Uniform Scanning andEnergy Stacking with
Proton Beams
AAPM Continuing Education Session
22-Jul-2010
Outline
Introduction to Technique - Moyers (15 min)
» description of delivery techniques and terminology
» radiobiology
» lessons from scanning electron beam incidents
» advantages and disadvantages of technique
Design and Implementation of Safe Delivery Systems - Anferov (15 min)
» potential hazards
» example hazard mitigations
Practical Aspects - Hsi (15 min)
» optimization of scan and stacking patterns
» multi-element detectors for measurements
» scanning and stacking specific QA
Questions - All (10 min)
Scanning Terminology
scanning modes (as defined by DICOM-RT ion)» none
» uniform scanning
» modulated scanning
repainting
uniform scanning patterns» Lissajous
» circular (single or multiple)
» raster (rectilinear)
» spiral
» triangle
History of Uniform Scanning in the Clinic
Early» Michael Reese e patient scan 1955
» Uppsala p Lissajous beam scan 1957
» Sagittaire e Lissajous beam scan 1970
» Berkeley He, Ne raster and circular beam
scans 1985
Recent
» Mitsubishi p, C circular beam scan 1995
» IUCF/MPRI p raster beam scan 2005
» IBA p triangle beam scan 2007
» Mitsubishi p spiral beam scan 2011
» Sumitomo p spiral beam scan 2011
Uniform Dose Coverage ofTarget in Depth Direction
energystacking
rotatingpropellors
ridgefilters
steppedcones
Energy Stacking Methods
direct extraction from accelerator
rangeshifter near accelerator
rangeshifter near gantry
rangeshifter in radiation head
Rangeshifter Types
binary slabs
linear double wedges
circular double wedges
circular steps
Number of Requestable Energies
Example Method Commentssynchrotron interpolation 18,000 energies between 70
and 250 MeVsynchrotron pre-programmed 256 energies (approximately
1 mm steps)cyclotron RS in SY 1 mm range stepscyclotron RS in head 30 range steps
Energy/Range Stability -During Treatment
acceleratorenergystability
RSthicknessstability
patientthicknessstability
Energy/Range Reproducibility -Day-to-Day
acceleratorenergyreprod.
RSthicknessreprod.
note green arearepresents 0.18 mm
Energy Switching Time
variable energy synchrotron» could change energy multiple times
during each cycle
» typically only a single energy isextracted each cycle
» verify energy before extraction
» need to reconfigure SY
fixed energy cyclotron» move RS ( 0.2 s)
» verify energy before delivery
» for SY RS, need to reconfigure SY
» for head RS, do not need toreconfigure SY
MU Considerations
Uniform dose coverage concerns
» flux non-uniformity during delivery
» shifting scanning patterns
» starting or stopping beam delivery inmiddle of scan pattern
Better dose uniformity with integralnumber of repaintings and largernumber of repaintings.
Shallow layers use a small fraction ofthe total MU; difficult to repaint.
Flux rate, scan pattern, scan speed,and number of repaintings must becarefully balanced.
Typically the MU per portal is restrictedto a minimum value.
Das et al., (1994)
Interplay with Patient Motion
motion of beam versusmotion of patient» if a person walks back and
forth through a scatteringsprinkler, will get a little wet
» if a person walks back andforth through a scanningsprinkler, may stay dry ormay get soaked
fast uniform scanning ofeach layer is typicallyfaster than respirationbut slow energy stackingmay be an issue.
