1

The issue of dosimetry and

uncertainties in the context of

internal emitters

Eric Blanchardon

EU scientific seminar « Issues with internal emitters »

Luxembourg, 23rd November 2010

INTERNAL DOSIMETRY

intake

distribution

in tissuesdeposition of

energyelimination

dose

i = m / M

biokinetic

model M

d = i x e

dosimetric

model e

in vivo / in vitro measurement m

i = c x b x t

environmental

measurement c

2

α

α risk of false positive

β risk of false negative

MEASUREMENT UNCERTAINTIES

M

DT

M > DT there is activity in the sample

M < DT activity is less than LD

β

sample

limit of detection LD

background

Poisson variability

MEASUREMENT UNCERTAINTIES

variability of Pu faecal excretion after inhalation of 239Pu for 24 h samples (Marsh et al, 2007)

relative uncertainty due to Poisson variability and chemical yield (Hurtgen and Cossonnet 2003)

in vitro measurement

3

Lung counting using 4 Ge detectorsCalibration using physical anthropometric phantoms

Livermore

Adequacy of calibration coefficients to real conditions?

=lungsA countsN

Restricted nature and distribution of activity

Rough representation

ε countingray tI ××

MEASUREMENT UNCERTAINTIES

in vivo counting

pixel

voxel

• Modelisation of the in vivo counting using numerical voxel phantoms

Simulated spectrum

• Measured worker• Internal contamination• Detectors

MEASUREMENT UNCERTAINTIES in vivo counting

• Simulation of particle transport using a Monte Carlo code

4

ICRP Publication 78 (1997)

IngestionInhalationExhalation

Urine

Blood

Otherorgans Kidney

Urinarybladder

Respiratory tract

LiverGastro

intestinaltract

Lymph nodes

Subcutaneoustissue

Faeces

Sweat

Direct absorption

Wound

Skin

Skin

BIOKINETIC MODELS

human respiratory tract modelICRP publication 66 (1994)

deposition, transport and absorption of particles into blood after inhalation

BIOKINETIC MODELS

5

Human alimentary tract modelICRP publication 100 (2006)

BIOKINETIC MODELS

Predicts fecal excretion, absorption into blood and retention in the alimentary tract.

BIOKINETIC MODELS

6

systemic model for plutonium

ICRPpublication 67 (1994)

BIOKINETIC MODELS

λλλλ(blood -> tissue) = outflow rate from circulation x fraction entering the tissueλλλλ(tissue 1 -> tissue(s) 2) = ln(2) / removal half-time

1.39 x 10-3Other kidney tissue to Blood

1.9 x 10-4Gonads to Blood1.386 x 10-2Kidneys (urinary path) to Bladder

2.11 x 10-4Liver 2 to Blood6.93 x 10-1ST0 to Blood

1.33 x 10-4Liver 1 to Small intestine1.29 x 10-2Blood to ST2

1.77 x 10-3Liver 1 to Liver 28.06 x 10-2Blood to ST1

7.6 x 10-3Cort/Trab marrow to Blood2.773 x 10-1Blood to ST0

8.21 x 10-5Cortical volume to Marrow7.1 x 10-5Blood to ovaries

4.93 x 10-4Trabecular volume to Marrow2.3 x 10-4Blood to testes

8.21 x 10-5Cortical surface to Marrow1.29 x 10-2Blood to ULI contents

4.11 x 10-5Cortical surface to Volume3.23 x 10-3Blood to Other kidney tissue

4.93 x 10-4Trabecular surface to Marrow6.47 x 10-3Blood to Kidney (urinary path)

2.47 x 10-4Trabecular surface to Volume1.29 x 10-2Blood to Urinary bladder content

1.9 x 10-5ST2 to Blood1.941 x 10-1Blood to Trabecular surface

4.75 x 10-4ST1 to Urinary bladder contents1.294 x 10-1Blood to Cortical surface

4.75 x 10-4ST1 to Blood1.941 x 10-1Blood to Liver 1

rate (d-1)Transferrate (d-1)Transfer

BIOKINETIC MODELS systemic model for plutonium

7

BIOKINETIC MODELS systemic model for plutonium

Leggett et al. 2005

BIOKINETIC MODELS systemic model for plutonium

Leggett et al. 2005

8

BIOKINETIC MODELS systemic model for plutonium

Leggett et al. 2005

UNCERTAINTY IN BIOKINETIC MODELS

Discussed in a series of authoritative articles

Leggett RW, Bouville A, Eckerman KF (1998) Reliability of the ICRP’ssystemic biokinetic models. Radiat Prot Dosim 79(1-4):335-342

