1. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
1
Master Thesis :
The clinical performance of the DAVID-system for the
in vivo verification of VMAT irradiation
Presented by Mustafa Saibu Danpullo
1st supervisor
Prof. Dr.B.Poppe
2nd supervisor
Dr. HK. Looe,
2. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
2
Layout
I Introduction
II Theory
VMAT and IMRT
MLC Design and Agility
MWPC and DAVID system
III Materials & Methods
Equipment, alignment, patient data, stability of DAVID chamber
Beam property, Error detection
Deconvolution
DAVID QA software
IV Results / Discussion
V Conclusion
3. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
3
I Introduction
IMRT (Intensity-modulated radiation therapy)
VMAT (Volumetric Modulated Arc Therapy):
Why In vivo verification ?
ICRU report 24 (1976) recommended that certain types of tumors
requires improve accuracy from 5% to 3.5%.
To detect Equipment-related errors and deviations from the initial plan
Complexity of planning and delivery techniques increases risk for
treatment-related error incidents.
In 1992 to 2007, more than 4,000 near misses without adverse
outcome to patient’s case were reported, more than 50% were related
to the planning or treatment delivery stage.
4. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
4
In vivo dosimetry methods:
in vivo intracavitary dosimetry with TLD
Diodes
DAVID (Device for Advanced Verification of IMRT deliveries
In-vivo verification during treatment
Online measurement of differences in dose to reference
Error detection of the Multi Leaf Collimator (MLC)
I Introduction
5. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
5
II Theory
Mostly Siemens, Elekta and Varian have introduced new LINAC
control systems that will be able to change the MLC leaf positions
IMRT uses many small fields to generated by beam-shaping
devices (MLC) to deliver a single dose of radiation
IMRT: Intensity-modulated Radiation Therapie
6. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
6
II Theory
VMAT : Volumetric Modulated Arc Therapy
VMAT is a rotational IMRT that can be delivered using
conventional LINAC with MLC
Elekta and Varian have introduced new LINAC control systems
that will be able to change the MLC leaf positions and dose rate
while the gantry is rotating.
Precise Beam infinity and Rapid Arc
7. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
7
A schematic drawing of the Siemens type A, Elekta type B and Varian type C MLC [18]
Stepped leafs for different manufacturers [34]
II Theory
8. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
8
II Theory: Agility MLC design
• 160 tungsten leafs,
• rounded arc edge,
• 5 mm width,
• High speed(2x normal MLC) of up to
3cm/sec,
• large field MLC enable clinicians to
shape radiation,
• extremely low transmission of about
<0.5%
• 45 cm isocenter clearance from
accessory holder.
MLC Motor
9. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
9
Inventor:
Prof. Georges Charpak,
France,1968
Nobel Prize in Physics (1992)
II Theory: Multi wire proportional chamber (MWPC)
The DAVID chamber is a multi-wire
ionization chamber designed by PTW
Freiburg based on Charpark's multi wire
proportional chamber.
10. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
10
Compton scattering Electric field causes electrons move to the
anode(wire) and ionizied atoms/molecules to the
cathode(plate)
Each detection wire accumulates charge which
loads a C.
After the voltage at the capacitor is read out, it is set
to zero and charged again
The voltage achieved is read out by the associated
amplifier at a rate of 1 Hz.
Performed by multi-channel electrometer
(MULTIDOS) + additional Software
.
II Theory: DAVID system functioning principle
11. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
11
Signal interpretation:
Ri: reading of a single channel (ion charge collected)
C: cross section of the lengthy collection volume along the wire
Ii: ionization density (x1 start of wire, x2 end of wire)
li1-li2: aperture of the associated leave pair
II Theory
II Theory
12. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
12
Front Plate
Back Plate
Air Volume
II Theory: DAVID system signal recording
3 groups of secondary electrons contributing to the signal:
a)“primary signal”
b)scattered signal
c)leakage radiation
(background signal)
13. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
13
VMAT Planning:
Treatment Planning System: ONCENTRA Masterplan Version 4.3
ELEKTA Synergy accelerator with an Agility 80 leaf-pair MLC
• Desktop Pro TM 7.011 is Elekta's third generation fully integrated
digital control system. MOSAIQ, DAVID software version 2.0
DAVID T34065
III Materials and Methods
14. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
14
Patient data configuration chart
III Materials & Methods
Reference
1st session
15. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
15
DAVID Analysis
• PTW: DAVID 2.0 software
III Materials & Methods
Warning level: 3%
Alarm level: 5%
16. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
16
III Materials & Methods:
•VMAT: 4 (1 H&N, 3 Prostates)
•180° to -180° Clockwise and anti clockwise
Stability of the DAVID system
•IMRT : 1 (Prostate)
•0°
•90°
•270°
(Prostate and Head and Neck) 14 days
17. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
17
III Materials & Methods:
The beam property of the DAVID chamber
•Percentage depth dose (PDD)
•Roos chamber 34001
•MP3 water phantom
•Transmission factor for 6 and 15 MV
•Semifex T31010 (Diff Field sizes)
Setup conditions
• With and without DAVID
• SSD 100 and 80 cm
• Photon energy 6 and 15
MV
• Different field sizes
18. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
18
1. Successive opening of 1 leaf on 1
side
III Materials & Methods:
VMAT Plan Editing for error detection
• MATLAB script to change the MLC-positions
3. Field shift of a leaf gap (size of leave
gap remains)
2. Successive shift of a leaf gap (size of
leave gap remains)
19. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
19
1. Successive opening of 1 leaf on 1
side
VMAT Plan Editing for error detection
• MATLAB script to change the MLC-positions
3. Field shift of a leaf gap (size of leave
gap remains)
2. Successive shift of a leaf gap (size of
leave gap remains)
III Materials & Methods:
20. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
20
Deconvolution
III Materials & Methods:
S(x) measured signal as convolution of
P(x) True „dose“ profile with
LRF fɛ(x).
