and inhalation and combine their sin curves

Question 16

why dont people really use gata data in clinics?

Answer 16

noise increases

Question 17

solutions for respiratory motion correction

Answer 17

1. estimate motion between gates from either
   - gate PET images
   - gated CT images
2. combine PET data from all gates
   - add registered PET images
   - incorporate motion into image reconstruction and estimate single motion-free image