beam delivery systems: scattering, scanning, w/wo - erice 2009

Beam Delivery Systems:

Scattering, Scanning, w/wo Gantries

or

Cost Effective Particle Therapy?

Ion Beam Therapy Workshop; Erice,

2009

Jay FlanzMGH/FBTC

Harvard Medical SchoolFlanz 2009 - Erice

What is a Beam Delivery System?• Convert the raw accelerator beam to a clinical treatment beam.

– What are the parameters needed for a clinical beam?

• Spread out the beam (5d) on the target according to prescription (Conformal)

• Deliver the prescribed dose in each of the 5 dimensions

– What are the associated accelerator beam parameters

• See Below

• Not Usually discussed: Beam timing, Emittance

• Monitor the parameters of the treatment beam.

• Direct the treatment beam to the Target in desired orientation

Accelerator Beam Treatment Beam

o Beam Energyo Beam Sizeo Beam Shapeo Beam Currento Beam Timing

o Beam Rangeo Dose Distributiono Dose Distributiono Dose Rateo Organ motion/Radiobiology

These are

connected!

Flanz 2009 - Erice

Longitudinal Beam Spreading:

Spread Out Bragg Peak (SOBP)

3

12

34

5

6

5

43

2 1

E0

6 5 4 3 2 1

A non-moving method to create an

SOBP for a specific Modulation Width

This is “Passive”

Ridge Filter

What if the target has a finite Longitudinal Extent?

Flanz 2009 - Erice

Beam is

scattered so

there is no

position

dependence

4

Range Modulator Wheels

Courtesy of

B.Gottshalk, Ph.D.

Range Manipulation

Range Compensator

Distal Edge Shaping

Courtesy of

B.Gottshalk, H. Kooy

Collimator

Flanz 2009 - Erice

Flanz 2009; Riyadh 5

Depth [mm]

0 20 40 60 80 100 120 140

Do

se

[%

]

0

20

40

60

80

100

120

Depth [mm]

0 50 100 150 200

Do

se

[%

]

0

20

40

60

80

100

120

Beam range: 17.19 cmModulation: 6.78 cm

0 20 40 60 80 100 120 140 160

Do

se

[%

]

0

20

40

60

80

100

120



Depth [mm]

Current Modulation - DYNAMIC

A Requirement of some Proton Therapy Accelerators

Paganetti

Flanz 2009 - Erice

Transverse Spreading Options

• Passive Scattering

– Single Scattering

– Double Scattering

• Wobbling (Beam Scanning with Scattered beam)

• Pure Magnetic Scanning

• Combined Magnetic and Mechanical Scanning

– Moving Magnet

– Moving Patient

• Etc.

6

Flanz 2009 - Erice

Simple form of Double Scattering

7

Flanz 2009 - Erice

Combined First Scatterer and Range Modulator

Usual reference to double scatteringReal birth of double scattering (Brahme)

What is Scattering? Proton Beam Properties

Dose and Dose Distribution

Transverse

Dose

Distribution

Depth Dose

Distribution

Uniformity

Penumbra

Distal Fall-off

Range

Dose, (Dose

Rate)

Width

Width

SAD

Current

Modulation

Errors for the most part

average out, or are easily

detectable.

SOBP Uniformity

• Tranverse

– Scatterers

– Single Scattering

– Double Scattering

• Longitudinal

– Range Modulator

– Ridge Filter

Flanz 2009 - Erice

Decades of Scanning - A New Generation

Scanning at Berkeley

Scanning at PSI

• 1978-9: Spot Scanning at NIRS - 30 Patients

– Range Modulator (Fast) + Lateral 2d Spot

• 1992 ish: B&W Scanning at BNL

• mid 1990‟s: Spot Scanning at PSI

• mid 1990‟s: Scanning at GSI

• 2008: Scanning at MDA (with Hitachi)

• 2008: Scanning at MGH (with IBA)Flanz 2009 - Erice

General Description of ScanningGeneral Description of Scanning: Pictorially, figure 1 describes the scanning process.

A beam at position A, at coordinate XA, is characterized by its current IA and it‟s beam size A. The beam

deposits a dose DA at location A. After that dose is deposited at location A, taking a time tA, the beam is

moved to location B. The time it takes to move from location A to location B is tAB. The beam current

during that movement is IAB. The velocity that the beam moves from position A to B is vAB = (xB-xA)/tAB,

and the current change between A and B is dI/dt = (IB-IA)/tAB. In this way we have defined all the terms

that are necessary in the delivery of beam scanning.

A

B

IA

IB

XB

XA

“Spot”

Time Driven

Dose Driven

Beam off/on

between spots

“Raster/Line/Segment”?

Digital World

Flanz 2009 - Erice

Time Structure in Pencil Beam Scanning

“Dose Driven” Scanning

Current

Dose

Position

Current

Dose

Position

Continuous

Pulsed

Continuous

Pulsed

Continuous Stable/Unstable. Pulsed Short or Long

“Time Driven” Scanning

PSIPSI

Flanz 2009 - Erice

More on Time; Not enough Time?• What if the beam is continuous?

