ASSESSMENT OF OCCUPATIONAL EXPOSURE DUE TO INTAKES OF RADIONUCLIDES

1
ASSESSMENT OF OCCUPATIONAL EXPOSURE DUE TO
INTAKES OF RADIONUCLIDES
Uncertainties and Performance Criteria
2
Interpretation of Measurement Results Unit
Objectives

• The objective of this unit to identify and define
  the criteria that are used to characterize the
  quality of the measurement process for both
  direct and indirect methods. It will also
  identify sources of uncertainty in measurement
  and interpretation and give an estimate of
  expected magnitudes.
• At the completion of this unit, the student
  should understand how to calculate Minimum
  Detectable Activity and establish adequate
  accuracy criteria for measurement bias and
  precision.

3
Interpretation of Measurement Results - Unit
Outline

• Measurement Uncertainties
• Intake and Dose Assessment Uncertainties
• Performance Criteria Accuracy
• Performance Criteria Sensitivity
• MDAs - Examples

4
Measurement Uncertainties
5
Dose determination uncertainties
Measurement
Interpretation
Direct or indirect measurements
?1
Body/organ content, M or Excretion rate, R
6
Measurement uncertainties

• Usually most straightforward to estimate
• Counting statistics dominate at low activities
• For radionuclides that are,
• Easily detected, and
• In sufficient quantity,
• counting statistics are small compared to other
  uncertainties
• Systematic uncertainties are important
• Correction for activity remaining previously
  measured intakes may be necessary

7
Common measurement uncertainties

• Statistical counting errors
• Distribution in the body
• Absorption by overlying tissue (low energy
  photons)
• External contamination of the subject or
  measurement system
• Calibration errors
• Source activity
• Simulation accuracy

8
Estimated Direct Measurement uncertainties
Source of uncertainty Estimated magnitude 1 s
Chest wall thickness determination 15 to 300 worst case for 17 keV
Geometry errors Subject size and shape departure from single-size calibration model 10 for good geometries (I m arc, linear w/front /back counts) 15-20 for common geometries (linear w/counts from 1 side, 50 cm arc) 40 for poor geometries (detector in contact w/body)
Positioning of subject 10-15 for whole body
From ANSI 13.30 (1996)
9
Typical uncertainties for assessing fission
product isotopes
Source of Uncertainty Estimated Uncertainty
Depth Length ? Width Height-Weight Analysis Technique Calibration Counting Statistics Total Estimated Uncertainty 12 5 7 3 5 7 40
From Toohey, et al,
10
Typical uncertainties for U lung counting
Source of Uncertainty Estimated Uncertainty
Chest Depth Chest Wall Thickness Activity Location Detector Placement Subject Background Calibration Counting Statistics Total Estimated Uncertainty 12 15 5 5 10 5 40 90
From Toohey, et al,
11
Typical uncertainties for Pu lung counting
Source of error Uncertainty
Subject background 50
Counting statistics 50
Chest wall thickness 40
Non-uniform distribution 70
Calibration 20
Overall uncertainty 110
From Toohey, et al,
12
Estimated Indirect Measurement uncertainties

• Several parameters contribute to indirect
  measurement uncertainties
• The uncertainty associated with most are highly
  variable
• Typical uncertainties associated with the
  radiochemistry are of the order of 3
• More details can be found in the USDOE Laboratory
  Accreditation Program report ANSI N 13.30 and
  ISO 12790-1

13
INTRODUCTION OF SF

• The recently developed IDEAS Guidelines for the
  assessment of internal doses from monitoring data
  suggest default measurement uncertainties (i.e.
  scattering factors, SF) to be used for different
  types of monitoring data.
• The SF values represent the geometric standard
  deviation of the distribution of all results,
  supposed to be approximated by log-normal
  distribution.

14
INTRODUCTION OF SF

• The IDEAS guidelines consider two types of
  uncertainty
• Type A connected to counting statistic and
  decreasing with the increasing of activity and
  counting time (Poisson distribution)
• Type B all other components of uncertainty also
  connected with inter and intra-subject
  variability (e.g. in excretion)

15
INTRODUCTION OF SF

• SF values are important. For these issues.
• They are needed to assess the uncertainty in the
  estimated intake and dose.
• They determine the relative weighting of data in
  fitting process and can effect the estimated
  intake when different types of monitoring data
  are used simultaneously.
• They enable rogue data to be identified
  objectively
• They enable objective (statistical) criteria
  (goodness-of-fit) to be calculated, which are
  used to determine whether the predictions of the
  biokinetic model (with a given set of parameter
  values) used to assess the intake and dose are
  inconsistent with the measurement data.