scanningsprinkler
scatteringsprinkler
Radiobiology of Scanned Beams
thus far no direct comparisonof uniform scanned protonbeams to scattered protonbeams
experiments with scannedelectron beams showed RBEup to 1.29 depending uponscan pattern
Meyn et al., (1991)
Lessons Learned from Previous Incidentswith Scanned Electron Beams
Sagittaire example» bending magnet power supply stuck at wrong high energy (32 MeV)
» energy feedback loop adjusted energy so beam would pass throughenergy analyzing slit in bending magnet stuck at wrong high energy
» scanning magnet power supply set at correct low energy (13 MeV)
» upstream dose monitor measured correct whole beam flux butfluence distribution downstream concentrated in middle of field
» one patient had parallel opposed posterior cervical strip fieldsresulting in 800 cGy to spinal cord in one fraction
» medical problems for patient within 45 minutes
Lessons for safety» energy interlocks for accidents - DAILY QA OF THE RANGE IS NOT
SUFFICIENT
» downstream fluence distribution detectors
Advantages and Disadvantages
advantages» uniform dose distribution for all energies and field sizes
» smaller loss of range for large fields compared to scatterer technique
» no need for electromechanical scatterer exchangers
» higher particle use efficiency / less neutron production
disadvantages» requires additional time to switch energies
» minimum MU constraint for portal
» increased interplay with patient motion
» requires diligent safety system
Part 2
Making UniformScanning safe
Safe Design Practices
If it can break –it will
i.e consider all failure modes and look at the outcome
Failure Modes & Effects Analysis (FMEA) process:
» Define failure modes and associated risks
» Add mitigations that– Reduce probability of a failure mode
– Detect failure and stop before any harm is done
Use KISS principle:
» Keep It Simple Stupid !
Uniform Scanning Features
Insensitive to beam misalignment
High instantaneous dose rate
Beam spot size can vary from 0.5 to 1.0Line Spacing without perturbing uniformity
Scan pattern can be started and validatedprior to delivering dose to the patient !
1
10
100
1000
Dose Rate
(Gy/min)
Double
Scattering
Uniform
Scanning
Spot
Scanning
Dose rate in a beam spot for average 2Gy/min to 103
cm3
0
1
2
3
4
5
6
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
Line Spacing / Sigma
Ov
ers
ca
n/
s
0
1
2
3
4
5
6
Rip
ple
[%]
Overscan
Ripple
Hazard Ratings
No perceptible effect
Small loss of performance
Loss of product function, but no damage touser, patient, equipment.
Possible injury without irreversible damage
Possible injury with permanent damage
Death of user or patient
MINOR
MODERATE
HIGH
CRITICAL
Insignificant
Catastrophic
New Hazards due to Scanning System
A. High dose rate in a beam spot can causelarge dose errors if scanning stops» 5% dose error can accumulate in 5 msec.
» 100% dose error can accumulate in 100msec
B. Non-uniform transverse dose distribution dueto errors in the scanning pattern.
» Accumulates over the course of the treatment.
Critical
High
Sensitivity to Beam Failures
strongweak*7. Beam Energy / per layer
strongweak6. Intensity Fluctuation
moderateweak5. Beam Intensity Rate
moderaten/a4. Beam spot halo
strongweak3. Beam spot shape
strongweak2. Beam spot size error
strongweak1. Beam misalignment
SpotScanning
UniformScanningBeam Failures
* Only if using passive range modulation (ridge filters, range shifter in the nozzle)
Safety Mitigations
Start scanning verification prior to dose delivery» Apply checks that validate scan profile, scan amplitudes and scan
accuracy.
Perform scanning system health validation at a fast rate(~1kHz) and interlock beam delivery.» Redundant hardware checking mitigates critical hazard of burning a
hole through the target.
Monitor Field Flatness, Size and Symmetry throughout thetreatment using segmented ion chamber.» This check validates accuracy of the dose delivery process.
Scanning System Health Checks
AGeneratorEvery 0.1 ms must receive a trigger pulseindicating Generator updated its output
A,BMagnetMagnet health: analog circuit monitorsvoltage from the pickup coils.
