SPECT gamma camera QA

what are the 3 stages of quality assurance

1. acceptance testing

2. commissioning

3. quality control

acceptance testing

ensuring equipment meets purchasing specification → done according to the national electrical manufacturer’s association (NEMA)

• sets out specific testing conditions to allow performances to be compared between different systems and manufacturers

• test conditions are designed to be replicated by manufacturers in a non lab condition

commissioning

making sure the system is ready for clinical use by measuring

• basic QC values

• optimising system and setting up protocols

quality control

multiple guidelines available by: IPEM…

periodic testing to find performance declines and ensure consistent performance

quality assurance vs quality control

• assurance: setting rules and standard for product quality

• control: inspection and testing of the product against the pre-set standards

what do you need to know for QC testing

• radionuclide specifications: energy peak(s)

• collimator type

• energy resolution and energy window

• camera area: useful field of view vs central FOV

• pixel/matrix size

• count rate/density

• whether scatter is included

difference between useful and central FOV

• useful: edges have degraded imaging capabilities

• central: 75% of the UFOV → best part of the camera with minimal degradation

intrinsic measurements

measurements performed without collimators, therefore:

1. there is no collimator blurring

2. smaller activity sources can be used

system measurements

measurements performed with collimators

1. mimics IRL

2. results are collimator specific

3. scatter is included: as collimators are used to decrease scatter

how does pixel/matrix size affect image

• large matrix → small pixels and vice versa

• balance between spatial resolution and sensitivity +noise

  • smaller pixels → increased resolution but decreased sensitivity → leads to more noise

  • noise = $\sqrt{N}$

what factors lead to poor image

1. non-uniform detector response

2. poor scatter rejection

3. spatial resolution

4. detector sensitivity

5. count-non-linearity

what can lead to detector non-uniformity

1. different PMT gain

2. positioning non-linearity

3. NaI(TI) crystal imperfections

4. collimator damage → septa damage. angulation errors, uneven hole size

5. PMT and crystal decoupling

intrinsic uniformity measuring

1. use a uniform flux source or phantom placed in a lead shield

2. opening is covered using a copper plate → reduces some of the scattering → acts as a point source

3. place the source at a distance producing an projection greater than 5 UFOV

4. Small pixel size is unnecessary as we are measuring the underlying system performance → large pixels are used to reduce noise

5. 9 point smoothing curve is then applied

why is a distance of 5 UFOV used in intrinsic uniformity testing?

due to inverse square law $I=\frac{Power}{4\pi r²}$

at close proximity:

• there is a large difference in distance from the source to the edge of the detector compared to the centre

• causes a massive drop in intensity at the edges

at 5 UFOV:

• difference from source to detector edge and centre is almost negligible as all of the rays are nearly parallel

systemic uniformity test

performed daily

a uniformly distributed activity source is used

• As rectangular technetium phantoms are difficult to produce a Co57 source absorbed into resin is used instead

• Has similar peak as Tc: 122 compared to 140keV

what are the UFOV and CFOV equations (uniformity types)

1. integral uniformity → measures global non-uniformity → max and min pixel values anywhere within the FOV

2. differential uniformity → measures regional non-uniformities → max and min are anywhere within a 5×5 pixel range

both uniformities use the same equation

$$uniformity=\frac{max-min}{max+min}*100$$

good uniformity is anything bellow 5%

what affects intrinsic detector resolution

1. statistical fluctuation of light photon distribution between photocathode and PMT amplification → causes no. of photons to not be proportional to radiation

2. spread of light as it travels to the PMT → loss of spatial resolution

3. multiple scattering within detector

how is intrinsic resolution measured

1. use the same set up as intrinsic uniformity testing

2. place 2 mask in front of the source with 1mm slits 30mm apart

3. we know we should be seeing perfectly straight lines but on the image they may appear distorted

4. measure the deviation of peak counts from the line of best fit

what is differential spatial linearity and integral spatial linearity

• Differential Spatial Linearity: Standard deviation of difference between peak locations and fit (typically < 0.2 mm)

• Integral Spatial Linearity: Maximum difference between peak location and fit (typically < 0.4 mm)

how is intrinsic system resolution improved

1. use thinner NaI(Ti) crystal → however decreases stopping power

2. use more smaller PMT

3. use higher energy gamma photons → more light is produced

4. use signal processing and position calculation algorithms

measure system resolution

•  small sources are placed 10cm from the collimator

  • system spatial resolution depends on distance from the detector

  • Measure at 10cm for consistency and because it is realistic

  • 0cm measurements can be effected by position of holes relative to source

or use a bar phantom:

• Cobalt-57 flood emitting through a Bar Phantom (strips of lead in resin)

• An easy test that can be performed more regularly

• Predominantly a qualitative assessment but can be performed quantitatively

• Bar phantom can be placed directly on crystal for intrinsic measurements

us $R_c$ equation in notes

what does system planar sensitivity depend on

1. crystal efficiency

→ increased with thickness and density of crystal

→ decreases with photon energy

2. energy window width

3. collimator height and opening width

$$sensitivity = \frac{counts~per~sec}{activity~in~source}=\frac{Cps}{MBq}$$

how is system planar sensitivity measured

a source with known amount of activity is scanned for a known amount of time

what is dead time

finite amount of time needed to process an event, depends on → light decay in the crystal + electronic processing time

• if 2 gamma rays enter the detector in an interval less than the dead time → one or both gamma rays are not detected (lost)

• sensitivity at high count rate decreases

• sets a maximum limit to detectable count rate

what are the type of dead times

1. paralysable:  When gamma ray enters the detector within the deadtime of the previous event, the deadtime ‘clock’ is restarted (dead time is prolonged_

   1. most gamma cameras are paralysable

2. Non- paralysable: When gamma ray enters the detector within the deadtime from the previous event, the gamma ray is ignored. Deadtime ‘clock’ is not extended

how is dead time measured

• Use a high activity decaying source which you acquire the count rate

• Look at the linear response region and then you extrapolate the data to higher count rates

• Measure 10%, 20% ect losses

• Almost all gamma camera imaging is performed in linear zone