16
INTRODUCTION OF SF

• The IDEAS Guidelines assume the overall
  uncertainty on an individual monitoring value can
  be described in terms of a lognormal distribution
  and the SF is defined as the geometric standard
  deviation (GSD).
• This approximation is valid if Type A errors are
  relatively small (lt30). Thus, it is assumed that
  if the measurements could be repeated,
  hypothetically at the same time, then the
  distribution of the measurement results could be
  described by a lognormal distribution.

17
INTRODUCTION OF SF

• SF values depend on type of monitoring
  measurement. Default values are reported in the
  following slides.
• When the type A component of the uncertainty is
  small (lt 30) the type B component alone could be
  used for uncertainty.

18
SF default values for in-vivo measurements

• SF values depend on type of monitoring
  measurement. For in-vivo measurement types

In vivo measurements SF values (Type B uncertainty)
Low photon energy E lt 20 keV 2.1
Intermediate photon energy 20 keV lt E lt 100 keV 1.3
High photon energy E gt 100 keV 1.2
19
SF default values for in-vivo measurements

• SF values depend on type of monitoring
  measurement. For in-vitro measurement types

In vitro measurements SF values (Type B uncertainty)
URINE For HTO after inhalation 1.1
URINE Normalized 24 h excretion 1.7
URINE Spot urine data 2.0
FECES Inhalation (Pu-Am) 2.5
FECES Wound (Pu) 3.1
20
Intake and Dose Assessment Uncertainties
21
Some sources of assessment uncertainty

• Mode of intake
• Physical and chemical form of material
• Particle size (AMAD) of the aerosol
• Time pattern of intake (acute vs. chronic)
• Errors in biokinetic and dosimetric models
• Individual variability in biokinetic and
  dosimetric parameters

22
Intake assessment uncertainties

• Difficult to quantify in routine monitoring -
  measurements are made at pre-determined times are
  unrelated to time of intakes
• Compromise between measurement interpretation
  quality and the practical limitations linked to
  measurement frequency
• Monitoring intervals should be selected so that
  underestimates due to unknown time of intake are
  3

23
Intake assessment uncertainties

• Practically, this is a maximum since the actual
  distribution of the exposure in time is unknown
• Statistically, the error is not systematically
  the same for all the assessments
• The random distribution of the exposure makes
  such an error clearly lower than a factor of 3
• If intake occurs just before sampling or
  measurement, it could be overestimated 3

24
Intake assessment uncertainties

• Particularly important for excreta monitoring ?
  daily fractions excreted can change rapidly
  immediately after intake
• If a high result is found in routine monitoring,
  it would be appropriate to repeat the sampling or
  measurement a few days later ? adjust the
  estimate of intake accordingly
• Samples could also be collected after a period of
  non-exposure, e.g. after weekend or holiday

25
Assessment uncertainties

• Models used to describe radionuclide behavior are
  used to assess intake and dose
• Reliability of dose estimates depends on the
  accuracy of the models, and limitations on their
  application
• This will depend upon many factors, including
• Knowledge of the time of intake, and
• Whether the intake was acute or chronic

26
Assessment uncertainties

• If the sampling period does not enable the
  estimation of the biological half-life,
  assumption of a long body retention may lead to
  an underestimate of the intake and the committed
  effective dose
• The degree of over- or under-estimation of the
  dose depends on the body retention pattern

27
Assessment uncertainties

• Radionuclide behavior in the body depends upon
  their physicochemical characteristics
• Particle size of inhaled radionuclides is a
  particularly important for influencing deposition
  in the respiratory system
• Gut absorption factor f1 substantially influences
  effective dose following ingestion

28
Assessment uncertainties

• When exposures during routine monitoring are well
  within limits on intake, default parameters may
  be sufficient to assess intake
• If exposures approach or exceed these limits,
  more specific information on
• Physical form and chemical form of the intake,
  and
• Characteristics of the individual,
• may be needed to improve the accuracy of the
  model predictions

29
Intake fraction, m(t) depends on several factors
60Co, inhalation type M
m(t)
Time after intake, d
Intake pattern (acut. vs. chr.) Deposition
site Time after intake Particle size
Absorption rate (F, M or S) Mode of intake
30
Performance CriteriaAccuracy
31
Performance criteria

• Accuracy
• Bias (Systematic errors)
• How well can a given measurement be reproduced.
• Repeatability or Precision (Random errors)
• How close is the mean of a series of
  measurements to the true value
• Sensitivity (MDA)
• What is the lowest value of a quantity that can
  be measured?

