Quantification of beam latency using AlignRT

Philip Yeo
Radiotherapy Engineer
University Hospital Southampton
UK
Rachel Barlow
Deputy Head of Radiotherapy Physics
University Hospital Southampton
UK

1. QUANTIFICATION OF BEAM
LATENCY USING ALIGNRT
Rachel Barlow & Philip Yeo

2. 2

3. 3

4. GUIDANCEDOCUMENTS
TG14
7
TG14
2
TG30
2
“III.B.4.b. Temporal accuracy (latency).
System latency and time (frequency) of
tracking should also be evaluated.”
4

5. GUIDANCE DOCUMENTS
TG302
5

6. PATIENTS
MOVE!

7. 7
CLINICAL IMPLICATION OF
LATENCY
600 MU/MIN

8. IPEM 81

9. 9

10. 20XX PITCH DECK 10

11. PHANTOM CHOICE
11
CIRS 4dCT

12. 20XX PITCH DECK 12

13. 01/12/2023 13

14. SEE NOTES
20XX PITCH DECK 14

15. 20XX PITCH DECK 15

16. 20XX PITCH DECK 16

18. Linac
•Beam off
•“MV ready” (TB)
Linac
•Beam on
•User
•Beam hold
•AlignRT
•Beam off
•User
Lead
•Static
•@0mm
Lead
•Start motion
Lead
•In motion
•@10mm
Lead
•Stopped
•Back @0mm
•Monitoring
•@0mm vert
•Vert RTD = RED
•@10mm
•Vert RTD =
GREEN
•<10mm

19. ANALYSIS
19

20. ANALYSIS
20
•Locate the pixel position of the lead marker on
the EPID image of the static phantom

21. 21
ANALYSIS

22. 22
ANALYSIS
TakeFirst Derivative

23. ANALYSIS
23
Distance travelled by the Lead Marker in pixels:
D = |LeadPosn
static–LeadPosn
moving|

24. RESULTS
20XX PITCH DECK 24

25. 01/12/2023 25
y = 0.1383x + 9.7904
R² = 0.8313
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
9.5
10
10.5
11
11.5
12
12.5
LAD Latency – EPID Method
138 ms

26. UNCERTAINTIES
26

27. y = 0.9993x + 0.123
R² = 1
0
5
10
15
20
25
30
35
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Measured cycle time
(mean, with stopwatch)
Set cycle time (in CIRS phantom software)
Set cycle time vs measured

31. •Pixel size26 cm
1024 pixels
•Frame Rate
EPID uncertainties
EPID =electronic portal imaging device aka MV
panel

32. Project aims and methods
Create an electronics-based method to characterise
gating latency in AlignRT.
Easily repeatable for routine QA.
UHS and BNHH: Elekta Linacs.
Compare results against portal imaging method and
literature.

33. Literature inspiration
(Lempart, et al., 2016)
Solenoid Phantom
Targ I
PIN diode circuit

34. Device development and data collection
System Diagram
Magnetron Current circuit and Solenoid phantom
System Circuit Diagram

35. Device development and data collection
Time stamps
generated when:
- Phantom moves
in and out of
gating tolerance.
- Every Mag I
pulse.
- Data sampled at
800Hz (every
1.25mS).
Microcontroller and Mag I
Conditioning Circuit
Mag I Signal Cable
Data Collection
PC
12V Supply
Solenoid
Phantom
Time
Mag I
145 A
Mag V
-40 kV
3.2 ms

36. Device development and data collection
Phantom in and out of
vertical tolerance

37. Data analysis method (beam-off latency)
130450000
130500000
130550000
130600000
130650000
130700000
1
10 19 28 37 46 55 64 73 82 91
100 109 118 127 136 145 154
Program Running Time uS
Printed Row of Time Stamps
Sample Beam-Off Latency Calculation
Solenoid Movement Mag I Pulses
Final Mag I pulse TS – Solenoid TS
=
130,648,638uS – 130,549,203uS
=
99.435mS Beam-off latency
Solenoid time stamp
Final Mag I time stamp

38. Results and discussion (beam-off)
Beam-Off Latency (mS)
LINAC Mean
Latency
Standard
Deviation
LAC 98.55 19.28
LAD 102.01 7.65
LAF 110 6.74
LAH 100.8 8.35

39. Data analysis method (beam-on latency)
150400000
150600000
150800000
151000000
151200000
151400000
151600000
1
33 65 97
129 161 193 225 257 289 321 353 385 417 449 481 513 545
Program Running Time
Printed Row of Time Stamps
Sample Beam-On Latency Calculation
Solenoid Movement Mag I Pulses
First Mag I pulse TS – Solenoid TS
=
151,290,551uS – 150,552,482uS
=
738.069mS Beam-on latency
Solenoid time stamp
First Mag I time stamp

40. Results and discussion (beam-on)
Beam-On Latency (mS)
LINAC Mean
Latency
Standard
Deviation
LAC 774.5 77.13
LAD 722 22.79
LAF 664.58 21.85
LAH 788.68 67.07

41. Electronics Method Future Prospects
-Refine hardware, and software: ease
of set up, more professional.
-Change phantom size (larger ROI),
more clinically relevant frame rate.
-Use with different linacs.
-Use with other SGRT systems
(Exactrac Dynamic).

42. 0
20
40
60
80
100
120
140
160
180
200
LAC LAD LAF LAH
Latency (ms) for each Linac
EPID Method Electronics Method

43. CONCLUSION

44. FUTURE PLANS
METHOD AND LINAC
COMPARISON
DUMMIES GUIDE!
44
BEAM ON

45. THANK YOU /
QUESTIONS?
[email protected]
philip.yeo@ uhs.nhs.uk