How large is your dosimetry error

1
How large is your dosimetry error? How certain is
your risk estimate?
Louis K. Wagner, Ph.D. Department of Radiology
2
Objectives

• Identify accuracies/precisions necessary for
  diagnostic and interventional imaging
• Identify sources of inaccuracies/uncertainties
• Assess relationship of dosimetry to risk in
• Mammography
• Computed tomgraphy
• Fluoroscopy
• Fetal dosimetry

3
Exposure meters
4
Exposure as a standard measurement
In 1908 Paul Ulrich Villard proposed that the
ionization density generated in air under normal
conditions of temperature and pressure (STP) be
used as a standard for the measurement of
radiation exposure.
5
Exposure as a standard measurement
The International Standard (SI) units of exposure
is Coulombs of charge of one sign generated per
kilogram of air at STP.
X dQ/(?airdV)
6
Units of Exposure
In the United States 1 R ? 2.58 x 10-4 C/kg of
air (1 mC/kg 3.876 R)
Most other countries No special unit, just mC/kg
7
Air kerma
Kerma kinetic energy released in
medium (matter) ? absorbed dose in
air (for diagnostic energies)
Buildup reaches equilibrium in a few centimeters
of air and bremsstrahlung and attenuation
losses in air are negligible.
8
Units of Air Kerma
The SI units of air kerma are Gy, mGy, etc. 1R
0.258 mC/kg 8.76 mGya 1 mGya 0.0295
mC/kg 114 mR
To distinguish air kerma from absorbed dose in
tissue, I use the following notation mGya for
air kerma mGyt for tissue dose
9
Goals of air kerma measurements
AAPM Task Group No. 6 Med Phys 19 (1) Jan/Feb
1992 231 - 241.
10
Goals of air kerma measurements
Within 99 confidence (3?), measurements must be
known to within 10 of the true value (i.e., the
value measured by a national standard).
This assures sufficient measurement accuracy for
risk assessment and for radiographic equipment
evaluation in diagnostic and interventional
radiology.
The 2? uncertainty of most calibrations of meters
is warranted at /- 5, placing the 3?
uncertainty at 7.5. This doesnt leave a lot of
room for others errors! Actual uncertainty of
calibration is probably better.
AAPM Task Group No. 6 Med Phys 19 (1) Jan/Feb
1992 231 - 241.
11
Risk

• Relative or comparative risk
• Usually uses a standard patient surrogate
• Mammo (ACR phantom)
• CT (16- or 32-cm diameter lucite cylinder)
• Fluoro (23-cm water)
• Real risk
• Pregnancy
• Fluoro Interventional Fluoro Procedures
• CT (Pediatric, fluoro, multiple)
• Stochastic

12
From MacKenzie I. Breast cancer following
multiple fluoroscopies. Br J Cancer 1965 19 1-8
13
Relationship of air kerma readout to absorbed
dose in tissue
D (mGyt) B ? f ? FTP ? Fcal ? FBQ ?
Kreadout(mGya)
f 1.06 mGyt /mGya
HVL in mm aluminum Backscatter factor
(Field size 30-cm x 30-cm) 2.5
1.34 4.0 1.41
FTP 0.98 - 1.02 FTP(70F, 770 mm)
0.984 FTP(74F, 750 mm) 1.018 Changes
significant with altitude, e.g., for Albuquerque
20 WARNING Airport pressures corrected for
altitude
Fcal ? FBQ 0.96 - 1.04
14
Temperature and pressure correction

• Some monitors can correct for temperature and/or
  pressure, others have no correction
• Accuracy and reproducibility of automatic
  correction should be verified
• Corrections are significant with altitude.
• WARNING Airport pressures are corrected for
  altitude