S(x) = P(x) * f(x)
„van Cittert“ iterative deconvolution
algorithm
21. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
21
opened MLC at every 10th interval from 1st to 80th
pairs.
Nine MLC slit through the entire DAVID chamber
IV Results and Discussion: Alignment
22. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
22
IV Results and Discussion: Stability of DAVID System
• IMRT: prostate
• Deviation of ±2% (+2%)
• VMAT: prostate
• Deviation of ±1% (-1%)
23. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
23
IV Results/Discussion: Stability of DAVID System
• IMRT: prostate
• Deviation of ±2% (+2%)
• VMAT: H&N
• Deviation of <0.5%
24. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
24
IV Results/Discussion: Transmission factor
The average KDAVID
• 0.939 ±0.003 for 6 MV
and
• 0.953 ± 0.004 for 15 MV
Reduction of dose at isocenter
due to 8mm of PMMA
By measuring the attenuation
factor the output value can be
corrected.
Attenuation of the beam by the DAVID chamber
25. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
25
100 cm SSD,
Increased about
0.32% with DAVID
No change of
Dmax
1.4 cm with
and without DAVID
80 cm SSD,
An increased about
4.26% with DAVID
slight change of
Dmax
1.3 cm with DAVID
1.4 cm without DAVID
IV Results/Discussion: Changes in PDD
26. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
26
100 cm SSD,
An increased about
0.67% with DAVID
Dmax
2.6 cm No change
80 cm SSD,
An increased about
18% with DAVID
slight change of
Dmax
1.8 cm with DAVID
2.4 cm without DAVID
IV Results/Discussion
27. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
27
Surface dose
Increase with increase in field size
Increase with increase in energy
Increase with decrease in SSD
IV Results/Discussion
28. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
28
Pronounce with
decrease in SSD and,
increase in photon energy and
increase in field size.
Increase in surface dose is
due to scattered secondary
electrons from the DAVID
chamber reaching the
water phantom surface
(electron contamination).
6 cm
45 cm
100 cm
IV Results/Discussion
29. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
29
Deconvolution test
by iteration method
IV Results/Discussion: Deconvolution
30. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
30
Deconvolution
•does not depend on the length of the slit.
•40 cm slit results showed a small decrease in the tail signals.
10 x 10 cm
10 cm slit
40 cm slit
20 cm slit
30 cm slit
IV Results/Discussion
31. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
31
IV Results/Discussion:
Deconvoluted slit signal at 40th and 65th wire
Deconvolution test
32. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
32
10x10cm fields before and after deconvolution
Single arc prostate plan before and after deconvolution (3 mm )
IV Results/Discussion: Deconvolution with DAVID software.
33. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
33
before deconvolution 6.1 mm
Gradients of the linear fit
before deconvolution : 0.47 and
after deconvolution 2.9 mm
Gradients of the linear fit
after deconvolution : 0.94
.
IV Results/Discussion:
Enhance sensitivity after deconvolution
34. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
34
before deconvolution after deconvolution
H&N 6mm error.
.
IV Results/Discussion
False alarm/warning effect before deconvolution.
The effect is eliminated after deconvolution
35. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
35
Undeconvoluted
deconvoluted
Measuring the deconvolution matrix with
the DAVID software as manual. LSF of single middle slit is measured and
used to generate the 80x80 LRF matrix with
MAT LAB and installed in the DAVID software
for deconvolution.
IV Results/Discussion: Limitations of DAVID-160 system
36. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
36
2mm MLC error, single bank shift
2mm MLC error, single leaf
Analyzing both the maximum deviation
and total deviation in two different plots
at the same time
IV Results/Discussion
37. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
37
Max Dose 75.99GyMax Dose 57.27Gc
IV Results/Discussion: Artificial MLC bang shift Error 2
cm
38. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
38
0
1
2
3
4
5
6
7
8
0 8 10 12 14 16 18 20
shift / mm
doseincrease/Gy
maximaldeviation/%
Undetectable error! Design of the chamber
DAVID: Gap-Shift (prostate with OAR: rectum back wall)
IV Results/Discussion
39. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
39
DAVID Quality Assurance software (DQA)
MAT LAB 2011b and 2012b
analyzes the daily session for all patient's data with 2 clicks online.
