Lecture 7: 12/12
PET quan)fica)on and tracer kine)c modeling for oncology
Sta$c PET imaging
- Looking at an image -> to make conclusions
- 6 cycles chemotherapy -> metasta7c lesion; take it up the therapy?
o You have to do quan7fica7on = measuring something
o Matrix on what you see -> done by SUV = standard uptake value = uniform
value that has no unit
- Pa7ent gets injected -> wait, 30 min for example -> brain
o Why wai7ng?
§ Tracer need to be taken up by the 7ssues
§ The non-specific uptake needs to washed out
§ Your ac7vity reaches a plateau = steady state where you look to the
tracer accumula7on in a region.
o How long wai7ng?
§ Needs to be consistent
§ Depend on tracer and applica7on
- Longer wai7ng -> gives image contrast
- Example: FDG
o First radioac7vity gets into the blood compartments
o FDG enters the cell by the glucose transporter -> glucose undergoes
hexokinases = phosphate aWached -> it can not be used by the mitochondrion
and it can not be converted -> stuck in the cell
- No temporal informa7on + mul7ple bed posi7ons to scan the whole pa7ent ->
difficult
SUV quan$fica$on
Ac#vity concentra#on
- Bq = one radioac7ve assay in one second
- Older units; curies
o 1 ci = 37 GBq
è Units are important
- Everything needs to be appropriate
o Correct for randoms, correct for aWenua7on, proper calibra7on of the system.
- What are the parameters that can vary?
SV quan#fica#on
- SUV takes these parameters into account
- Simple normaliza7on factor
- Example
- Radioac7ve decay
o Time point (post injec7on); at the moment we do the scan we want to know
the radioac7vity -> a tracer decays over 7me => needs to be taken account
o Calculate SUV -> calculate at the same 7me, normally at the beginning of the
scan
o PET tracers; short lives <-> SPECT tracers; longer half-lives
, § Zirconium 89 = long half live PET tracers
Sta#c quan#fica#on
- SUV; distribu7on volume normaliza7on factor
o Inject 5mCi in a baby compared to a big guy -> difference
o Body weight (kg)
§ Tracer is not uniform distributed over the pa7ent -> not correct
normaliza7on.
o Alterna7ve distribu7on volumes
§ Lean body mass
§ Body surface aera
Glucose
- Glucose level at the moment of the scan
- FDG compe77ve substrate with glucose
o The more glucose -> the more compe77on -> the less uptake => correct
standard uptake value with plasma glucose levels
- Higher plasma glucose levels -> more compe77on -> less FDG
Tumor SUV
- Tumor grows that much and quick -> central necro7c parts, because lack of oxygen
- Different quint of suv
o Mean suv
o Suv at 80 or 90%
o Suv max
o Suv peak
è Oncological applica7on
- Par7al volume effect
o If you look to a small tumor compared to a big tumor you will assume that the
darker spots take the most uptake -> but same ac7vity concentra7on? But
why darker? Par7al volume effect: every point spread out by gaussian curve->
7ny sphere gets smeared out by spa7al resolu7on -> you lose the intensity
due to the limit of the spa7al resolu7on
Tumor to … ra#o
- Normalize to blood
- Normalize to live
- Tumor to background
Problems sta#c imaging
Example 1
- Which one is prolifera7on the most, the blue or the red?
Example 2
- Tracer not slowly metabolized in the liver -> more uptake in the tumor due it become
longer available.
Solu7on
- Measure concentra7on in intact plasma
- Measure temporal; profile of ac7vity distribu7on
è Dynamic imaging
Dynamic PET imaging
- Inject -> wait -> scan
PET quan)fica)on and tracer kine)c modeling for oncology
Sta$c PET imaging
- Looking at an image -> to make conclusions
- 6 cycles chemotherapy -> metasta7c lesion; take it up the therapy?
o You have to do quan7fica7on = measuring something
o Matrix on what you see -> done by SUV = standard uptake value = uniform
value that has no unit
- Pa7ent gets injected -> wait, 30 min for example -> brain
o Why wai7ng?
§ Tracer need to be taken up by the 7ssues
§ The non-specific uptake needs to washed out
§ Your ac7vity reaches a plateau = steady state where you look to the
tracer accumula7on in a region.
o How long wai7ng?
§ Needs to be consistent
§ Depend on tracer and applica7on
- Longer wai7ng -> gives image contrast
- Example: FDG
o First radioac7vity gets into the blood compartments
o FDG enters the cell by the glucose transporter -> glucose undergoes
hexokinases = phosphate aWached -> it can not be used by the mitochondrion
and it can not be converted -> stuck in the cell
- No temporal informa7on + mul7ple bed posi7ons to scan the whole pa7ent ->
difficult
SUV quan$fica$on
Ac#vity concentra#on
- Bq = one radioac7ve assay in one second
- Older units; curies
o 1 ci = 37 GBq
è Units are important
- Everything needs to be appropriate
o Correct for randoms, correct for aWenua7on, proper calibra7on of the system.
- What are the parameters that can vary?
SV quan#fica#on
- SUV takes these parameters into account
- Simple normaliza7on factor
- Example
- Radioac7ve decay
o Time point (post injec7on); at the moment we do the scan we want to know
the radioac7vity -> a tracer decays over 7me => needs to be taken account
o Calculate SUV -> calculate at the same 7me, normally at the beginning of the
scan
o PET tracers; short lives <-> SPECT tracers; longer half-lives
, § Zirconium 89 = long half live PET tracers
Sta#c quan#fica#on
- SUV; distribu7on volume normaliza7on factor
o Inject 5mCi in a baby compared to a big guy -> difference
o Body weight (kg)
§ Tracer is not uniform distributed over the pa7ent -> not correct
normaliza7on.
o Alterna7ve distribu7on volumes
§ Lean body mass
§ Body surface aera
Glucose
- Glucose level at the moment of the scan
- FDG compe77ve substrate with glucose
o The more glucose -> the more compe77on -> the less uptake => correct
standard uptake value with plasma glucose levels
- Higher plasma glucose levels -> more compe77on -> less FDG
Tumor SUV
- Tumor grows that much and quick -> central necro7c parts, because lack of oxygen
- Different quint of suv
o Mean suv
o Suv at 80 or 90%
o Suv max
o Suv peak
è Oncological applica7on
- Par7al volume effect
o If you look to a small tumor compared to a big tumor you will assume that the
darker spots take the most uptake -> but same ac7vity concentra7on? But
why darker? Par7al volume effect: every point spread out by gaussian curve->
7ny sphere gets smeared out by spa7al resolu7on -> you lose the intensity
due to the limit of the spa7al resolu7on
Tumor to … ra#o
- Normalize to blood
- Normalize to live
- Tumor to background
Problems sta#c imaging
Example 1
- Which one is prolifera7on the most, the blue or the red?
Example 2
- Tracer not slowly metabolized in the liver -> more uptake in the tumor due it become
longer available.
Solu7on
- Measure concentra7on in intact plasma
- Measure temporal; profile of ac7vity distribu7on
è Dynamic imaging
Dynamic PET imaging
- Inject -> wait -> scan