– No change from previous slide

• What if the beam is pulsed?

– Example: 6cm x 6cm x 6cm = 0.21 liters

– Example: 8cm x 8cm x 8cm = 0.5 liters

– Beam Size 3.3mm

– Bragg peak width 3.3mm

– Assume Dose Driven Spot Scanning

– #Spots = 6000, 14250

– @ 30 pps 200 seconds = 3.3 minutes (Once)

– @ 30 pps 500 seconds = 8 minutes (Once)

Flanz 2009 - Erice

RadioBiology Time Effects? (sec, nsec, psec)e.g. Scanning with E then x, or x then E, Short pulses vs. Rf structure

Methods of Dose Driven Scanning: DSS and DET

Not for Distribution Flanz 2009

DET Spot Locations (~20)

- For DET multiple

directions or arc

therapy and

intensity

modulation

required to obtain

uniform dose

distributions.

Distal Edge of Target

Mackey

SS Spot Locations (~300)

Flanz 2009 - Erice

Scanning Proton Beam Properties

Dose and Dose Distribution

Transverse

Dose

Distribution

Depth Dose

Distribution

Penumbra

Distal Fall-off

Range

Dose, (Dose Rate),

Weight, Gradient

Size

Shape

Conformity/

Gradient

SAD• Generalized Scanning

• Uniform Scanning

• Wobbling

Errors are less „intuitive‟ and

may not average out. (Harder

to measure?) Need a basis for

estimating the effects of

errors and tolerances !Flanz 2009 - Erice

Evolution of Conformation

Tumor

Treated

Volume

Organ

Target Volume

Collimator

"Classical" Conformation

Treated

Volume

Tumor

Organ

Target

Volume

Intensity Modulation

Photon

Conformal:

Multiple

fields up to

dynamic arc

vs.

Charged

Particle:

Single or few

fields?

Scanning

SOBP (Scattering)

Flanz 2009 - Erice

What are the some features of

Proton Beam Scanning?

• Allow a non-uniform Dose Delivery

• Reduce # fields for a uniform Dose Delivery (Clinical effects?).

• Reduce unwanted Dose (e.g. Proximal Primary Dose)

• Reduce the need for Patient Specific Equipment

– Apertures

– Compensators

• Reduce radiation from primary beam intercepting machine components (n)

Therefore “Fluence Modulation” is

required for optimized Uniform Dose !

Flanz 2009 - Erice

Example

without a

Compensator

Reference 3D (almost) Cube Plan• Cube of 12(50%)x12(50%) x10 cm

(14 layers), 2.7 Gy

• Planned with ASTROID

– Penumbra Optimization (ala

PSI/Berkeley)

– This results in a balance between

penumbra and overall uniformity.

(There will be ears.)

– TPS provides the map which must be

compared with the measurement

Therefore “Fluence Modulation” is

required for optimized Uniform Dose !Flanz 2009 - Erice

Pedroni et. al.

Proton Beam Scanning (PBS) is NOT IMRT

• IMRT requires multiple fields of non-uniform dose delivery to

obtain a conformal uniform dose distribution.

• Particle Beam Scanning, can create a highly conformal dose

distribution with ONE field, but layers or pencils (DET) of non-

uniform field are required.

– Dose Modulation can be required for uniform field – NOT IMRT

• Non-uniform overall (single field) dose distributions using

proton beam scanning can be delivered like IMRT if needed, but

in general fewer fields are needed.

– Fewer fields „are‟ needed for highly conformal distributions – NOT IMRT

• Using the Term IMPT does NOT convey the power of PBS

– Power = Efficiency, conformality, speed, etc.

Flanz 2009 - Erice

First Patient Treatment – 4 liter Sarcoma

Sarcoma, layer 5

Sarcoma, layer 2

Photon Proton

Scanned

Beam

Flanz 2009 - Erice

Full

Irradiation

from ONE

direction

and no PSH

Another way Scanning can help?With scanning, Geometry becomes less

intuitive, but more flexibleAngle of entry

does not have to be

as constrained as

with Scattered

beams, and can still

have a „square‟

edge and uniform

dose. (Proximal

dose is shifted.)

- Matching

- SAD may not be

as much of an issue

Flanz 2009 - Erice

Normal Delivery

Delivery from

another angle

Proximal Dose issue?:

Alternate Fields

Do we need a Gantry?

Direction – What‟s needed to control

the trajectory of the Particle Beam• What beam angles are needed?

• Gantry

– Full Gantry

– Partial Gantry

• Fixed Beamline(s)

– Single

– Multiple Beams

• Reisenrad or equivalent

• Patient Orientation

• Later

Size of a Gantry, if needed

Enclosure

What drives the size of the

enclosure

Space for the patient

Other equipment

Optics

Beam Gantry Imaging PPS Immobilization TargetFlanz 2009 - Erice

Max roll or

tilt = 15 ˚

Horizontal Beam

Max roll or

tilt = 15 ˚

Pure Vertical Beam

30 Degrees off Vertical Beam

30 Degrees + Horizontal Beam

Niek Schreuder

ProCure

BPTC Gantry Angle Summary

Over 60% of all Treatment

fields were delivered

within +/- 10 degrees of

cardinal angles.

