personal online dosimetry using computational methods · suppose we can use monte-carlo simulations...

Copyright © 2017 SCK•CEN

BVS-ABR Young Scientist Event: Digital tools to support ALARA

11th of October 2019, Hasselt

1

Mahmoud Abdelrahman, Pasquale Lombardo, Olivier Van Hoey, Filip Vanhavere

Personal Online DosImetry Using CoMputational Methods


Framework for individual monitoring: why is dosimetry needed

2

Individual monitoring of workers

Control occupational exposure

Dose limits and ALARA principle

Inform workers of their exposure

Dose Annual Limit

Effective Dose 20 mSv

Eye Lens 150 mSv

Extremities 500 mSv/year


Problems with individual dosimetry

Workers don’t like to wear dosimeter

Workers especially don’t like to wear more than one dosimeter

Still not all parts of body covered

What if other parts of body need dosimetry in future (brain, heart,…)?

Not always strict use of dosimeters:

Forgetting

Not correct place

Not taking into account personal differences

3


Uncertainties in personal dosimetry

4

Risk is given by effective dose

Complicated system of operational quantities to estimate effective dose

Hp(10) is only estimation of E

No dosimeter is perfect for Hp(10)

Non-linearity, fading, …

Energy and angular dependence….

Loosing dosemeter: all data lost…

Not wearing correctly

Dependent on homogeneity of the fieldFactor 1.5 in either direction for doses near the limit

Factor 2.0 for lower doses (ICRP 75)


Personal Dosimetry: Interventional Radiology

5


Personal Dosimetry: Interventional Radiology

6


Personal Dosimetry: Inhomogeneous fields

7


Personal Dosimetry: what brings the future?

8

Development in direction of small Active dosimetry systems with wireless data transfer

No more need to send dosimeters monthly to read out service

May be no need for physical dosimeters?

Suppose we can use Monte-Carlo simulations to calculate on-line all doses

Advantages:

No more need for physical dosimeter

No more loosing dosimeters

No more need for operational quantities

No more worries for changing quantities/weighting factors

Doses to all organs can be known

Personalized dosimetry possible

Better accuracy possible

Faster feedback to workers

….

mSv

Copyright © 2017 SCK•CEN9

Dose Simulations

Motion tracking

Computer Vision

Machine Learning

Computational Phantoms

Monte-Carlo Methods

2000’s

Hybrid

1980’s

Voxel

1960’s

Stylized

1950’s

Simple

Computational Power

Position/Pose

Radiation sourceGeometries


PODIUM Objectives

10

Improve occupational dosimetry via an online dosimetry application using computer

simulations: without the use of physical dosemeters

Develop an online application in which we will calculate individual occupational doses

In a limited time frame, simultaneously use an intermediate approach with pre-

calculated fluence to dose conversion coefficients for phantoms of different statures

and postures

Apply and validate the methodology for two situations where improvements in

dosimetry are urgently needed: neutron workplaces and interventional radiology

The legal aspects to introduce this or similar techniques as an official dosimetry method

will also be established


PODIUM Workplan

11

WP1 WP2

WP0

Management

WP3

Dosimetry Online Calculation Application

Dose simulations input Computational phantoms and MC

Validation in Interventional Radiology

WP4 WP5Validation in Neutron workplaces

WP6

Results

Dissemination


PODIUM Application

12

𝐃𝐨𝐬𝐞

𝐅𝐥𝐮𝐞𝐧𝐜𝐞

MCNP

Effective Dose

Organ Doses

MC-GPUmSv


Dose

Calculation

Motion

Tracking

Input

Radiation

Source

Input

Geometry

Input

13

WP1: Dose Simulations Input

Staff movement monitoring and Radiation field mapping


Depth Measurement Techniques

Microwaves Light Waves

Triangulation

(Two Camera

views)

Time-of-Flight

(Kinect V2)

Structured light

(Kinect V1)

Laser Scan

(Linear Scan)

Ultrasonic Waves

WP1: Motion Tracking

Human Motion Tracking

Non-visual

Tracking

Inertial based

Magnetic based

Glove based

Other sensors

Visual Tracking

Visual Marker

based

Visual

Marker-less

based

Combinatorial

tracking

Robot-Aided

Tracking

Depth cameras

RGB-D

14


Staff Motion Tracking

Markerless tracking based on computer vision

RGB-D cameras: combine color information with per-pixel depth information

Existed for years for high prices (~ $10k to $30k), very cheap nowadays…

Two technologies: Structrued light & Time-of-Flight (ToF)

Microsoft® Kinect V2 .0

15


Storing XYZ coordinates or send to a cloud

Real-time processing

Depth Image

Staff Motion Tracking

Skeleton Tracking

Tracking system based on single depth camera


Tracking to computational phantom

17

Realistic Anthropomorphic Flexible Phantom (RAF)


Animation of RAF phantom

18


Geometry Input

Define of the workplace geometry for the calculations

Complex geometries can be prepared by:

Scanning of the workplace

Converting CAD files to different formats

Modeling and tracking of important moving objects (shielding)

is also needed

19


Interventional Radiology and Cardiology Parameters

Parameter Range

High Voltage 60-120 kVp

Intensity 5-1000 mA

Inherent filtration 3-6 mm Aleq

Additional filtration 0.2-0.9 mm Cu

Energy range of scattered spectra 20 keV – 100 keV

X-Ray spectrum

• Tube potential (kVp value)

