You are on page 1of 66

Quality Control and the ACR’s

MRI Accreditation Program
Geoffrey D. Clarke, Ph.D.
University of Texas Health Science Center
at San Antonio
Radiological Sciences Division

Overview
• ACR MRI Accreditation Program
• MRI Quality Control Program
– Definitions
– Technologist’s Daily & Weekly QC Tasks
– Physicist’s Responsibilities
– Annual Tests for the Physicist
ACR MRI Accreditation Program
Features
• Evaluates effectiveness of quality control
measures
• Will collect findings to further the
development of quality control
information
• Qualified Medical Physicist should be
responsible for overseeing the
equipment quality control program
ACR MRI Accreditation Program
MRI Survey Agreement
• Official request for ACR Accreditation
• Site agrees to provide all
documentation, including but not limited
to quality control logs, films, records, or
any necessary information requested by
the survey team
• Agree to use the ACR MRI phantom
• Tests run in a short time
• Pulse sequences as compatible
as possible with all commercial
MRI scanners
• Specific measurements to
account for :
– Geometric Distortion
– Slice thickness & Position
– Factors Affecting Image SNR
(resolution, bandwidth, ghosting)
– Image Uniformity
ACR Phantom
• Affordable
($730)
• Results
easily
evaluated
Documents for Using Phantom
ACR MRI Accreditation Program
Quality Control Section
• Refers to tests put forth in the ACR
Standard for MRI
• All tests to be carried out in accordance
with written procedures and methods
• Preventative maintenance:
– Documented by qualified service engineer
– Repairs documented and records
maintained by the MR site
ACR
Magnetic
Resonance
Imaging
Quality
Control
Manual
2001
Radiologist’s Responsibilities
Eleven specific responsibilities including:
• To ensure that an effective quality control program
exists for all MRI
• To select the technologist to be the primary quality
control technologist
• To ensure that appropriate test equipment and
materials are available to perform the technologist's
QC tests.
• To arrange staffing and scheduling so that quality
control tests can be carried out.
• To select a qualified medical physicist or MRI
scientist To ensure that records are properly
maintained and updated in the MRI QC procedures
manual.
MRI QC Technologist’s
Responsibilities
• Weekly magnetic resonance image
quality control procedures (daily
recommended)
• Weekly quality control of hard copy and
soft copy Images
• Routine visual inspection of equipment
Weekly Technologist’s Tests
• Center frequency
• Table positioning
• Setup & Scanning
• Geometric accuracy
• High contrast resolution
• Low contrast detectability
• Artifact analysis
• Film quality control
• Visual Checklist
Technologist QC
• Daily (weekly)
tests
• Cheat sheets
• Review data after
ten days to
establish
baseline values
and variability
Measurement of Central
Frequency
• Can be performed by technologist
• Record center frequency value on
ACR phantom or manufacturer’s
phantom
• Automated analysis and recording
often available on modern MRI
systems
Central Frequency
•Specific but not sensitive
•Action Criteria:
Change in Hz from previous day
> 2 * resonant frequency in MHz
• Suggestive of:
Magnet drift
RF instability
Transmitter Gain/Attenuation
• Reflects power required to optimize
RF pulse:
–Depends on coil, phantom, pulse
sequence, etc.
–Should remain constant over time if
nothing in pulse sequence or hardware
has changed
Terminology & Units Vary with Vendor
• GE: displayed on screen (dB)
• Philips: under system performance
parameters
– rf_act_drivescale
• Siemens: “options” – “adjustments”
– Frequency
– Transmitter amplitude (temp) (V)
• Toshiba: “acquisition window”
– Center frequency (MHz)
– RF level
Geometric Accuracy
• Measure
distance
along main
axes of
phantom
• Compare
with known
values
Potential Causes of Geometric
Accuracy Failures
1. Phantom mispositioning
2. Gradient miscalibration
3. B
o
inhomogeneity
a. Ferromagnetic objects in magnet
b. Poor magnet shimming
4. Gradient non-linearity
5. Inappropriate receiver bandwidth
6. Poor eddy current compensation
7. Combination of two or more of above
Geometric Accuracy - Axial
Failure Due To: • Miscalibrated Gradients
• Low Receiver Bandwidth
• High B
o
Inhomogeneities
BW = 7.4 kHz
BW = 3.6 kHz
Slice #1
Gradient Correction
• If gradients
are
inherently
non-linear
gradient
correction
may be
applied
Percent Geometric Distortion
T1-Weighted, Central Slice
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
1-
Oct
31-
Oct
30-
Nov
30-
Dec
29-
Jan
28-
Feb
29-
Mar
28-
Apr
28-
May
27-
Jun
27-
Jul
26-
Aug
25-
Sep
25-
Oct
Date of Measurement
System #1 System #2 System #3
Poor Gradient
Calibration
Assessment of
MR Image Quality
• Performed by
Technologist
• Using ACR
Phantom
• High Contrast
Resolution
• Low Contrast
Detectability
RF Noise/Leaks/ Spikes
Single
frequency
artifact
shows up as
zipper in
middle of
image.
0.90 mm 1.0 mm 1.1 mm
Chem
Shift
Slice Thickness
Resolution Insert