BGenerator,Magnet
Waveform stability: waveform parameters donot change during treatment
BGenerator,waveform
Measure waveform parameters: Min and Maxvalues of Currents, Voltages, Frequencies
BPower Supplyerrors
Every 1 ms PS output is within tolerance fromGenerator
AGenerator orPower Supply
Every 1 ms PS output change indicate beamspot motion 1 cm or more
HazardFailure ModeHardware Health checks
Scanning is only part of the picture
Treatment energy setup validate beam penetration range
Lateral Beam spreading validate scanning safety
Dose modulation in depth validate ridge filters / range shifters
Dose conformation to target validate collimator & bolus
Measure the dose Redundant dose counters,MUs agreement
Dose delivery system
Safety Checks
Safety Summary
Compared to a Double scattering systemUniform scanning adds two new hazards:» Stopped scanning
» Incorrectly executed scanning
With dedicated safety electronics monitoring health of thescanning system uniform scanning can be safe and robustalternative to both double scattering and pencil beamscanning
Part 3
Practical aspects
–utilize uniform
scanning & discreteenergy stackingprotons for treatments
Maglev train at China with maximumspeed of 431 km/h (268 mph)
Virtual source distances forvarious scanning magnets
Combined scanning magnet» single virtual source distance
Dual scanning magnets» two virtual source distances
» effective source size and distance aresame for both axes
Parallel scanning» single "infinite“virtual source
distance
Requirements of clinical performance
Dose Reference Volume(DRV)
Transverse –lateral extentInside 2-times penumbra
Penumbra width
Depth –longitudinal extentWithin modulation width
Requirements for lateralextent at above are onlyapplied for depths with thelongitudinal extent.
Scan Pattern Optimization -spot shape & size in air
Spot shape in air, i.e. proton fluencedistribution at the entrance of patient.
Large section near patient is under vacuumfor IBA delicated nozzle at Essen. Measuredspot sizes for this nozzle are ~4 and 6mm forbeam ranges of 32 and 20 cm in circularshape. However, when a beam-positionmonitor located at entrance of gantry wasinserted as extra material into beamline,elongated beam spot was observed as shownat most left of top panel.
Spot sizes in air also measured for auniversal nozzle at MPRI are 6 to 14mm forvarious ranges as shown at bottom panel.Larger spot size for universal nozzle is due tono applied vacuum from scanning magnet topatient.
Scan Pattern Optimization -spot size in patient
Spot sizes in air σair measured atMPRI was fit as a function ofrange R; shown dash line.
Spot sizes σpt due to scatters inpatient is calculated by 0.02275R +0.12085E-4R2 as in Hong et alpublication as open circles.
Total spot size including initialspot size and scatters in patient asclose circles is calculated by foreach beam range
σtot (R) = (σair2+ σpt
2)1/2
Scan Pattern Optimization -Path spacing & over-scan inside field
Optimization based path spacing & over-scan inside field
Effects of collimation -Over-scan beyond field edge
Over-scan distance beyond edges of beam-limiting devices
- upstream trimming collimators & patient aperture
Scan Pattern Optimization -various depths
Scan pattern optimization of spot size and path spacing focuson depths within longitudinal extent and the center ofmodulated protons at beamline isocenter.
Energy Stacking Optimization -Depth Spacing
Widths of non-modulated Bragg-Peak (BP) depth dosesvaries with proton energy (i.e. beam range) and delivery system
3 options for stacking energy layersto form required flatness overlongitudinal extent are needed forMPRI TR2. For options with rangesof 12-20cm and 20-27m, depthspacing between consequent layers is0.6 cm. However, 0.3 cm depthspacing is required for ranges of 4-12cm when the width is only 0.5 mm for4 cm range. For widths from 3cm to1.5cm, 4 options are needed forOKC-IBA US beam-line with ~0.6cmdepth spacing for all beam ranges.
Energy Stacking Optimization -Weights of energy layers
After depth-spacing is chosen for each option,original weights of energy layers were obtainedfor standard range of 16cm by theoretical modelas shown blue points. During commission, depthdoses with original weight were measured asshown blue points at bottom panel. A correctingalgorithm was used to adjust ~8% title.Optimized weights of energy layers were thenobtained as shown red points in both panels.
Because widths of non-modulated protons variessignificant between options, using optimizedweights from standard options results significanttilt on depth doses for non-standard options.The correcting algorithm described above isused for obtaining optimized weights for variousoptions during commission. Similar procedurehas been internally performed by IBA vendor.