• NDD- Non deconvoluted data
• ˆDD-Deconvoluted data
• ED- Electrical data
IV Results/Discussion
40. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
40
Analysis only specific data on
specific date, session and only print
out the deconvoluted data (DD)
IV Results/Discussion
41. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
41
Patient text file from DQA software
Sample patient 2014-01-12
Display only data's with MLC error
indicating the data type (DD),Beam
number, Session, segment and
particular MLC with error.
IV Results/Discussion:
42. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
42
The DQA software displays the warning and alarm errors
Warning and error dialogs at future date entry and invalid date entry respectively
year-month-day `yyyy-mm-dd'
IV Results/Discussion
43. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
43
VMAT and IMRT plans sessions
IV Results/Discussion:
44. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
44
DAVID chamber:
• Linear dependency on leaf opening
• Sensitivity dependent on leaf gap opening
• How much radiation pass through the opening
• Deconvolution double the sensitivity
DAVID is design for specifics LINACS
Independent from the LINAC
In-vivo verification of MLC malfunction during VMAT
• Undetectability field shifts due to chamber design
(Suggestion: perpendicular wires or gradient)
V Conclusions
45. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
45
V Conclusion
Single MLC bank shift error
Maximum and total deviation to be analysed
Deconvolution matrix
To be generated for each linac
To be generated by single single slit
• In comparison to other techniques measurement of undisturbed
signals
-> no dependence on patient position(EPID)
-> measurement of the complete delivered dose(TLD, diode or MOSFET
detectors)
• Suggestions for future development: Standard design for software
• Deconvolution program to be integrated,
• DQA to be integrated
• Design two chambers perpendicular to each other
Co-operate directly with LINAC vendors for specifics designs
46. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
46
Sources
1. [1] Ezzel GA, Galvin JM, Low D, Palta JR, Rosen I , Sharpe MB, Xia P, Xiao Y, Xing L and Yu CX. Guidance on delivery,
treatment planning, and clinical implementation of IMRT: report of the IMRT subcommittee of the AAPM radiation
therapy committee Med Phys 2003; 30:2089-115.
2. [2] ESTRO Booklet No. 9, 2008. Guidelines for the verification of IMRT, edited by Ben Mijnheer and Dietmar Georg.
ISBN 90-804532-9.
3. [3] Schneider F, Polednik M, Wol D, Steil V, Delana A, Wenz F, Menegotti L. Optimization of the gafchromic EBT
protocol for IMRT QA. Z Med Phys 2009; 19(1):29-37.
4. [4] Poppe B, Blechschmidt A, Djouguela A, Kollho R, Rubach A and Harder D. Two-dimensional ionisation-chamber
arrays for IMRT plan verication. Med Phys 2006; 33:1005-15
5. [5] Poppe B, Thieke C, Beyer D, Kollho R, Djouguela A, Rühmann A, Willborn KC and Harder D. DAVID-a translucent
multi-wire transmission ionization chamber for in vivo verication of IMRT and conformal irradiation techniques. Phys
Med Biol 2006; 51:1237-48.
6. [9] Poppe B, Looe H K, Chofor N, Rühmann A, Harder D and Willborn K. Clinical Performance of a Transmission
Detector Array for the Permanent Supervision of IMRT Deliveries. Radiother. Oncol. 2010; 95:158-65
7. [10] Looe H K, Harder D, Rühmann A, Willborn K and Poppe B. Enhanced accuracy of the permanent surveillance of
IMRT deliveries by iterative deconvolution of DAVID chamber signal proles Phys. Med. Biol. 2010; 55:3981-92
8. [11] Heukelom, S., el al.Wedge factor constituents of high-energy photon beams: Head and phantom scatter dose
components Radiother. Oncol. 32: (1994) 66-3
9. 12] Jursinic, P. A. Changes in incident photon uence of 6 and 18 MV x rays caused by blocks and block trays Med
Phys 26 (1999) 2092-8
10.[13] v. Klevens, H. Dependence of the tray transmission factor on collimator setting and sourcesurface distance Med.
Phys. 27 (2000) 2117-3
11.[14] Sharma, S.C., el al., Recommendations for measurement of tray and wedge factors for high energy photons Med
Phys 21 (1994) 573-5
47. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
47
Thank you for your
attention
48. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
48
Additional Slides
49. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
49
50. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
Description of the Elekta synergy DAVID160 system
50
David58
David160
51. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
51
•Φ constant with depth
(small # interactions)
•Same # electrons set in
motion in each square
•i.e., interactions per
volume constant through
target
•dose reaches a
maximum at R (kerma
constant with depth,
equals absorbed dose
beyond )
Number of electron tracks set in motion by
photon interaction
52. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
52
53. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
53
54. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
54
55. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
55
56. WG Medical Radiation Physics, Pius-Hospital and Carl von Ossietzky University, Oldenburg
56