ACR MRI Phantom
Resolution Insert
ZOOM = 2
Spatial Resolution Matrix:
Registration with Phantom
Image
Matrix
Resolution
Holes
PARTIAL VOLUME ARTIFACT
High Contrast Spatial
Resolution
• Evaluate conspicuity of holes arranged in
two square arrays
• Has to avoid partial-volume artifact…
UL
LR
High Contrast Resolution
•Specific but not sensitive
•Action Criteria:
Any reduction in # of holes seen
• Suggestive of:
Increased eddy currents
Poor gradient calibration
Poor B
o
uniformity
Reduced stability of system
Slice #10
Slice #12
Low Contrast Detectability
5.1% Contrast Detectability
3.7% Contrast Detectability
Low Contrast Detectability
1.5 T System: T1-Weighted Scan
0.00
5.00
10.00
15.00
20.00
25.00
1-May 31-May 30-Jun 30-Jul 29-Aug 28-Sep 28-Oct 27-Nov 27-Dec 26-Jan 25-Feb 27-Mar 26-Apr
Date
H
o
l
e

S
i
z
e

V
i
s
u
a
l
i
z
e
d

(
m
m
)
1.4% contrast
2.4% contrast
3.7% contrast
5.1% contrast
LCD and Signal-to-Noise
T
o
t
a
l

N
u
m
b
e
r

o
f

S
p
o
k
e
s

Signal-to-Noise Ratio
0 50 100 150 200 250 300 350

40

35

30

25

20

15

10

5

0
Low Contrast Detectability
•Sensitive but not specific
•Action Criteria:
•Slice used dependent on B
o

field strength
•Sustained 5 row decrease in
number of hole sets seen
• Suggestive of:
•Reduced stability of system
ACR MRI Standard
Hardcopy Image QC Tests
• Sensitometric Measurements for Film
Processors
• Hard Copy of SMPTE test pattern
• Similar to process used for
mammography program
• Laser camera film less sensitive to
temperature changes

Laser Film QC
Weekly:
View SMPTE pattern
Verify gray levels
• 0/5% & 95/100%
patches
Film 6 on 1
• 4 on 1 if necessary
Plot OD of
• 10%, 40% & 90&
patches
Observe film for
artifacts
SMPTE Test Pattern
Action Limits
SMPTE
patch
OD Control
Limit
0 2.45 ±0.15
10% 2.10 ±0.15
40% 1.15 ±0.15
90% 0.30 ±0.08
Documenting the
Quality Control Program
• Data form for daily equipment quality
control
– ACR MRI QC manual, pg. 64
• MRI Facility quality control visual
checklist
– ACR MRI QC manual, pg. 65
• Laser film printer control chart
– ACR MRI QC manual, pg. 66
If QC Test Fails
• Common errors-
– Check for magnetic objects in bore
– Re-seat head coil
– Reposition & landmark phantom
– Make sure scan room door securely closed
• Repeat daily QC scan procedures
• Record results again in QC notebook

