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