7/21/2014
National Institute of Standards and Technology
(NIST) Quantitative Imaging Initiatives
Stephen Russek
Project Leader: Biomagnetic Imaging Standards, NIST, Boulder, CO
NIST Boulder
F2 time standard
NIST Gaithersburg
Josephson junction
voltage standard
AAPM MO-C-12A-6 July 21, 2014
Outline
1. Role of NIST
2. NIST Medical Imaging Standards: what's new
• Ionizing radiation standards (Radiation Physics Division)
 CT / PET phantoms
• Optical imaging standards
• Computational standards (Information Technology Lab)
 Virtual/ numerical phantoms
3. MRI standards/ phantoms (Electromagnetics Division)
•
•
•
NIST/ISMRM MRI system phantom
NIST/USCF breast phantom
NIST/RSNA QIBA isotropic diffusion phantom (Mike Boss
TU-C-12A-8 Tuesday 10:15AM)
NIST’s Role in Quantitative Medical Imaging
NIST is a National Metrology Institute: measurement & standards
We are good at:
•
•
•
•
•
•
Standard reference materials (SRMs)
Standard reference artifacts (phantoms)
Enabling traceability
Establishment of “ground truth”
Long term monitoring
Measurement development/ basic
metrology research
Not good at:
• Setting standards, most standards are
consensus
• Phantom mass production
• Moving very fast
Universities
NIH
NIST
Clinical Sites
Pharma
FDA
Professional
Societies
Venders
1
7/21/2014
Evolution of NIST medical imaging standards
Safety based standards
NIST radiation
dose standards

Imaging as quantitative
measurement of biomarkers
Mammography
Quality Standards Act
NIST Metrology
PET/CT & MRI
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Imaging as a Biomarker: Standards for
Change Measurements in Therapy
Workshop Summary
September 14-16, 2006,Laurence Clarke
Ram D. Sriram NISTIR 7434
Congressionally
funded NIST
imaging initiatives
Patient Protection and
Affordable Care Act
Metrology for Computed Tomography (CT)
•
•
•
Calorimetry-based dosimetry (MO-E-17A-12)
SRMs with calibrated attenuation coefficients
Simple dimensional phantoms
Foam lung
mimics
SRM 2087 Dimensional
Standard for Medical
Computed Tomography
SRM 2088 Density
Standard for Medical
Computed Tomography
Heather Chen-Mayer at the
PET/CT scanner with HDPE
phantoms.
Z.H. Levine
standards
standards
Brian Zimmerman
2
7/21/2014
Brian Zimmerman
PET Phantoms
68Ge
in epoxy cylinders
Brian Zimmerman
• Monitor scanner performance during clinical trials
• Comparison across scanners and clinical sites
• Accuracy of reconstruction and
scatter/attenuation corrections
Standard uncertainty on activity ~ 1 % Compatible with Jaczszak or ACR IQ phantoms
Optical Medical Imaging Program at NIST
Jeeseong Hwang, David Allen, Toni Litorja (PML) Antonio Possolo (ITL)
Emerging Application Areas:
Calibrated Hyperspectral imaging:
( surgery, combat and
diabetic wounds)
“Wet” and digital
phantoms
[HbO2]/[Hb]
Tissue Oximetric
Imaging
imaging with a palette of 100s of contiguous spectral
bands. Each pixel has a full spectrum and can reveal
chemical information about a region.
Optical coherence tomography:
Near-IR 3d
imaging technique that collects scattered light that
reveals sub-surface features, 1mm resolution
Fast
tunable white
light laser source *instrument is currently in demo mode
NIST Workshop on “Standards for the Advancement of Optical Medical
Imaging,” August 26-27, 2014 NIST Gaithersburg, Maryland
3
7/21/2014
Virtual/ numerical phantom for modeling
clinical tumors
Adele Peskin, Alden Dima, Charles Fenimore, James Filliben,
Joseph Chen, Richard Rivello (Information Technology Laboratory)