Responsibilities of the
Qualified Medical Physicist
or MRI Scientist
• Write Purchase Specifications
• Perform Acceptance Testing
– Baseline Measurements
• Determine Action Limits
• Set up Daily/Weekly QC Tests
• MRI equipment performance review



Annual Survey Tests
• Magnetic Field Homogeneity
• Slice Position Accuracy
• Slice Thickness Accuracy
• Radio Frequency Coil Checks
– Volume coils
1. Signal-to-noise ratio
2. Percent integral uniformity
3. Percent signal ghosting
– Surface Coils SNR Tests
• Inter-Slice RF Interference
• Soft Copy Displays (monitors)
Magnetic Field Homogeneity
e
o

e
o
e
o

Fourier transform of
signal produces a
Lorentzian peak in
well-shimmed
magnet
Magnet firld
homogenity can be
characterized using
FWHM of resonance
peak
Denotes a totally
uniform magnetic
field.
All signal is at
resonant
frequency, e
o.

Ideal Homogeneity Good Homogeneity Poor Homogeneity
FWHM
FWHM
Magnetic Field Homogeneity
Phase images from
GRE sequences
with 13.9 ms
difference in TE’s
Phase and
Unwrapped
Phase Images
John Pauley, Stanford Univ.
Magnetic Field Homogeneity
• Overall, the phase mapping technique provides
the best mechanism for evaluating field
homogeneity.
• For some system service personnel may have to
provide use of phase-mapping acquisition and
analysis tools.
• Filmed copy of vendor’s final homogeneity map
and shim coefficients is useful for documentation
and establishing a baseline.
Slice Position Accuracy
• Uses Crossed-Wedges as Reference for
Positioning and Slice Spacing Accuracy
• MRAP pass criterion: magnitude of bar
length difference s 5 mm.
• The actual displacement is ½ of the
measured difference
• ACR Accreditation criteria are very weak,
physicist may want to hold manufacturer to a
higher standard
Slice Position Accuracy
SLICE #1
SLICE #11
Slice Position
Slice Spacing
Crossed wedges should be of equal length
if position and spacing are accurate
Slice Position Accuracy
• Causes of poor performance:
• Operator error
• Table positioning shift
• Miscalibrated gradients
• High B
o
inhomogeneities
MRI Slice Thickness
• Signal ramps have a slope of 10:1
• Signal from ramp is 10 x slice thickness
• Two ramps are used to compensate for in-
plane rotation of the phantom
• Phantom does not compensate for tilting
backwards or swaying left-right
Slice Thickness Measurement
1. Use slice 1 of ACR T1.
2. Magnify by 2 to 4. Adjust window/level
to see signal ramps. (Set window at
minimum.)
3. Use rectangular ROI to measure mean
of middle of each signal ramp. Take
average.
Slice Thickness
– Lower display to
one/half the
average ramp
signal.
– Measure lengths
of top and
bottom ramps
and calculate
slice thickness.

( )
( ) bottom top
bottom top
0.2 Thickness Slice
+
×
× =
Slice Thickness
• ACR-MRAP: slice thickness measured should
be + 0.7 mm of prescribed value
– + 14% error on 5mm slice, may be too generous
• Corrective actions:
– Check Axial Site Series Images
– Replace cables & connectors, look for other sources
of distorted RF pulse shape in RF electronics
– Try switching RF coils
– Check gradient calibration
MRI Equipment Performance Evaluation

Site: _____________________________ Date: ________
MRAP Number: ____________________ Serial Number: ___________


Equipment:
MRI System Manuafacturer: _________________ Model : ________
Processor Manufacturer : _________________ Model: _________
PACS Manufacturer: _________________ Model: _________
ACR MRAP Phantom Number used: _________

1. Magnetic Field Homogeneity

Method Used (check one): Spectral Peak ___ Phase Difference ___
Other (desc ribe) __________________________

Measured Homogeneity: Diameter of Spherical Homogeneity
Volume (cm) (ppm)