15
Calibration and beam quality
For acceptance testing of chambers
Mammography 1) 0.25 mm Al 2) 0.35 mm Al 3) 0.5
mm Al
General diagnostic 1) 1.5 mm Al 2) 3.5 mm Al 3)
5.0 mm Al 4) 10 mm Al (my addition)
Computed Tomography 1) 3.5 mm Al 2) 10 mm Al
AAPM Task Group No. 6 Med Phys 19 (1) Jan/Feb
1992 231 - 241.
16
Acceptable energy performance
For acceptance testing of chambers
Mammography, General diagnostic , Computed
Tomography lt 5 change in correction factor
Survey lt 30 change in correction factor
AAPM Task Group No. 6 Med Phys 19 (1) Jan/Feb
1992 231 - 241.
17
(No Transcript)
18
Acceptable energy performance
For acceptance testing of chambers
19
Acceptable energy performance
For acceptance testing of chambers
20
Energy dependence of f-factor
21
Back scatter factor
22
Other sources of measurement uncertainty

• Linearity with rate and magnitude (test by
  intercomparison)
• Stem effect
• Leakage
• Selection mode (e.g., Exp rate Exp Pulsed Exp
  etc.)
• Battery
• Condition of equipment

23
Other sources of measurement uncertainty
Test by intercomparing meter responses with
escalating tube current
24
Malfunction errors
Meters can malfunction without warning.
1.00 mGy
0.70 mGy
25
Blunders

• After repair of fluoro unit physicist found it
  exceeded max allowable rate
• Service engineer found output well under maximum
  rate
• Both dosemeter systems manufactured by the same
  company had sensors for chamber used
• Unfortunately, engineers dosemeter misidentified
  3-ml ionization chamber as a 60-ml chamber, which
  produced a reading that was roughly a factor of
  twenty too low.
• Every system that that engineer calibrated would
  have had outputs a factor of 20 higher than what
  he measured.
• The manufacturers literature for the dosemeter
  states that his series of chambers could be
  compatible with his electrometer, but a special
  adapter is required and not all chambers in that
  series have provisions for automatic recognition.
  The user must verify that the displayed chamber
  designation matches the detector that is in use.

26
Blunder errors
The display reading must be multiplied by 2 when
used with the 1015C.
27
Mammography dose
28
Mammography dose
Dg DgN XESE
DgN F (Target material, filtration material,
kVp, HVL, breast thickness and composition)
XESE F (setup and exposure meter measurement)
kVp depends on accuracy of the kVp meter HVL
depends on aluminum purity, accuracy of thickness
guage, energy dependence of exposure meter
other things
29
Important factors for accurate measurement of
dose in mammography
Exposure meter energy dependence a) affects
measurement of XESE b) affects measurement of HVL
Aluminum purity and thickness measurement affects
measurement of HVL
kVp meter accuracy affects measurement of kVp
Exposure meter calibration /- 7.5
Machine settings (e.g., AEC mode and
compression), and equipment (e.g., filter, grid,
paddle) affect actual x-ray exposure
30
kVp Dependence of DgN (mrad/R) Compression 4.2
cm HVL 0.34 mm Al
kVp DgN (50,50)
Conclusion A 2 kVp error can result in a 1.5
error in dose.
Data from Wu, Barnes, Tucker Radiology179
143-148, 1991
31
HVL Dependence of DgN (mrad/R) Compression 4.2
cm kVp 26
HVL (mm Al) DgN (50,50)
Conclusion An error of 10 in HVL can result in
an 8 error in dose.
Data from Wu, Barnes, Tucker Radiology179
143-148, 1991
32
(No Transcript)
33
(No Transcript)
34
Mammography HVL, error in measured filter
thickness
Results Systematic error of up to 4 when
thickness measured by some micrometers.
Conclusion Measure thickness by mass per unit
area using physical density of 2.699 milligrams
per cubic millimeter.
35
Mammography HVL, error due to impurities in plates
Results 1) 1100 aluminum alloy yields variations
of ?4 in measured HVL 2) in pure aluminum
measured HVL is 3 - 10 larger than with 1100
alloy 3) variation in HVL measured in pure
aluminum is ?0.7 . Conclusion Measure HVL with
aluminum plates of at least 99.9 purity.
36
Important factors for accurate measurement of
dose in mammography
Exposure meter energy dependence a) affects
measurement of XESE b) affects measurement of HVL
Aluminum purity and thickness measurement affects
measurement of HVL
kVp meter accuracy affects measurement of kVp
Exposure meter calibration /- 7.5
Machine settings (e.g., AEC mode and
compression), and equipment (e.g., filter, grid,
paddle) affect actual x-ray exposure
37
Kerma-Length-Product
KLP ? -? ? K(z) dz
38
Kerma-Length-Product
39
(No Transcript)
40
(No Transcript)
41
Kerma-Length Product (KLP)

