Quality Assurance In Diagnostic

Radiology

Robert G. Gould, Sc.D.

Professor and Vice Chair
Department of Radiology
University of California
San Francisco, California

Why Do Quality Control?

•  Improve clinical results

•  Preempt quality or safety problems
•  Maintain standard of care
•  Minimize patient radiation dose
•  Satisfy government regulations

1

 QC Testing
•  Acceptance testing
–  Upon installation prior to patient use
–  Medical physicist
•  Annual inspection
–  Medical physicist
–  Equipment vendor/service provider
•  Daily and weekly tests
–  QC technologist

Quality Control (QC)

•  Team approach
–  Radiologists, Medical Physicists,
Technologists
•  Use eyes and experience
•  Don’t “work around” problems
•  Try to be preemptive

2

 Mechanical Integrity

•  Fix problems as soon as possible

–  They only get worse
•  If things become loose, tighten them!

Regulations
•  Are more better?
•  Are all of equal value?
•  Do they cover all aspects of IQ and
safety?
–  Should I stop when all the regulatory tests
are complete?

3

 Quality Control
•  Emphasize those tests that are
important to IQ and/or safety
–  Concentrate on those functions that effect
quality and safety
–  Minimize time on activities done primarily
to meet regulations

Digital Projection Imaging: QC

X-ray tube!

Collimator!

X-rays!

Patient!

Grid!
Detector!

4

 X-Ray Tube Concerns
•  Focal spot size
–  Component in spatial resolution
•  Worn anode
–  Variation in intensity across field
–  Increase in HVL due to metal coating on
inside of glass
•  Instabilities, arching

Focal Spot Measurement

•  When?
–  Acceptance
–  Annually
–  Tube replacement
•  How?
–  Star pattern-measure spatial
resolution
–  Pinhole camera

5

 Spatial Resolution Measurement
•  Image a lead bar
test pattern
•  Assess using vendor
QC software to
determine contrast
of specific line pairs
–  MTF can be obtained
•  Determine along
both axis or at an
angle of 45°

Pinhole Camera

•  The best Focal spot, F

–  Shows emission distribution L1

–  Difficult and time
consuming Pinhole

–  Not possible for some tubes

•  Use CR and large L2

magnification factor (≥ 5x) Image, S

•  Careful alignment Detector

F = S (L1/L2)
= S [1/(M-1)]

6

 Generator QC
•  Consistent x-ray output for same
technical factors (KVp, mA, exposure
time)
•  mA and time settings
–  Should be linear
–  Should be consistent
–  mR/mAs should be a constant
•  KVp calibration

Collimators
•  Restrict primary X-ray beam to
detector size or a smaller
anatomic region-of-interest Tube housing

•  Restriction and alignment X-ray tube

of X-ray beam to detector Collimator

Added
•  Major component in radiation filtration

protection Light

Mirror

•  Reduces scatter

X-rays

•  Lower personnel exposure

Patient

•  Improved image quality

•  Reduction in patient radiation Grid

burden Detector

•  Component in beam filtration

7

 Collimator
•  What?
–  X-ray field - detector alignment
–  X-ray - light field alignment
•  When?
–  Acceptance
–  Annually
–  Tube replacement

X-ray - Light Field Alignment

•  Sum on opposite
sides < 2% of the
source-image
distance (SID) Mirror

Light

D

source

D

8

 Filtration

•  Minimum amount set by

regulation
•  Determined by measuring
the HVL

Intensity
Peak tube

–  Thickness required to Potential

reduce X-ray intensity to
half its initial value
–  Measured in mm of Al
–  Measure of X-ray beam 0 20
40 60 80 100 120

penetrance (hardness) Energy, KeV

Filtration/Beam Quality
Tube Minimum
•  Indicated by
measuring half Potential, HVL,
KVp Mm Al
value layer (HVL)
•  Need to measure at 50 1.5
only a single KVp 71 2.1
–  Tube potential
indication should be 80 2.3
calibrated 100 2.7
120 3.2

9

 Filtration
•  Current digital R/F and angio systems
have variable filtration
–  Combinations of Al and Cu of various
thicknesses
–  Anatomic protocols automatically change
•  Measure HVL at minimum filtration

Basic Imaging Geometry: Detector

Tube housing

•  Converts X-ray intensity to X-ray tube

electrical signal Collimator

Added
•  Major component of spatial filtration

resolution Light

•  Major determinate of patient Mirror

dose X-rays

•  Component of automatic Patient

exposure control system
Grid

Detector

10

 Digital Detectors: Radiography
•  What?
–  Uniformity
–  Artifacts
–  AEC
•  When?
–  Acceptance
–  Annually
–  Component replacement
–  Manufacturers recomendation