________ _______
________ _______
________ _______

2. Slice Position Accuracy

From Sl ice Positionss #1 and #11 of the ACR Phantom:

Wedge (mm)

Slice Location #1 ________
Slice Location #11 ________

3. Slice Thickness Accuracy

From Slice Position #1 of the ACR Phantom:

Slice Thickness Top ______ Calculated slice
(fwhm in mm) Thickness (mm) ______
Bottom ______

Duplicate these forms so they will be available for repeated use.
=+ = -


=+ =
-
Site &
Equipment
Data
B
o
Homogeneity
Slice Position
Accuracy
Slice Thickness
Accuracy
Bird-Cage Type Head Coil
RF coils produce uniformity patterns characteristic of
their design.
Uniformity Pattern
Birdcage Coil
High Field
Solenoid Coil
Low Field
ACR
Phantom
Slice #7
Volume RF Coil Measurements
Image Intensity Uniformity
• Performance criteria: PIU > 90%


• Measurement Considerations:
• Display may not show signal values
• Display may not allow user to set signal
display level
• There may not be a well-defined
high/low intensity level
percent integral uniformity =100× ÷
÷
+
|
\

|
.
| 1
( )
( )
high low
high low
Image Intensity Uniformity
• Causes of failure:
• Poor phantom centering in head coil
(usually AP)
• Ghosting
• Motion or vibration
• Mechanical failure in head coil
Phased- Array Coils
•The signal is viewed from more “angles”
•Using four channels does not produce \4 * S/N
Abdomen-pelvis
phased-array
Cervical-Cranial Phased Array
Torso
Phased Array
General Features of MRI
Phantoms
– Nonsignal-producing
container
– Proton density
similar to water
– Shorten T1: NiCl &
CuSO
4

– Mimic Conductivity
of tissues: NaCl
Agarose
Gel
Phantom
with Inserts
Uniform
Spherical
Phantoms
Phantoms Developed by Users
Surface RF Coil Measurements
Volume Coil Data
% Image Uniformity
Max Signal
Min Signal
Signal-to-Noise
Mean Signal
SD of Background Signal
Percent Signal
Ghosting
Ghost Signal
Mean Signal
Background Signal
Surface Coil Data
Maximum Signal-to-
Noise
Maximum signal
SD of Background Signal
-
Volume Coil
Data Recorded
Calculated Values:
Uniformity
SNR
Ghosting
Surface Coil

Signal-to-Noise
4. RF Coil Performance Evaluation
A. VOLUME RF COIL -
RF Coil Description: __________________________ Date: ____________
Phantom Description: ___________________________________________
Pulse Sequence: Type: ____ TR: _____ TE: ______ flip angle _____
degrees
FOV: _____ cm
2
Matrix: ___________ BW: _________kHz ; NSA ___
Slice thickness ______mm; spacing _______ mm
TX attenuation (or gain) __________
Data Collected:

Mean
Signal
Maximum
Signal
Minimum
Signal
Background
Signal
Noise
Standard
Deviation
Ghost
Signal


Calculated Values:
Signal-to-Noise
Ratio
Percent
Image Uniformity
Percent
Signal Ghosting


B. RF SURFACE COIL -
RF Coil Description: __________________________ Date: ____________
Phantom Description: ___________________________________________
Pulse Sequence: Type: ____ TR: _____ TE: ______ FOV: _____ cm
2
Matrix: ___________ BW: _________kHz ; NSA ___
Slice thickness ______mm; spacing _______ mm
TX attenuation (or gain) __________
Maximum
Signal
Noise Standard
Deviation
Maximum Signal-to-
Noise Ratio


Image uniformity distribution OK? ________
Image ghosting OK? ________
HARD COPY IMAGE: Window width ________ Window level _______
Several copies of this page may be required to report on all RF coils.
ACR MRI QC Manual, pg. 125
Slice Cross-Talk Measurements
1. Position 5mm slices on the uniform volume
2. Repeat measurements decreasing the slice
gap :
Series # 1 2 3 4
No. Slices 11 11 11 11
Slice Gap (mm) min 0.5 1.0 5
3. Measure the signal-to-noise ratio (SNR) for
each of the image sets.
4. Plot the SNR vs. percentage slice gap
Signal-to-Noise
vs. Inter-slice Gap
SE, TE = 20
FE, TE =8
TFE, TE =4
100%