• Properties of KLP meters
• Partial volume detector shaped like long right
  circular cylinder
• Intended for use inside a phantom
• Readout depends on dose, slice width and scatter
• Measurement may be dependent on length of meter

• Utility of KLP meters
• Quality control device for CT
• May be used to estimate dose to patient

42
Kerma-Length-Product
Average air kerma , Ks, spanning length of slice
interval for a series of slices is
(approximately) Ks KLP/S Kavg L/S where
Kavg is corrected air kerma measurement L is
active length of chamber S is slice
increment (Assumes L gtgt beam profile)
43

Computed Tomgraphy Dose Index
CTDI f ? Fcal ? FTP ? FBQ ? (1/nT)? -7T7T
K(z) dz
CTDI f ? Fcal ? FTP ? FBQ ? (1/nT) ? KLP
Using a Kerma length meter with L 100 mm, CTDI
is best approximated for T 7mm where T is the
nominal slice width for contiguous slices and n
is the number of slices per scan.
Note f 0.90 mGypmm/mGya for polymethyl-methacry
late and 1.06 for tissue.
44

Computed Tomgraphy Dose Index
CTDI100mm f ? Fcal ? FTP ? FBQ ? (1/nT) ?
-5050 K(z) dz
CTDI100mm f ? Fcal ? FTP ? FBQ ? (1/nT) ?
KLP(100mm)
where T is the nominal slice width for contiguous
slices and n is the number of slices per scan.
Note f 0.90 mGypmm /mGya for
polymethyl-methacrylate and 1.06 for tissue.
45
Computed tomography dose
CTDI100mm 16-cm PMMA phantom 120 kV 220 mA 1
s 10-mm slice
Dose at center PMMA dose 40.0 mGy Water dose
47.1 mGy Difference is solely due to f-factor
for PMMA which is 15 too low
46
Computed tomography dose
47
CT head dose - bolus versus no bolus for
otherwise identical scan conditions
Dose (100-mm) at center (120 kVp, 270 mA, 1 sec,
10-mm slice thickness)
48
Computed tomography dose
The effects of bolus on dose measurement
100-mm dose at center (120 kVp, 270 mA, 1 sec,
10-mm slice thickness) Water phantom dose (with
teflon ring but no bolus) 17.4 mGy Water
phantom dose (with teflon ring and with bolus)
20.4 mGy Difference of 17 is solely due to
scatter from bolus
49
Computed tomography dose
CT scan of 18-cm water-equivalent head
phantom (CT -5)
CT scan of 16-cm PMMA head phantom (CT 115)
50
Computed tomography dose
CT scan of 18-cm water-equivalent head phantom
with teflon ring ( -5)
CT scan of human head 20.3 cm x 14.9 cm Fluid
4 W/GM 35- 40
51
Computed tomography dose
CTDI100mm 120 kV 200 mA 1 s 10-mm image
slice dose at center
52
Kerma-Length-Product
L Active length of the pencil chamber (100
mm) S slice incrementation interval (for
contiguous slices, S is the same as the nominal
slice width)
53
Kerma-Length-Product
L Active length of the pencil chamber (100
mm) S slice incrementation interval (for
contiguous slices, S is the same as the nominal
slice width)
54
Kerma-Length-Product
55
Dose (mGyt) f(mGyt/mGya) ? Fcal ? 8.76 (mGya/R)
? 100 mm/s(mm) ? Xcorr(R) 1.06 mGyt/mGya ?