Digital Detectors: Flat Field Uniformity

•  Digital detectors do not

respond uniformly across field
–  Produces density variations
within the image
–  ‘Structured’ noise
•  Assessed by uniformity of pixel
values (eg. Mean and
standard deviation)
•  Most systems have software Mean,

that automates testing standard deviation

11

Digital Detector Structured Noise
Normalized “flat-field”
correction matrix

“Flat field”
algorithm

Flat-field “corrected” image

Raw image

•  Periodically generate new correction

matrix
–  Follow manufacturers recommendation
–  Often done by technologist

Digital Detectors: Artifacts

•  Non-uniformities
•  Dropouts and dead pixels
•  Determined by imaging uniform plastic
block
–  View with narrow W/L

12

 Automatic Exposure Control (AEC)

•  Should be able to maintain a pixel mean

value within ~15%
–  Track with changes in KVp
•  Clinically used range (~ 50 - 120 KVp)
–  Track with changes in patient thickness
•  5 - 35 cm of water equivalent

Annual Testing - Key Measurements

•  Mechanical integrity
•  Linearity of mAs
•  Half value layer
•  X-ray field - detector size
•  Light - x-ray field alignment
•  Spatial resolution
•  Artifacts/uniformity
•  AEC consistency

13

 Fluoroscopy QC

•  What?
–  Table-top exposure rate
–  Automatic brightness control
•  When?
–  Installation
–  Annually
–  Major component changes
–  Manufacturer’s recommendation

Typical Regulations
Fluoroscopic Equipment
•  Table-top exposure rate cannot exceed 10
R/min
•  During routine fluoroscopy the table-top
(patient entrance) exposure rate shall not
exceed 5 R/min for a typical patient
• Determined by use of a phantom equivalent to
8” of water

14

 Automatic ExposureControl (AEC)
•  Feedback mechanism that attempts to maintain a
constant brightness level from the center portion of the
output screen
–  center weighted exposure meter
–  adjusts the X-ray technique factors (mA and/or KV)
•  Determinate in patient dose

Fluctuating patient entrance dose! Constant input exposure!

Constant
light level!

X-ray tube!
+ KV! Patient! Image intensifier!
+ mA!
Feedback loop!

Automatic Exposure Control

•  For a given object size, should require
same kVp and mA
•  California:
–  8 inch plastic (lucite) phantom
–  12 inch table-top to entrance surface
distance
–  6.25 x 6.25 inch field at table-top
–  Record kVp and mA weekly

15

 Fluoroscopy - Image Quality
•  Image resolution pattern
–  Bar pattern (line pairs/ mm)
•  Contrast sensitivity
–  Low contrast phantom

Problem: very subjective

Computed Tomography QC
•  What?
–  Dose
–  Slice thickness/sensitivity profile
–  Table incrementation accuracy
–  Image quality factors
•  When?
–  Installation
–  Annually
–  Major component changes
–  Manufacturer’s recommendation

16

 Image Quality CT
•  Uniformity
•  Artifacts
•  Linearity
•  Noise
•  Spatial resolution
•  Contrast sensitivity AAPM phantom

CT Dose Measurements
•  CTDI
•  In air at isocenter (mR/mAs)
7.8  

7.8   4.3   7.8  

120  KVp  
100  mAs  
Cylindrical  PMMA   360°  rota<on   7.8  
64  slice  CT   Body  
 
32  cm  diameter

17

 QC Challenges
•  Man-machine interfaces
–  What goes on in the software black box?
–  How to test?

Cedars-Sinai CT Overexposures
•  What happened?
•  Brain perfusion procedures
–  Used in stroke assessment
•  Over-rode ‘default’ protocol
settings
–  Protocols come with the
machine
–  Changed technique factors that
effect dose
•  Eight times the protocol dose

18

 Cedars-Sinai CT Overexposure

•  Went on for 18 months because no one

made the association of hair loss and
skin reddening with CT procedure
–  2-3 weeks after exposure before onset of
hair loss

Cedars-Sinai CT Overexposure

•  Errors at multiple levels

–  Originally caused by changing default
protocol
–  Dose indicators appear at time of scan:
should have been recognized at time of
scan
–  Radiologist should have realized overdose
from the images
•  Not found during any QC testing

19

 Conclusions
•  QC is a necessary and valuable aspect of
x-ray imaging
•  QC should be a meaningful endeavor not
just going through the motions
–  React to problems before they interfere with
patient images
•  Not all QC tasks are of equal value
–  Concentrate on the important ones (those
that effect patient safety and image quality)

20