32
Performance criteria - Bias
Definition
where Bri relative bias for the ith
measurement Ai measured activity Aai
actual activity for the ith measurement
33
Performance criteria - Bias

• For a test or measurement category,
• Where Br Relative bias for the category
• n number of replicate measurements

34
Performance criteria Repeatability

• Definition
• where SBr measurement repeatability for the
  test or measurement category

Also termed Precision
35
Accuracy - How close is close enough?

• When the activity Aai is at or above the
  specified Minimum Testing Level (MTL),
• Relative bias, Br
• - 0.25 ? Br ? 0.50
• Relative repeatability, SBr
• SBr ? 0.40
• These values used by ISO and USDOE Laboratory
  Accreditation Program

36
MTL Values for Direct Measurements
Measurement Category Type Radionuclide MTL
I. Transuranium elements via L x-rays Lung 238Pu 9 kBq
II. Americium-241 Lung 241Am 0.1 kBq
III. Thorium 234 Lung 234Th in equilibrium w/ parent 238U 0.5 kBq
IV. Uranium-235 Lung 235U 30 kBq
V. Fission and activation products Lung Any two 54Mn, 58Co, 60Co, 144Ce 134Cs 137Cs/137Ba 3 kBq 30 kBq 3 kBq
VI. Fission and activation products Total body All of 134Cs, 137Cs/137mBa, 60Co 54Mn 3 kBq
VII. Radionuclides in the thyroid Thyroid 131I or 125I 3 kBq
37
MTL Values for Indirect Measurements
Measurement category Radionuclide MTL (per L or per sample)
I. BETA activity average energy lt 100keV 3H, 14C 35S 228Ra 2 kBq 20 kBq 0.9 kBq
II. BETA activity average energy 100 keV 32P 89, 90Sr or 90Sr 4 Bq
III. ALPHA activity isotopic analysis 228,/230Th or 232Th 234/235U or 238U 237Np 238Pu or 239/240Pu 241Am 0.02 Bq 0.02 Bq 0.01 Bq 0.01 Bq 0.01 Bq
IV. Elements (mass/volume) Uranium 20 µg
V. GAMMA (photon) activity 137Cs/137mBa 60Co 125I 2 Bq 2 Bq 0.4 kBq
38
Accuracy - How close is close enough?

• ICRP Publication 75, General Principles for the
  Radiation Protection of Workers
• For external dosimetry a factor of 1.5 at the
  limits (20 mSv/year)
• The overall uncertainty in the dose from internal
  exposure, is likely to be greater than for
  external exposure

39
Accuracy - How close is close enough?

• ICRP Publication 75, General Principles for the
  Radiation Protection of Workers
• Sampling frequencies should be chosen to avoid
  errors due to intake uncertainties of more than
  about a factor 3
• For less simple programs, e.g. for insoluble
  plutonium, total uncertainties may be about one
  order of magnitude.

40
Performance Criteria Sensitivity
41
Two terms describe sensitivity

• Minimum Detectable Activity (MDA) (a priori)
• Minimum activity that can be detected
• Probability, a, of false positive (Type I error)
• Probability, ß, of false negative (Type II error)
• Decision level, LC, (a posteriori)
• The total count value or final measurement of a
  statistical quantity, LC, at or above which the
  decision is made that the result is positive
• Probability, a, of false positive (Type I error)

42
Confidence levels and k values
a 1-ß k
0.001 0.999 3.090
0.005 0.995 2.576
0.010 0.990 2.326
0.025 0.975 1.960
0.050 0.950 1.645
0.100 0.900 1.282
0.200 0.800 0.842
0.250 0.750 0.675
0.300 0.700 0.525
0.400 0.600 0.254
0.500 0.500 0
43
Standard Deviation
where s standard deviation of a set of N
measurements xi ith measurement in
the set ?x mean of the set of
measurements Estimate of the relative standard
deviation for a single measurement sB
standard deviation of the appropriate blank
sample s0 standard deviation of the net subject
or sample count
44
Illustration of LC and MDA relationship
0
Lc
MDA
0 Value of background distribution LC The
likelihood that the sample distribution
characterized by LC was not really positive
(false positive) is a MDA The likelihood that a
sample distribution characterized by the MDA will
be missed (false negative) is ß and is not really
positive (false positive) is a
sB
a
(b)
(a)
kasB
?
s0
(c)
kßs0
Background
Detected
Not detected
May be
Will be
45
Minimum detectable activity - MDA

• Values assigned to MDA depends on the risk of
  making an error, false positive or false
  negative.
• Simplification Assume ß a, and ß a 0.05
• Then ka k1-ß 1.645 k
• where s0 standard deviation of net subject
  counts
• K efficiency
• T subject counting time