1.86 ? 8.76 (mGya/R)? 100 mm/s(mm) ? Xcorr (R)
1727 mGyt ? mm/R ? Xcorr (R)/ s(mm) Conditions
120 kVp 220 mA 1-sec rotation speed.
56
Dose (mGyt) f(mGyt/mGya) ? Fcal ? 8.76 (mGya/R)
? 100 mm/s(mm) ? Xcorr(R) 1.06 mGyt/mGya ?
1.86 ? 8.76 (mGya/R)? 100 mm/s(mm) ? Xcorr (R)
1727 mGyt ? mm/R ? Xcorr (R)/ s(mm) Conditions
120 kVp 220 mA 1-sec rotation speed.
57
Computed tomography dose
Head phantoms
Body phantom
58
Computed tomography dose
CT scan of 33-cm solid-water body phantom
CT scan of 33-wk prenancy
59
Fluoroscopic dose
FDA Public Health Advisory Avoidance of
Serious X-Ray-Induced Skin Injuries to
Patients During
Fluoroscopically-Guided Procedures September 30,
1994 To Healthcare Administrators
Risk Managers
Radiology Department Directors Cardiology
Department Directors The Food and Drug
Administration (FDA) Center for Devices and
Radiological Health (CDRH) has received reports
of occasional, but at times severe,
radiation-induced skin injuries to patients
resulting from prolonged, fluoroscopically-guided,
invasive procedures...
60
Fluoroscopic entrance air kerma measurements
Patient simulated attenuator
mm Cu attenuator
Ref Lin P-JP. Periodic testing of equipment. In
Balter S, Shope TB. Syllabus A categorical
Course in Physics. Physical and Technical Aspects
of Angiography and Interventional Radiology.
Radiological Society of North America, Inc., Oak
Brook, IL, 1995, pp-241.
kV mA
61
Fluoroscopic dose
From Anderson JA, Wang J, Clarke GD. Choice of
phantom material and test protocols to determine
radiation exposure rates for fluoroscopy.
RadioGraphics 2000 20 1033 - 1042.
Conclusion PMMA and water are linear at a ratio
of 0.91 and mostly independent of beam quality.
62
Fluoroscopic dose
From Anderson JA, Wang J, Clarke GD. Choice of
phantom material and test protocols to determine
radiation exposure rates for fluoroscopy.
RadioGraphics 2000 20 1033 - 1042.
Conclusion The correspondence of copper sheets
to water thickness is strongly dependent on beam
quality.
63
Fluoroscopic dose
Graph applies to 20 cm of water.
From Anderson JA, Wang J, Clarke GD. Choice of
phantom material and test protocols to determine
radiation exposure rates for fluoroscopy.
RadioGraphics 2000 20 1033 - 1042.
Conclusion The correspondence of PMMA and Cu to
water thickness is a simple matter of differences
in penetration spectra.
64
Air Kerma-Area-Product Meters(Dose-Area-Product
or DAP meters)

• Utility of DAP meters
• Quality control device for
• Heavy feet
• Collimation
• Patient monitoring for stochastic risk
• Portal monitor for dose if collimation area is
  factored out (SSD known)

• Properties of DAP meters
• Partial volume detector
• Usually located on or in housing assembly
• Readout depends on dose and collimation
• Measurement mostly independent of SSD
• Primarily related to stochastic effects
• Table attenuation a?