90%




80%




70%
0% 25% 50% 75% 100%
Percentage Slice Gap
S
i
g
n
a
l
-
t
o
-
N
o
i
s
e

R
a
t
i
o


(
p
e
r
c
e
n
t

o
f

M
a
x
i
m
u
m
)

Soft Copy Displays
• Requires precision
luminance meter
• Four tests
– Maximum and minimum
luminance
– Luminance uniformity
– Resolution
– Spatial Accuracy
-
RF Slice
Interference
Soft Copy
Displays
Review of
Routine QC
Program
5. Interslice RF Interference
Phantom Description: ___________________________________________
Pulse Sequence: Type: ____ TR: _____ TE: ______ FOV: _____ cm
2
Matrix: ___________ BW: _________kHz ; NSA ___
Number of slices______









6. Soft Copy Displays
Monitor Description: __________________________________________
Maximum Luminance: ________________________ Cd m
-2
.
Minimum Luminance: _________________________ Cd m
-2
.
Luminance Uniformity:
Average of values obtained in four corners of screen: ______ Cd m
-2
.
Luminance measured in center of screen: ______ Cd m
-2
.
Percent difference: ________ %
|(Center – Average Corners)/(Center) x 100% < 30%|
7. Evaluation of Site’s Technologist QC Program
4) Set up and positioning accuracy: (daily) _________
5) Center Frequency: (daily) _________
6) Transmitter Attenuation or Gain: (daily) _________
7) Geometric Accuracy Measurements: (daily) _________
8) Spatial Resolution Measurements: (daily) _________
9) Low Contrast Detectability: (daily) _________
10) Film Quality Control (weekly) _________
Visual Checklist: (weekly) _________
Seri es Sli ce Si gnal-
Number Gap to- Noi se
( mm) Rati o
1
2
M
e
a
s
u
r
e
d

S
N
R
100 %
90 %
80 %
70 %
0% 25% 50% 75% 100%
I n te r - sl i ce Ga p (p e rce n t o f sl i ce th i ckn e ss)
3
4
Seri es Sli ce Si gnal-
Number Gap to- Noi se
( mm) Rati o
1
2
M
e
a
s
u
r
e
d

S
N
R
100 %
90 %
80 %
70 %
0% 25% 50% 75% 100%
I n te r - sl i ce Ga p (p e rce n t o f sl i ce th i ckn e ss)
3
4
ACR MRI QC Manual, pg. 127
MRI Equipment Evaluation Summary
Site ___________________ Report Date: __________
System MRAP #_____________ Survey Date: __________
MRI System Manufacturer ___________ Model: __________
Physicist/MRI Scientist: ____________________
Signature: ________________________________

Equipment Evaluation Tests
Pass / Fail
1. Magnetic Field Homogeneity: _________
2. Slice Position Accuracy _________
3. Slice Thickness Accuracy _________
4. RF Coils’ Performance
a. Volume Coils’ Signal-to-Noise Ratio _________
b. Volume Coils’ Image Uniformity _________
c. Volume Coils’ Ghosting Ratios _________
d. Surface Coils’ Signal-to-Noise Ratio _________
5. Inter-slice RF Interference _________
6. Soft copy displays _________
Medical Physicist’s or MRI Scientist’s Recommendations for
Quality Improvement:
______________________________________________
______________________________________________
______________________________________________
______________________________________________
______________________________________________
______________________________________________
______________________________________________
______________________________________________
______________________________________________
Summary
Sheet
(pg. 131)
Successful MRI QC Program
1. Technologists run QC scans on a
daily basis
2. If exceed action criteria – repeat QC
procedure
3. Physicist reviews QC data annually
4. Record data - report problems to
service
5. Have service record problems and
solutions in a service log