1. NCS Report 24 http://radiationdosimetry.org/ j.vd.kamer@nki.nl
VMAT QA
NCS Report 24

2. Contents What is the NCS NCS Report 24 (VMAT QA) Starting subcommittee
Writing a report
Status of a report
NCS Report 24 (VMAT QA)
Based on earlier reports
Machine QA
VMAT and the TPS
Patient QA for VMAT

3. The NCS Website: http://radiationdosimetry.org/
Would be nice to have some extra pictures of persons doing linac QA or similar
Founded 3 September 1982
Aim:
Promote appropriate use of dosimetry for both scientific and practical applications
Tasks
Participation in dosimetry standardisations
Promotion of dosimetry intercomparisons
Drafting of dosimetry protocols
Collection and evaluation of physical data related to dosimetry

4. Typical Reports CoP QA Dosimetry TPS Audits
VMAT (NCS24); IMRT (NCS 22); Radiology (NCS19); Brachy (NCS20,14,13,4); CT (NCS11); Conventional (NCS9,8)
Dosimetry
I-125 brachy (NCS20); Radiology (NCS17,6); Photons/electrons (NCS18,5,2); Orthovolt (NCS10); Ir-192 (NCS7)
TPS
Monte Carlo (NCS16); Conventional (NCS15,12)
Audits
NCS 18 (NCS23); DRN (NCS21);

5. Starting an NCS subcommittee
Usually somebody signals a problem
New technique, how to do QA
Initially each institute for itself
Blind spots? What can go wrong? What do others do (cover your arse)
Propose setting up an NCS subcommittee
Invite colleagues
First meeting  definitive task of subcie
Final meeting: farewell dinner
Usual throughput 4 years (but could be > 10yrs)

6. Composition of a subcie
Both large and small institutes
Academic and peripheral
Different vendors
Geographical spread (Belgium)

7. General Process of Writing
Enthusiastic first meeting!
Discussions
What do you do, visiting each others sites
Have you also had this problem?
How do you deal with…
2nd meeting
Assigning chapters to couples
Start writing
Xth meeting  depression
Writing is difficult
A lot has already been published
Why bother…
Yth meeting  new boost
One chapter (almost) finished
Focus, boost
Almost there…

8. Finally, after 4 years… Finalizing report If NCS boards accepts
Submitting to external reviewers (optional)
Presenting to NCS board
If NCS boards accepts
Report in NCS format
Published on website
Informing stakeholders
Dinner
Finalizing IMRT QA report

9. Status of an NCS Report But: Keep Thinking!
According to the Law: Field Norm:
You don’t have to follow it
But motivate why + alternative
When there is an incident
Did you follow guidelines?
If not
Why?
Do fellow colleagues think that you did okay?
But: Keep Thinking!

10. Present subcommittees
Stereotactic RT
Tomotherapy
Cone-beam CT
IMRT & VMAT audit
Afterloaders for Brachytherapy
Radiation Doses & Risk Estimation
Total Body and Skin Irradiation
Electron Audit

11. NCS 24
Code of Practice for the Quality Assurance and Control for Volumetric Modulated Arc Therapy
Anton Mans AVL, Amsterdam Danny Schuring CZE, Eindhoven Mark Arends RIF, Leeuwarden Lia Vugts AMC, Amsterdam (initially: BVI) Jochem Wolthaus UMC-U, Utrecht Heidi Lotz UMCG, Groningen Marjan Admiraal VUmc, Amsterdam Rob Louwe Wellington Blood and Cancer Centre Wellington, New Zealand Michel Öllers MAASTRO, Maastricht Jeroen van de Kamer AVL, Amsterdam

12. Report Started December 2010 Report: February 2015 ~15-20 meetings
Note download date!
Report
Started December 2010
Report: February 2015
~15-20 meetings
4 chapters
65 pages
28 references
Including to NCS 8 & 22
9 figures
6 tables
3 formulas

13. # patients in AVL
T1w gd
T2w
T2-FLAIR
3D CRT
IMRT
VMAT
FFF VMAT

14. # patients in AVL 2009  20014 Decrease simple IMRT (3D CRT)
3100  2400
Also decrease IMRT
2350  1850
Strong increase VMAT
13  1730
Staring increase VMAT + FFF
0  125

15. # patients in AVL Current practice:
All VMAT, except if IMRT is better (e.g. conventional lung)
Simple plans all IMRT Breast, palliation  no wedges anymore!

16. NCS 24, VMAT QA Machine QA VMAT and the TPS Patient QA for VMAT
Static
Dynamic
VMAT and the TPS
Discrete computations, continuous delivery
Patient QA for VMAT
Gamma evaluation

17. Machine QA: ‘static’ (§2.2-4)
Static gantry
Gantry (1°), collimator (1°), MLC position (<1mm), output (0.5%), flatness1 (3%) & symmetry2 (3%)
Speed determination
Gantry, MLC leaf
Dose rate dependence
Output (0.5%); flatness3 (2%) & symmetry3 (2%)
1 Over profile
2 Over flattened area
3 Of TPS profile

18. Machine QA: ‘dynamic’ (§2.5)
Flatness & symmetry (2% 1)
5 constant doserates
Output
More than 1000MU (0.5%); dose rate changes (0.5%)
Combined Dose Rate + Gantry Speed
Dose rate and gantry speed ‘interchangeable’
Profile should be <2% of reference
1 Of TPS profile

19. Measuring Field Flatness - AVL
StarCheckMAXI gantry mounted
Gantry angle sensor
Movie recording
Analysis using ‘Home Grown’ software