Leggett RW (2001) Reliability of the ICRP's dose coefficients for members of the public. I. Sources of uncertainty in the biokinetic models. Radiat Prot Dosim 95(3): 199-213

Harrison JD, Leggett RW, Nosske D, Paquet F, Phipps AW, Taylor DM, Métivier H (2001) Reliability of the ICRP's dose coefficients for the members of the public. II. Uncertainties in the absorption of ingested radionuclides and the effect on dose estimates. RadiatProt Dosim 95: 295-308

Leggett RW (2003) Reliability of the ICRP's dose coefficients for members of the public. III. Plutonium as a case study of uncertainties in the systemic biokinetics of radionuclides RadiatProt Dosim 106: 103-120

9

UNCERTAINTY IN BIOKINETIC MODELS

Mostly dependant on the availability of relevant data

behaviour of the element in human subjects (H1)

behaviour of the element in other mammalian species (A1)

chemical analogue in human subjects (H2)

chemical analogue in other mammalian species (A2)

basic physiological data (P)

Quantification of reliability of the model

quantity of interest in [A,B] with roughly 90% probability

uncertainty factor UF = (B/A)½

reliability :

� high if UF < 2.2 (category I)

� moderate to high if 2.2 < UF < 3.3 (category II)

� low to moderate if 3.3 < UF < 8 (category III)

� low if UF > 8 (category IV)

UNCERTAINTY IN BIOKINETIC MODELS

Leggett et al. 1998

adults

children

10

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

f r

in vitro dog monkey mouse rat human

PuO2 MOX Pu

-Ma

rP

u-P

al

Pu

-Na

Pu

-CP

u-C

lP

u-M

gP

u-T

BP

Pu

-re

s

Pu nitrate Pu

cit

Type F

Type M

Type S

absorption of Pu : rapidly dissolved fraction fr

1.0 x 10-3Type S

11.2GSD

2.0 x 10-3MedianMOX

14.3GSD

1.1 x 10-3MedianPuO2

frCompound

UNCERTAINTY IN BIOKINETIC MODELS

Davesne et al. 2010

DOSIMETRIC MODELS

application of a Monte Carlo particle transport code to reference a anthropomorphic phantom

ICRP publication 107 (2009) decay data: energies and intensities of emissions

Photon energy (MeV)

Yie

ld p

er n

ucle

ar tr

ansf

orm

atio

n

U-235 Photon Line Spectrum

5E-50,0001 0,001 0,005 0,01 0,02 0,05 0,1 0,20,3 0,5 11E-7

1E-6

1E-5

0,0001

0,001

0,01

0,1

1

1020

PhotonsX-raysGamma rays

ICRP publication 110 (2010) : reference computational phantoms of the adult male and female

11

LOCAL GEOMETRY

spongious region of right scapulae with 90% cellularity (50 µm thick voxels)

red, trabecular bone

black, active haematopoietic marrow : targetfor leukaemia induction

yellow, inactive marrow

green, endosteum : target for bone cancerinduction

(50 µm layer instead of former 10 µm)

Target cells are also identified in the respiratory and alimentary tracts

Source and target regions in the skeleton

main sources (NCRP commentary No. 15,1998)

incomplete information on masses, compositions, shapes and locations of the organs and tissue of the human body

oversimplifications of the representations of certain complex anatomical structures in the body when calculating the energy deposition

limitations in the physical data (e.g. energy and intensity of radiations emitted by the radionuclides, photon interaction coefficients; etc.)

limitations in computational procedures for evaluating the energy deposition of penetrating radiations

UNCERTAINTY IN DOSIMETRIC MODELS

12

location of short-range (α, β, Auger) emitters in tissues

UNCERTAINTY IN DOSIMETRIC MODELS

thorium in hamster liver (Brooks et al. 1985)

uranium in rat tissues

0,0E+00

5,0E-05

1,0E-04

1,5E-04

2,0E-04

2,5E-04

3,0E-04

0 100 200 300 400 500

Max. depth of 20 µm annulus

Abs

orbe

d F

ract

ion

1 MeV1.5 MeV2 MeV0.5 MeV0.1 MeV

location of target cells for cancer induction

UNCERTAINTY IN DOSIMETRIC MODELS

13

APPLICATION OF THE MODELS

“the problem”

APPLICATION OF THE MODELS

time of contamination ?