Realistic CT lung tumor data (virtual phantom) with
known tumor volumes based on clinical tumors
Embed synthetic tumors in DICOM data sets from
NCI RIDER at two time points to determine accuracy
of volume change measurements
Tumor +
blood vessels
NV lung tissue
Clinical tumor
vascular/
partial volume
Synthetic tumor
Pixel value (Hounsfield units)
Peskin Lecture Notes in Computer Science
(LNCS) series pp. 736 - 746 2010
NIST/ISMRM MRI System Phantom
First MRI phantom
with NIST traceability,
temperature and field
corrections, stability
monitoring
Measures:
• Geometric distortion, B1 uniformity, B0 uniformity, T1, T2,
Proton density, resolution, slice thickness, SNR
Purpose: Scanner QC and inter-scanner comparison, verify T1
T2 mapping protocols, off-the-shelf validation for some clinical trials
MRI Phantoms: must cover large parameter space
1000
PVP
10%
T2-T1 plot @ 1.5T for:
NiCl2 & MnCl2 array @ 20C
T1 Array
25%
and selection of tissues @ 37 C
T2 Array
40%
Skin
T2 (ms)
Fiducial Array
CSF
Blood
NiCl2 array
PD Array
Fibroglandular mimic
MnCl2 array
Gray matter
100
Fat
Fat mimic
Olive oil
Heavy mineral oil
White Matter
Connective
Glial Matter
Optic nerve
Spinal cord
White matter
Kidney
Kidney
Cartilage 55°
Muscle
Liver
Skeletal muscle
Liver
Heart
Cartilage 0°
contrast enhanced
10
100
1000
• Phantoms contain
materials with well
defined parameters!
• Many other dimensions
required to mimic tissue:
diffusion, conductivity,
susceptibility!
T1(ms)
4
7/21/2014
T1-Inversion Recovery gold standard:
Need to understand variability
0.65
NiCl2 @ 1.5 T
20
0.64
Error = 100*(T1measured-T1)/T1
r1 (1/s)
0.63
0.62
15
5%
0.61
10
0.60
0.58
0.57
17 18 19 20 21 22 23 24 25 26 27 28
Temperature (C)
• Arrays have large ranges with
short and long T1s, T2s that can
be challenging to measure
• Need in-situ thermometry for
accurate phantom measurement
comparisons.
Error (%)
0.59
NIST 1.5 T 20.12 C 11TI
MDACC 1.5T-11TI
MGH 1.5T-12TI
Cin 1.5T-6TI
MGH 3TB6-11TI
MGH 3TB4-10TI
CUINC3T-9TI
UCD BIC 3T-10TI
5
0
-5
-10
-15
-20
10
272.9 4.1 ms
CoV=1.5%
100
1000
Target T1 relaxation time (ms)
Katy Keenan
T1 Variable Flip Angle: Large variations in
practical mapping sequences
NIST/ISMRM
system phantom
T1 array
Standard protocol:
7 flip angles
Katy Keenan
NIST/UCSF Universal Breast Phantom
Katy Keenan NIST, Nola Hilton UCSF
normal
benign
malignant
• For ACRIN 6698/ISPY 2 DWI Biomarkers
for Assessment of Breast Cancer
Response to Neoadjuvant Treatment
• T1, diffusion, geometric distortion, and
tissue mimics
5
7/21/2014
T2 of olive oil used in
breast phantom
CPMG (NMR)

SE (NMR/MRI)

FSE (MRI)


signal
Do precise measurements of NMR parameters
make sense in tissue?
TE (s)
Many peaks each
with its own T1, T2
MRI measures apparent T1,
apparent T2 and apparent
diffusion coefficient!
Need pragmatic but
rigorous definitions?
Mike Boss
Breast Phantom: T2 ground truth?
Spin echo includes chemical exchange r2 Dw2  B02
1.5 T
3.0 T
MRI MultiEcho Spin
Echo
NMR
CPMG*
t = 1.0ms
NMR MultiEcho Spin
Echo*
TE=15ms
2.26 mM NiCl2
& 0.25 mM
MnCl2 in water
55 ms
57.8 ms
39.0 ms
38.8 ms
35% w/w Corn
Syrup in water
261 ms
266 ms
47.3 ms
84.7 ms
40 ms
160 ms
32.5 ms
171.8 ms
Material
Grapeseed Oil
NMR
CPMG*
MRI measurements at 16.5 deg C. NMR measurements at ~20 deg C.
* Integrated over all peaks
Mike Boss/ Katy Keenan
SI traceability for MRI?
PET dose calibrator
•Traceability in MRI not established
(exception dimensional traceability through
optical interferometry)
directly traceability to NIST
•composition traceability for Ni and Mn
concentrations NIST Nickel SRM 3136 and
Manganese SRM 3132 Standard Solutions using
inductively coupled plasma optical emission
spectroscopy (ICPOES).
• NIST can offer traceable measurements of T1,
T2, ADC, susceptibility … using calibrated
variable field, variable temperature NMR ,
magnetometry if we can agree on suitable
definitions!
System phantom reference
libraries
phantom reference libraries
6
7/21/2014
NIST Perspective/ Goals
• NIST is ramping up biomedical imaging standards for
quantitative biomarkers
• Goal: extend precise traceable measurements inside the
human
• Assist developing/ validation MR phantoms: anisotropic
diffusion, active flow/perfusion
• NIST will help facilitate a roadmap for standards for
quantitative MR
• NIST will investigate a study of economic impact of
standards-based quantitative MR
Workshop on Standards for Quantitative MR
NIST Boulder July, 2014
NIST MRI standards team: Mike Boss, Katy Keenan, Karl Stupic
7