20. 6 MV - S2
Variation 2% tov 0°

21. Machine QA: ‘dynamic’ (§2.5) (cont’d)
doserate & gantry speed variation
Inertia  to avoid TPS yielding undeliverable plans
E.g. small field. VMAT: 2° low speed (high dose); 38° high speed (low dose)

22. Machine QA: ‘dynamic’ (§2.5) (cont’d)
Stress Test
Inertia  to avoid TPS yielding undeliverable plans
High Dose Rate + High Gantry Speed followed by Low Dose Rate + Low Gantry Speed
Too slow switching  spill of dose to succeeding segment

23. MLC Leaf Positioning Dynamic Picket Fence Test Orthogonal to beam
Leaf position <1mm) 2 4 6 1 3 5 7
Linac head
Measurement device
Radiation

24. MLC Leaf Positioning Dynamic Picket Fence Test Orthogonal to beam
Leaf position <1mm) 1 2 3
Arrangement ensures maximum leaf speed 4 5
Slits just as Picket-Fence test 6

25. Test Frequencies See document
Many tests at acceptance or on indication
Redo if major changes!

26. VMAT + TPS VMAT: continuous dose delivery with
Moving leaves Same as sliding window IMRT (Varian)
Moving Gantry (+collimator)
Varying dose rate irrelevant for TPS
Is what the machine decides, in combination with gantry/collimator/leaf speed

27. DICOM file Control Points (CP’s) Describes the linac state
Leaf/jaw positions, couch position/angle, collimator, gantry angle, beam energy, delivered MU’s,…
Leaf/gantry/collimator move between CP’s
Linac state between CPs ‘undefined’
Usually linear interpolation between CP’s
One leading parameter
Elekta: # delivered MU’s (every 40ms)
Varian gantry angle (every 50ms)
MLC controller for MLC
Machine controller for other parameters (#MU)

28. Tolerated Deviations Two things to control Maximum deviation
Maximum duration of a (smaller) deviation

29. Interpretation DICOM file by Linac

30. VMAT Model Something needed by most planning systems
Discrete description of continuous VMAT
Used both for computation and DICOM transfer

31. (too) Simple model Compute dose @ CP’s Is okay for step&shoot IMRT
Linac delivers dose, TPS does not compute dose  incorrect
Control Point
No MU in TPS
Yes MU in TPS
# MU Linac
#MU
Control Point

32. Dosimetric Averaging
Distribute dose values between CP’s
#MU

33. Dosimetric Averaging model
Distribute dose values between CP’s
Control Point
No MU in TPS
Yes MU in TPS
# MU Linac
#MU
Control Point

34. Geometric Averaging Distribute leaf/gantry/coll. values between CP’s
#MU

35. Geometric Averaging Distribute leaf/gantry/coll. values between CP’s
Control Point
No MU in TPS
Yes MU in TPS
# MU Linac
#MU
Control Point

36. Optimal Model Compute all steps between CP’s (0.01° instead of 4°)
Monte Carlo
Control Point
Yes MU in TPS
# MU Linac
#MU
Control Point

37. TPS modelling Know your VMAT model! CP spacing (4° or 2°)
Limit modulation
Limitations linac
Description of treatment couch
Leaf tip modelling
Small fields, Minimal leaf separation

38. Patient Specific QA Class solutions
For a certain treatment site (e.g. prostate only)
Define #arcs + start/stop angle, coll.angle etc.
Define computation resolution (e.g. 4°)
Thoroughly measure class solution before clinical introduction
Can the linac handle it?
Dosimetry accuracy (3D)

39. QA Tools Ionisation chambers Film dosimetry Detector arrays
Few points, large volumes
Film dosimetry
Becoming available, still difficult for high accuracy
Detector arrays
Low resolution, sampled points
EPID dosimetry
Poorly available

40. Gamma Evaluation (1/2) 2D comparison measurement and computation:
Same place: what’s the dose (deviation in %)
Same dose: at what place (deviation in mm)
Gamma 3%/3mm
Measured and computed dose profile P1 P2

41. Gamma Evaluation (2/2) Measure grid Computation grid 206.0 302.1 272.3
258.6
294.8
222.0
205.5
206.0
202.0
228.0
Dose
Grid size 1 mm
Reference dose 200 cGy
= 0.34
< 1.0 is OK

42. Example Gamma Evaluation (1/2)
Computed dose
Profile Comparison
Measured dose (film)
Gamma Comparison

43. Example Gamma Evaluation (2/2)
Statistics
Cut-off (which values are considered? 50% of dmax?)
Pass rate (percentage of points gamma < 1)
Mean g
g 1% (near highest)
3D or 2D?
Composite plan?
Orthogonal to gantry, fluency, phantom???

44. g Evaluation with EPID @NKI 2012 - 2014

45. Induced Errors, Found with EPID dosimetry
Errors at the edge of the field
Usually no centre
Small effect on pas rate (only border g values)
Consider averaged g or g-1%)
Wrong energy (6MV vs 10MV): 10% error for prostate and lung, and 15% for H&N VMAT
Wrong collimator angle: huge effect determined with gamma evaluation, but no effect at isocenter for prostate and lung VMAT

46. Testers dummy run ≠ developers
Horizontal , two arcs
Developers of ARTFOCE
Testers dummy run
Vertical, two arcs
Checks with Octavius Phantom

47. Why Testers dummy run ≠ developers
Different planner, different constraints
Reduce plan delivery time in Pinnacle
Backup
Tester can act as backup of developer
Needed for all disciplines!
Independent test of procedures + clarity

48. Summary NCS Reports are guidelines
VMAT QA, the dynamic equivalent of IMRT QA
Patient-specific QA:
Difficult to find good measures
 Make sure basic QA is OK!

49. Thank you for Your attention