1 Jan 1 Mar 1 May 1 Jul 1 Sep 1 Nov 31 Dec

M1 M2 M3 M4 M5 M6

individual variability : deviation from the reference biokinetic and dosimetric model

14

1.E-07

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1 10 100 1000

Temps (j)

Excretion fécale (Bq/Bq incorporé et par jour)

Type S, AMAD = 1 µm

Type S, AMAD = 10 µm

Type M, AMAD = 1 µm

Type M, AMAD = 10 µm

physico-chemical form (AMAD, absorption type) ?

APPLICATION OF THE MODELS

time (d)

Faecal excretion (Bqper incorporated Bq) faecal excretion

of 239Pu

For AMAD = 5 µm, dose coefficients:

Type M: 3.2 x 10-5 Sv.Bq-1

Type S: 8.3 x 10-6 Sv.Bq-1

Harmonisation:by following the procedures any two assessors should obtain the same estimate of dose from a given data set

Harmonisation:by following the procedures any two assessors should obtain the same estimate of dose from a given data set

General philosophy of the IDEAS guidelines (Doerfel et al. 2006)

Optimisation:the “best” estimate of dose should be obtained from the available data

Optimisation:the “best” estimate of dose should be obtained from the available data

Proportionality:the effort applied to the evaluation should be proportionate to the dose – the lower the dose, the simpler the process should be.

Proportionality:the effort applied to the evaluation should be proportionate to the dose – the lower the dose, the simpler the process should be.

APPLICATION OF THE MODELS

15

E = i x e(L)

p(L)p(i) p(t)

p(S | i, L, t)

p(M | S)

biokinetic modelL

intakei

time of intaket

doseE

sampled activityS

measured activityM

error model

prospective dosimetry =

direct propagation

retrospective dosimetry =

inverse problem

p(L|M)p(i|M) p(t|M)

p(E|M)

Techniques for solving inverse problemsClassical method (Molokanov et al, 2010)WeLMoS method (Puncher and Birchall, 2008)Bayesian network (Davesne et al, 2010)Markov chain Monte Carlo (Miller et al, 2002)

p(E | i, L)

PROPAGATION OF UNCERTAINTY

Birchall et al. 2010

Uncertainty on lung dose from Pu inhalation for epidemiological study of nuclear workers

PROPAGATION OF UNCERTAINTY

from urine bioassay measurement, taking account of biokinetic uncertainty in the human respiratory tract model, applying the WeLMoS method

16

0

2

4

6

8

10

T = 720

d, faeces

T = 360

d, faeces

T = 180

d, faeces

T = 180

d, urine

T = 180

d, faeces

fol urine

T = 180

d, urine

fol

faeces

T = 180

d, urine

+ faeces

T = 90 d,

faeces

T = 90 d,

urines

MDD (mSv)

1 mSv

Davesne et al. 2010PROPAGATION OF UNCERTAINTY

minimum effective dose detectable with 95% confidence by routine Pu monitoring at AREVA La Hague

DRAFT CONCLUSIONS

Internal dosimetry is complex but relies on sophisticated measurement techniques and dosimetric models which are upgraded with scientific progresses.

Model complexity warrants guidance in their application and reminder of their limitations and unavoidable associated uncertainties.

Quantification of uncertainty is important for epidemiological studies, retrospective assessment of individual risk, nuclear medicine and quality assurance of monitoring programs.

Robust mathematical methods have recently been applied to this issue. NCRP report 164 on uncertainties in internal radiation dose assessment was released this month. The harmonization of approach to uncertainty at the European level is a challenge for the years to come.

Further research is desirable to investigate the respective location of internal emitters and target regions for health effects in the human body ; and to link the outcome of dosimetry and microdosimetry with the observation of biological responses in the various situations of exposure.