Instrumentation noise equivalent exposure INEE for routine quality assurance: INEE measurements on a

1
Instrumentation noise equivalent exposure (INEE)
for routine quality assurance INEE measurements
on a clinical flat panel detector

• T. Szczykutowicz, A. Kuhls-Gilcrist,
• D. Bednarek, S. Rudin

2
Instrumentation Noise Equivalent Exposure (INEE)

• INEE provides a measure of total instrumentation
  noise in terms of detector entrance exposure.
• INEE model N2 k(Exposure INEE)
• Noise (N2) is the variance in pixel gray values
  from flat-field images
• k is a proportionality constant relating pixel
  gray values to units of exposure (µR)
• If Exposure gt INEE ? Quantum Limited
• If Exposure lt INEE ? Instrumentation Noise
  Limited

3
Instrumentation Noise Equivalent Exposure (INEE)

• INEE provides a measure of total instrumentation
  noise in terms of detector entrance exposure.
• INEE model N2 k(Exposure INEE)
• Noise (N2) is the variance in pixel gray values
  from flat-field images
• k is a proportionality constant relating pixel
  gray values to units of exposure (µR)
• If Exposure gt INEE ? Quantum Limited
• If Exposure lt INEE ? Instrumentation Noise
  Limited

4
Instrumentation Noise Equivalent Exposure (INEE)

• INEE provides a measure of total instrumentation
  noise in terms of detector entrance exposure.
• INEE model N2 k(Exposure INEE)
• Noise (N2) is the variance in pixel gray values
  from flat-field images
• k is a proportionality constant relating pixel
  gray values to units of exposure (µR)
• If Exposure gt INEE ? Quantum Limited
• If Exposure lt INEE ? Instrumentation Noise
  Limited

5
Instrumentation Noise Equivalent Exposure (INEE)

• INEE provides a measure of total instrumentation
  noise in terms of detector entrance exposure.
• INEE model N2 k(Exposure INEE)
• Noise (N2) is the variance in pixel gray values
  from flat-field images
• k is a proportionality constant relating pixel
  gray values to units of exposure (µR)
• If Exposure gt INEE ? Quantum Limited
• If Exposure lt INEE ? Instrumentation Noise
  Limited

6
Procedure for Quality Assurance

• Remove anti-scatter grid
• If grid cannot be removed, the grid transmission
  factor can be used to acquire the detector
  entrance exposure
• Removing the grid eliminates grid line noise
  patterns

7
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)

8
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)
• Measure detector entrance exposure for range of
  mAs values

9
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)
• Measure detector entrance exposure for range of
  mAs values
• Disable image processing

10
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)
• Measure detector entrance exposure for range of
  mAs values
• Disable image processing
• Acquire image sequence

11
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)
• Measure detector entrance exposure for range of
  mAs values
• Disable image processing
• Acquire image sequence
• Calculate the variance for each image sequence
  for each mAs value used and plot versus detector
  entrance exposure

12
Procedure for Quality Assurance

• Remove anti-scatter grid
• Add filtration to get a standardized spectrum
• (e.g. RQA IEC 62220-1)
• Measure detector entrance exposure for range of
  mAs values
• Disable image processing
• Acquire image sequence
• Calculate the variance for each image sequence
  for each mAs value used and plot versus detector
  entrance exposure
• Fit plot using linear regression and obtain the
  slope and intercept INEE Intercept / Slope

13
Varian PaxScan 2020 FPD
14
Varian PaxScan 2020 FPD

• 1 frame / sec was used to minimize temporal
  effects.
• Spectrum Used 70kV, 51.2 mm of AL (includes
  inherent tube filtration)

15
Typical Result for Digital Angiography (DA) Mode
16
Typical Result for Digital Subtraction
Angiography (DSA) mode
17
Comparison of DA and DSA Modes
DA
DSA
18
Comparison of DA and DSA Modes
Integration Capacitance 0.5 pF
DA
Integration Capacitance 4 pF
DSA
19
INEE Stability

• Over a four month period, our system was tested
  three times and showed maximum deviations of 5
  and 2 for DA and DSA modes.

20
Effects of Image processingDA Mode

• In DA mode, non linear image processing resulted
  in unusual variance behavior.

21
Effects of Image processingDA Mode

• Even with linearization of the mean signal to
  detector entrance exposure and recalculation of
  the variance, the resulting trend was still found
  to be non linear.

• DA mode, with (non-linear) image processing
  enabled, yielded no INEE.

22
Effects of Image processingDSA Mode

• The un-subtracted images were used to calculate
  the mean signal and variance.
• In DSA mode, no linearization was required (i.e.
  the image processing was linear).
• INEE 39.2 0.4 µR.

23
Summary

• INEE was determined for DA and DSA modes their
  differences could be due to integration
  capacitance differences.
• DA and DSA modes implemented different image
  processing. In DA mode, the imaging processing
  was non-linear and the INEE could not be
  measured. In DSA mode, the image processing was
  linear and had no effect on the INEE.

24
Conclusions

• The INEE metric was applied to a clinical FPD to
  obtain the minimal exposure at which the detector
  was quantum noise limited.
• If, Exposure lt INEE detector performance is
  instrumentation noise limited
• If, Exposure gt INEE detector performance is
  quantum noise limited
• The INEE is a simple and direct measure,
  utilizing the variance in pixel gray values of a
  flat image over a range of exposures and
  requiring no special phantoms, and can be readily
  implemented in a clinical setting.

25
Future Work Incorporating Spatial Frequency

• Use NPS(u,v) Instead of Signal Variance
  

Spatial Frequency
A Kuhls-Gilcrist et al., Instrumentation Noise
Equivalent Exposure (INEE) An Investigation of
Spatial Frequency Effects, AAPM 2008, SU-DD-A4-6
26
Future Work Incorporating Spatial Frequency

• Relating to Frequency Independent Case

Varian Paxscan 2020
Quantum Limited Performance _at_ Exposures gt 3 µR
for all Spatial Frequencies (up to Nyquist)
A Kuhls-Gilcrist et al., Instrumentation Noise
Equivalent Exposure (INEE) An Investigation of
Spatial Frequency Effects, AAPM 2008, SU-DD-A4-6
27
Acknowledgements

• NIH Grant R01-NS43924
• NIH Grant R01-EB002873
• Toshiba Medical Systems Corporation