• Tube current

• Added filtration

• Target material

• Voltage waveform

Tube Angulation

• C-arm projections

20

Input• Radiation dose structured report (RDSR)

extracted from the X-ray machine

• Time synchronization with tracking

• DAP meter for normalization

Radiation Source Input: Radiology Case


WP2: Dose Simulations

Provide fast dose calculations for workers moving in realistic workplace fields

Based on Monte Carlo methods

Using variety of computational body phantoms

Two different approaches:

Library of pre-calculated conversion coefficients

Real on-line simulation

21


WP2: Dose Simulations

22

Export to mesh

format or

voxelization

Protective

garments

Easy

Posture

Control


MCNP Framework

23


MCNP Framework

24

~ 1 min. per irradiation event per single core

Intel Xeon E3-1270 v5 3.60 GHz


Validation

25


Controlled Experiment - Malmö

26

DC16

EDDDC20

EPD1 EPD2

EPD1

EPD2

4 cm

14.5

cm

24.5

cm DC20 DC16

EDD

Operator

Experiment 1 Experiment 2 Experiment 3


Controlled Experiment – IR Malmö

27

ExperimentTube

Angle

DICOM

KAP

[mGy.cm2]

Measured Hp(10) [µSv]

EPD-1 EPD-2 TLD (Ref.) TLD (Ref.) EDD

1 0° 2804.9 11 16 9 (DC16) 7 (DC20) 0

2 0° 4102.6 73 72 85 (DC48) 134 (DC52) 196

3 15.1° 4209.5 73 63 105 (DC36) 156.3

Ratio: Measured / Simulated

Experiment EPD1 EPD2 DC16 DC20

1 1.21 1.13 1.42 1.60

2 1.45 1.29 0.97 1.18

3 0.99 1.01 0.79

Results:

EPD1

EPD2

4 cm

14.5

cm

24.5

cm DC20 DC16

EDD

Operator phantom


Test at UZ-VUB - Brussels

28

Kinect 2

View from Kinect 1


Test at UZ-VUB - Brussels


Test at CHU-Liège

30


Test at CHU-Liège

31


Validation Case: Angioplasty Procedure

32

Measurement of accumulated dose Hp(10) of operators with Thermo EPD Mk2.3

Estimation of dosimeter location by the tracking system


Procedure Dose Report

Procedure parameters:

• kVp:

- 80 – 125 kVp

• DAP:- 15935.6 μGy.m2

• Projections:

- 0LAO / 0CRA

- 75RAO / 0CRA

- 27RAO / 0CRA

33


Clinical Validation: Results

34

Validation Case

Simulations

Accumulated

Hp(10)

Measured EPD

Accumulated

Hp(10)

Diff.

Clinical Experiments

CHU-Liège Case 4 38 μSv 23 μSv ~1.65

SJH Case B 7.7 μSv 5 μSv ~1.54

SJH Case C 68.3 μSv 55 μSv ~1.24


Conclusion

Part of the future will be dosimetry without physical

dosemeters

Although dosimeters still will exist for many applications

Very exciting topic: novel approach

New technologies can do amazing things

Results show the validity of the method in interventional

radiology and some neutron workplaces

Still some challenges to be solved

Shielding tracking, worker identification

35

www.podium-concerth2020.eu


PODIUM Team

36

PODIUM is part of the CONCERT project. This project has received funding from the

Euratom research and training programme 2014-2018 under grant agreement No 662287.


Thank you for your attention!

Join PODIUM Workshop in Athens

At the19th EAN WORKSHOP“INNOVATIVE ALARA TOOLS”

26th – 29th November 2019Jointly organized with the European ALARA Network (EAN)

Copyright © 2017

SCK•CEN

Copyright © 2014 - SCKCEN

PLEASE NOTE!

This presentation contains data, information and formats for dedicated use ONLY and may not be copied,

distributed or cited without the explicit permission of the SCK•CEN. If this has been obtained, please reference it

as a “personal communication. By courtesy of SCK•CEN”.

SCK•CEN

Studiecentrum voor Kernenergie

Centre d'Etude de l'Energie Nucléaire

Belgian Nuclear Research Centre

Stichting van Openbaar Nut

Fondation d'Utilité Publique

Foundation of Public Utility

Registered Office: Avenue Herrmann-Debrouxlaan 40 – BE-1160 BRUSSELS

Operational Office: Boeretang 200 – BE-2400 MOL


13:19:37


13:19:38


13:19:39


13:19:40


13:19:41


13:19:42


Skeleton tracking of the first operator


(1) Roguin et al., Brain tumours among interventional cardiologists: a cause for alarm? Eurointervention. 1086-1081:7;2

(2) Vano E, et al. Radiation-associated lens opacities in catheterization personnel: results of a survey and direct assessments. Journal of Vascular Interventional Radiology

2013;2: 197-204.

(3) Dehmer G, et al. Occupational Hazards for Interventional Cardiologists, The Society for Cardiovascular Angiography and Interventions. Catheterization and

Cardiovascular Interventions; 2006;68, 6, 974-976.

46

personal online dosimetry using computational methods · suppose we can use monte-carlo simulations...

Documents