Standards for Quantitative PETBased Imaging at NIST
toward harmonization
Lisa R. Karam
Chief, Ionizing Radiation Division
National Institute of Standards and Technology
301-975-5561
lisa.karam@nist.gov
NIST Measurement Standards
Health Care
Radiation Physics
Theoretical dosimetry
Codes and models
Measurements for dosimetry
Brachytherapy
X-ray calibrations
Standards for mammography
Radionuclide metrology
Radionuclide standardization
Standard Reference Materials
Calibrations
Metrology for Diagnostics
•
•
•
•
•
Mammography
Conventional x rays
Spiral CT
Theoretical Dosimetry
Imaging radionuclides,
etc.
Instrument Calibrations Required by
Mammography Quality Standards Act
•
•
•
•
1992 Law requires proficiency testing of mammography
instruments
NIST and FDA developed necessary calibration facilities
4 manufacturers and FDA now meet the legal requirements
3 AAPM Accredited Dosimetry Calibration Labs
Conformal radiation therapy
CT and conventional plan
How We Got Here
key challenges and potential solutions
• Current imaging predominantly qualitative
(good for “yes/no” answers)
• Semi-quantitative results give semireliable answers
• Quantitative imaging
–
–
–
–
Treatment planning
Patient evaluation
Drug development
Clinical trials
Do you see it? What is it? Where is it?
Has it changed? What’s it up to?
Just how much radiation do I have to get???
Quantitative PET-CT
• Usefulness of PET-CT in trials depends on consistent
subject data over time and distance
• Persistent variability in results from PET-CT images
(in addition to subject variability)
– Between clinical sites
• Activity calibration (injected or in phantom)
• Conversion of image intensity to activity
• Protocols for acquisition, reconstruction, analysis
– Between scanners
• Conversion of image intensity to activity
• Different reconstruction algorithms
– Between scans
• Activity calibration (injected or in phantom)
• Conversion of image intensity to activity
Calibration traceable to national standards for
more quantitative results in patient assessment, drug
development, and treatment planning
Calibrating the Tools
•
Development of standardized, calibrated phantoms
–
–
–
–
Traceable calibration of scanners consistently over time
Means for QA/QC
Comparability of data from multiple sites during clinical trials
Comparability of reconstruction techniques
•
Close collaboration with Rensselaer Polytechnic Institute
to develop more anthropomorphic phantoms
•
Status
– Methods to convert CT image-based phantoms to
segmented/grouped voxels, organ specific polygon mesh,
and stereolithography (STL) formats for 3D visualization
– CT scans of NIST ref. BOMAB phantoms and LLNL thorax
– Anthropomorphic phantoms establish a “virtual calibration
phantom library” to quantify dosimetric difference with other
phantom geometries.
– Adding more CT scanned phantoms to the compilation.
Traceability Despite 2-hour Half-Life
“closing the loop”
•
Input from stakeholders from FDA to industry
(including drug developers) show needs for
– Traceability in activity measurement for instrument
calibration
– Longer-lived than conventional PET isotopes
– Phantoms to more closely mimic human body
– Protocols for use, training
•
Activity calibrators calibrated against
NIST 18F standard, linked to 68Ge
(geometry-specific calibration factors)
Ge-68 (t1/2 271 d) traditional for PET attenuation
– No National standard
– Calibration lacking for different geometries
(syringe, phantoms)
•
Injected activity
Status
– F-18 calibrations for the NIH PET Center and
PETNET Solutions (regional distributor)
– Calibration transferred to NIST 4πγ Secondary
Standard Ionization Chamber (future calibrations
to be more routine)
– New calibration 68Ge (on-going work with
RadQual, LLC for other geometries like syringe)
Scanners calibrated against NIST 68Ge
standard using calibrated phantom
Calculated activity
Standardized protocols,
analysis methodologies
Simplified Decay Schemes
68Ge-68Ga
and 18F
68Ge
100 % EC
(270.95 d)
1.2% β+
1.79% EC
68Ga
(1.130 h)
1077 keV
87.94% β+ (1899.1 keV)
8.70% EC
68Zn
(stable)
How will differences in Eβmax, photon
emission affect assay and image
quantitation?
96.86 % β+ (633.5 keV)
3.14 % EC
18F
(1.83 h)
18O
(stable)
Preparation of 68Ge Standard
Stock solution: dose vial
containing ~126 MBq GeCl4 in
5 mL of 0.5 mol L-1 HCl
Dilute*
Dilute*
GeE1A2
GeE1A1
Dilute
Counting in NIST
“4π”γ ionization
chamber, radionuclide
activity calibrators
Counting in NIST
“4π”γ ionization
chamber
GeE1A1D1
Dilute
*Carrier consisted of nominally 45
µg each Ge+4 and Ga+3 in 1 g
solution containing 0.5 mol L-1 HCl
LSC sources measured with
TDCR, 3H-standard efficiency
tracing, and 4πβ-γ
anticoincidence counting
GeE1A1D2
Point sources
measured by HPGe
Counting in NIST
“4π”γ ionization
chamber, Capintec
dose calibrators
68Ge
Epoxy-Based Sources
simulating 18F in a 5 mL syringe
Sources (liquid and epoxy
in mock-syringes)
prepared using 68Ge
standard solution
Measured in 4 activity
calibrators and
secondary standard
ionization chamber
Calibration factor derived
for liquid-filled syringe,
applied individually to
epoxy-filled syringes
*Conducted in collaboration
with RadQual, LLC
Determination of Response Factor
18F
relative to 68Ge
Total of 5 mock syringes
+ 1 NIST ampoule
gravimetrically prepared
from [18F]DG
Activity conc. determined
from ampoule (traceable
to National Standard)
Activity in each syringe
calculated; each source
measured in ionization
chambers.
Response of 18F at
different calibration factors
compared with liquid and
solid 68Ge sources.
Measured relative response = 1.054 ± 0.020;
Monte Carlo predicts 1.053
Measurement Error Uncovered for 18F
• Traceable 68Ge mock-syringes sent to 3
clinical sites
• Measurements made at manufacturerrecommended settings for 18F in 10 different
types of dose calibrator
• Comparison with NIST traceable activity value
for 18F (from relative response factor)
• Total of 31 independent measurements
• Results were all 5-6 % high
Calibrated 68Ge Phantom Standards
• Constructed using same technology as mock
syringes
• Calibrated against NIST liquid 68Ge standard
using HPGe spectrometry
• Standard uncertainty on activity ~ 0.7 %
Linking Activity to Common Standard
18F
Primary
Standard
68Ge
uc = 0.61 %
uc = 0.34 %
Primary
Standard
uc = 0.83 %
18F
18F
Syringe
Phantom
Response
factor=1.054
Response
factor?
68Ge
Syringe
Mock Standard
68Ge
Phantom
Goal at beginning: uncertainty in phantom calibration < 1%
uc = 0.52 %
uc = 0.7 %
Expanding Scope for the Future
•
Moving forward in activities in
– standards for characterization and performance (phantoms,
protocols)
– comparisons for change analysis (contrast agents,
radionuclides)
•
Expanding focus to additional areas in medical imaging (PET-MRI,
DXA) in which current expertise can leverage
•
Available resources (e.g., RIDER) for continued development and
implementation
•
On-going partnership with community to develop resources for
validation, harmonization, and optimization
•
Initiate efforts in medical physics beyond diagnostics
Measurement foundation to enable standardization among
centers:
Quantitative imaging traceable to National standards