65
Relationship of KAP (DAP) to SSD
KAP Readout Area Kerma Areas Aream (ds
/dm)2 Kermas Kermam(dm/ds)2 Ergo
KAPs ? KAPm
Level of skin
ds
Level of DAP meter
dm
66
Fluoroscopic dose
Skin dose f BSF FBQ Fcal DAP/ area of field
at skin
Factors to consider 1) variation of entrance
skin site through beam angulation 2) variation of
field size through electronic magnification or
collimation 3) variation of SSD
Note DAP meters with which I am familiar use
sealed chambers and require no TP correction.
67
Fluoroscopic dose - verification of DAP
68
Fluoroscopic dose
69

• Characteristics of the monitor
• Stem effect
• Cable coiling
• Directivity
• Reproducibility (with disconnect)
• Energy response
• Linearity with kerma
• Linearity with kerma rate

70
Stem Effect
71
Set up for measuring entrance exposure rate (EER)
under various fluoroscopy conditions
72
Scintillation detector as it appears in
fluoroscopic image
73
Scintillator is completely removed in all
subtracted images
74
Monitor mounted on c-arm and wrapped in lead for
shielding from scatter
75
Demonstration of standard setup with table at
maximum height and II down to surface of patient
Detector at port
Monitor mounted on c-arm and wrapped in lead to
shield from scatter
76
(No Transcript)
77
Test using 25-cm plastic phantom

• Conclusion
• At 50-cm SSD dose increased by 40 over 70-cm
  SSD
• If 500 rad at 50 cm, only 360 rad at 70 cm
• Difference 140 rad!

78
Sample Skin Doses for TIPS Procedures
79
ICRP
During the first 10 days following the onset of
a menstrual period, there can be no risk to any
conceptus, since no conception will have
occurred. The risk to a child who had previously
been irradiated in utero during the remainder of
a 4-week period following the onset of
menstruation is likely to be so small that there
need be no special limitation on exposures
required within these 4 weeks.
80
Potential Risks at 2 - 8 weeks PC

• Some forms of malformation
• Small Head Size
• Neoplasm

A threshold in excess of 50 mGy is likely for
this effect.
81
Potential Risks at 8 -15 weeks PC

• Mental retardation
• Small head size
• Neoplasm
• Hypothyroidism (131-Iodine)

82
Potential Risks at gt15 weeks PC

• Neoplasm
• Hypothyroidism (131-Iodine)

83
Important Attitude

• The inaccuracy of the estimate is often more
  important than the estimate!

84
Sources of Inaccuracy and Uncertainty

• Fluoroscopy kVp mA-min (recall, ABC, etc.)
• Fraction of fluoro mA-min over conceptus
• Conceptus depth (ante- vs retroverted bladder
  full or empty upright, supine, prone obliquity)
• Patient composition
• mAs for AEC
• Discarded films (retakes)
• CT - patient size, beam-on time for reference

85
My Goals in Dose Estimate

• Provide the physician with my best estimate of
  the dose
• Provide the physician with the uncertainty of my
  estimate with the maximum likely dose to the
  conceptus
• Provide the physician with documented evidence of
  the potential risk to the conceptus

86
Methods for X-ray Dose Estimates

• Normalized depth dose
• Percentage depth dose
• Tissue-air-ratios
• Patient simulated normalized depth-dose

Since this technique assumes a fixed location of
the conceptus in the pelvis, deviations of this
depth from actual introduces error.
87
Percent Depth-Dose Technique

• Dose
• A(corrected for TP, BQ, calibration) x f x B x
  P/100 x (inverse square correction)

88
Conceptus depths in anteverted uteruses
Patient AP thickness 19 cm 26 cm 20 cm
Central depth 9 cm 13 cm 10 cm
Conceptus depth 6 cm (avg of 16 early
pregnancies) 6 cm 4 cm (partially filled
bladder) 6 cm (full bladder)
Refs Ragozzino et al. 1981, 1984 Wagner et al.
1983.
89
Dependence of Dose to Conceptus on Depth (70 kVp)
90
Out-of-field Dose to Conceptus
91
CT Dose to Conceptus
92
CT Dose to Conceptus
93
CT Dose Estimate Techniques

• Free-in-air reference dose model
• CTDI reference dose model

Dose Dref x F(0) (Ti / 10mm) x Si F(zi)
94
From MacKenzie I. Breast cancer following
multiple fluoroscopies. Br J Cancer 1965 19 1-8