46
Minimum detectable activity - MDA
where sB1 standard deviation in subject
counts with no actual activity sB0
standard deviation in unadjusted blank
counts It can be assumed that sB1 sB0 s0,
and m 1 Then, s0 sB?2 1.415sB, where sB is
the standard deviation of a total blank count
47
Minimum detectable activity - MDA
For direct measurements, MDA becomes For
indirect measurements where R chemical
recovery ? radiological decay constant ?t
elapsed time between reference time and
time of count
48
Direct measurement MDAs
Measurement category Organ MDA
I. Transuranium elements via x-rays Lungs 185 Bq/A
II. 241Am Lungs 26 Bq
III. 234Th Lungs 110 Bq
IV. 235U Lungs 7.4 Bq
V. Fission and activation products Lungs 740 Bq/A
VI. Fission and activation products Whole body 740 Bq/A
VII. Radionuclides in the thyroid Thyroid 740 Bq/A
From ANSI 13.30 A is the number of photons
per nuclear transformation L x-rays

for transuranium elements, and gamma rays for
fission and activation products
49
Indirect measurement MDCs (urine)
Measurement Category Nuclide MDC
I Beta - Average energy 100 keV 3H, 14C, 35S 147Pm 210Pb, 228Ra, 241Pu 370 Bq/L 0.37 Bq/L 0.19 Bq/L
II. Beta Average energy gt 100 KeV 32P, 89/90Sr or 90Sr 131I 0.74 Bq/L 3.7 Bq/L
III. Alpha Isotopic specific measurements 210Po, 226Ra, 228/230/232Th, 234/235/238U 237Np, 238/239/240Pu, 241Am, 242/244Cm 3.7 mBq/L 2.2 mBq/L
IV. Mass determination Uranium (natural) 5 µg/L
V. Gamma or x-rays Emitters with photons 100 keV 2 Bq L-1/A
VI. Gamma or x-rays Emitters with photons gt 100 keV 2 Bq L-1/A
From ANSI 13.30 A is the number of photons per
nuclear transformation L x-rays for
transuranium elements, and gamma rays for fission
and activation products
50
Indirect measurement MDAs (faeces)
Measurement Category Nuclide MDA
VII. Alpha Isotope specific measurements 234/235/238U, 228/230/232Th, 238/239/240Pu, 241Am 37 mBq/sample
VIII. Beta Average energy gt 100 keV 89/90Sr or 90Sr 0.74 Bq/sample
IX. Gamma or x-rays Emitters with photons 100 keV 2/A Bq/sample
X. Gamma or x-rays Emitters with photons gt 100 keV 2/A Bq/sample
Minimum detectable concentration - From ANSI
13.30 A is the number of photons per nuclear
transformation L x-rays for transuranium
elements, and gamma rays for fission and
activation products
51
MDAs Examples
52
Determination of MDA - Example

• 90Sr by Beta Gas Flow Proportional Counting
• 20 reagent blanks were counted for 1 hour each
  3600 s

Total counts Total counts Total counts Total counts Total counts
83 69 53 72 59
77 70 62 88 53
66 73 59 55 74
72 70 65 68 61
53
Determination of MDA - Example

• 90Sr by Beta Gas Flow Proportional Counting

?B 67.4 counts SB ? ?(Xi 67.4)2/19
9.4 Counting efficiency, K 0.36 Chemical
yield 0.81
54
Determination of MDA - Example

• Whole body counting for fission and activation
  products

Radionuclide 137Cs 60Co
Organ Body Lungs
Counts in peak region - B 9 8
SB ?B 3 2.8
Count time, T s 600 600
Calibration factor, K 1.35?10-4 2.97?10-4
MDA - Bq 209 90
55
References

• HEALTH PHYSICS SOCIETY, Performance Criteria for
  Radiobioassy, An American National Standard, HPS
  N13.30-1996 (1996).
• INTERNATIONAL ATOMIC ENERGY AGENCY, Occupational
  Radiation Protection, Safety Guide No. RS-G-1.1,
  ISBN 92-0-102299-9 (1999).
• INTERNATIONAL ATOMIC ENERGY AGENCY, Assessment of
  Occupational Exposure Due to Intakes of
  Radionuclides, Safety Guide No. RS-G-1.2, ISBN
  92-0-101999-8 (1999).
• INTERNATIONAL ATOMIC ENERGY AGENCY, Indirect
  Methods for Assessing Intakes of Radionuclides
  Causing Occupational Exposure, Safety Guide,
  Safety Reports Series No. 18, ISBN 92-0-100600-4
  (2002).
• International Standards Organization, Radiation
  Protection Performance Criteria for
  Radiobioassay Part 1 General Principles, ISO
  TC 85/SC2 (1999).