Energy Stacking Optimization -between proton delivery systems
Weights of energy layers Energy options –range-shifter thickness and MU courting per proton
Scattering generated by materials used in energy stacking
Weights of optimized energy-stacking layers vary betweenenergy options; depend on range-shifter thickness and MU courtingper proton. Although the trend ofweights used energy layers issimilar between different protonbeam-lines, subtle difference canbe related scattering generated byenergy layers in different beam-lines.
Depth Dose Measurements
Use of a single chamber to measure the dose at all depths requires repeatingthe whole delivery sequence for each depth.
Use of a MLIC (multiple layer ionization chamber) allows measurement ofdose at many depths using a single delivery.
0
10
20
30
40
50
60
70
80
90
100
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Depth [cm]
DD
[%]
R 27cm Scan 3cm
CircularPt-by-Pt 3cmcircular
MLIC F.S. 10cm
Pt-by-Pt, F.S. 10cm
MLIC F.S. 10.cm
Dmitri et. al. 2007
Lateral Dose Profile Measurements
Film
Step-by-step method with mini-chamber
MPIC (multiple pad ionization chamber) - 1Dor 2D configuration
In-plane Profiles
0
10
20
30
40
50
60
70
80
90
100
110
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8
X (cm)
Pro
file
(%)
Water Phantom
MPIC
Treatment failure recover
Partial treatmentdelivery
Routine QA -Range-shifter of energy layer
Check beam range at off-axis positions to verify constancy ofrange-shifter material thickness used for energy stacking.
Routine Modulation QA
Ensure that files storing weights of energy layers forvarious modulations of each option are not corrupted.
» Check weights of energy layers for standard condition daily.
» measure depth dose distributions monthly.
» compare routinely used files with secure master filesannually.
For a proton system that weight and scanning aptitude ofeach energy layer are determined by an algorithm asfunction of beam range and modulation, ensure that thecalculation algorithm is not corrupted
» Check weights of energy layers for standard condition daily
» measure depth dose distributions and lateral profiles monthly.
Summary of practical aspectsfor utilizing uniform-scanning protons
Optimizations of scanning pattern and energy-stacking need beperformed to satisfy clinic requirements on lateral andlongitudinal extent for providing good treatments.
Specific dosimeters for depth doses and lateral profiles are neededfor commissioning a US proton beamline efficiently.
Recovery treatment for partial delivery is required.
Beam range verify at off-axis positions is necessary when largearea of range-shifter is used for energy stacking duringcommission.
Routine QA for off-axis range constancy as hardware andmodulation as software should be performed to assure correctdose delivery.
References
Moyers, M. F. “Proton Therapy” The Modern Technology of Radiation Oncology: ACompendium for Medical Physicists and Radiation Oncologists ed. van Dyk, J.(Wisconsin: Medical Physics Publishing, 1999) p. 823 - 869.
Meyn, R. E. Peters, L. J. Mills, M. D. Moyers, M. F. Fields, R. S. Withers, H. R. Mason,K. A. "Radiobiological aspects of electron beams" Frontiers of Radiation Therapy andOncology 25 eds. Vaeth, J. M. and Meyer, J. L. (S. Karger AG Basel, Switzerland, 1991)p. 53 - 60.
Moyers, M. F. "LLUPTF: eleven years and beyond" Nuclear Physics in the 21st Century(New York: American Institute of Physics, 2002) p. 305 - 309.
Moyers, M. F. Vatnitsky, S. M. Practical Implementation of Light Ion Beam Treatments(Wisconsin: Medical Physics Publishing, 2011).
V. Anferov, “Scan pattern optimization for uniform proton beam scanning”, Med. Phys.36(8), 3560 (2009).
J.B. Farr et al., “Clinical characterization of a proton beam continuous uniform scanningsystem with dose layer stacking”, Med. Phys. 35(11), 4945 (2008).
Das, I. J. et al., "Dosimetric problems at low monitor unit settings for scanned andscattering foil electron beams" Med. Phys. 21(6), 821 (1994).
Nichiporov, D. et al., “Multichannel detectors for profile measurements in clinical protonfields”, Med. Phys. 34(7), 2683 (2007).
Questions
Interesting Stuff
http://www.ngsir.netfirms.com/englishhtm/Lissajous.htm