What sites can be

treated with this?

No Time/Postdocs

Flanz 2009 - Erice

What if you didn‟t have all 360 degrees?

Two Fixed Beam Concept Drawings

Couch or

Chair !

Angle TBD?

Flanz 2009 - Erice

Two Fixed Beam Concept Drawings

Inclined beam from BELOW

Couch or

Chair !

Angle TBD?

Flanz 2009 - Erice

A perspective on Proton Therapy

TechnologyWhere did we come from?

Where to go from here?

Does anyone have a map or menu?

PTCOG XXXXI

MPRI / IUCF

Jay Flanz

10/11/04

Another Idea (Jay) – PTCOG 41

Complexity vs.

Simplicity

Another Idea for

MGH

Flanz 2009 - Erice

Beam angles DEPEND on patient orientation

• Lying, Sitting, Standing

• Reproducibility (Day to Day?)

• How long can position be

maintained (sec, min??)

• Knowledge of existing

anatomy/tumor location and

shape and required treatment

plan du jour

• Imaging, True adaptive planning

Flanz 2009 - Erice

Siemens ProCureIBA, Hitachi

The Evolution of the Medical Accelerator

Flanz 2009 - Erice

Gantries• Photon Gantries

– Needed for conformality

• Proton Gantries

– Needed for conformality for Scattered Beam Delivery

– The Gantries and Beam Delivery are the highest costs,

NOT the accelerator

– Development of lower cost accelerator is usually

associated with mounting it on a Gantry

• Implies Single room system

– Either Design a Lower cost Gantry, or

– Design OUT the need for them

• Imaging

• Adaptive Planning + Scanning

• These will take a while so in the meantime… Flanz 2009 - Erice

Scaling DOWN Machines - Quantitatively

Smaller / Cheaper

LLNL

ProTom

30

Varian

Flanz 2009 - Erice

http://www.tomotherapy.com/

If we need a Gantry, what are

some of the considerations?

In-plane Gantry Color Choices

Gantries for Scattered Beams have dominated (Scatter after Bend)!

Scanning Gantry Geometries

Using Two or One Combined Dipole e.g. IUCF

Flanz 2009 - Erice

Scanning Magnets, Gantry Optics – New Opportunities

Minimizes the

Dipole Size

Innovative Gantry DesignSee Dejan‟s talk

Some Gantry Design Considerations (If you really need/want one)

Flanz 2009 - Erice

Beam Size Rqmt or Penumbra ?

35

0

0.2

0.4

0.6

0.8

1

1.2

0 1 2 3 4 5 6

Sigma Things:

FWHM/Sigma = 2.35

80-20/Sigma = 1.13

Not 1.6

90-50/Sigma = 0.72Therefore, if we have a

5mm spacing, we need a

sigma of 7mm

(FWHM=16.4mm)

(10%-10%=29.9mm).

2

2

1

x

ey

Flanz 2009 - Erice

Organ

at Risk

Target

Assume: Dose to Target within +/-

2.5% and Organ at Risk < 50%

Beam Sigma (why?) Beam Edge (why not?)

• Use Is a Sharp edge needed??

– For cases of penetration less than 5cm.

• Is a small beam needed ???

– when absolutely necessary to pass BETWEEN

two critical structures. (What is the percentage of

these cases?) Then one needs small sigma.

• Is a sharp edge bad??

• When abutting beams (spot scanning).

• How to achieve sharp edge?

– Small sigma

– Use the Distal Edge ???

– Modified MLC

– ANOTHER WAY

Flanz PTCOG 44

Tolerance:

10% variation in sigma --> 2% VARIATION IN DOSE !

Flanz 2009 - Erice

How can one achieve a sharp edge beam

Without a collimator right at the patient?

• With a selectable effective drift one can also control the „penumbra‟ which could be useful in matching.

Upstream CollimatorGaussian

Beam

Focussing Elements

Rectangular Beam

PTCOG 36, Catania J. Flanz What Questions should we

be asking about Scanning Systems.

Flanz 2009 - Erice

PSI Gantry 2

• The Ultimate PBS Device?

– Infinite SAD

• (Adjustable?)

– EDGE CONTROL

– Not 360 degrees

– Not Cheap!Proscan 2

Flanz 2009 - Erice

New Opportunities REALIZED?

Beam Size, 3 m drift, From Gantry to Isocenter

Emit = 18mm mrad, sigma = 3mm Emit = 18, sigma = 9mm

Emit = 5, sigma = 3mm Emit = 5, sigma = 9mm

ISO

CE

NT

ER

ISO

CE

NT

ER

ISO

CE

NT

ER

ISO

CE

NT

ER

Typical Cyclotron Degraded Beam

X2DIP ~ X2

ISO + L2*2ISO

Flanz 2009 - Erice

Effect of Beam Size on Gantry Dipole and PowerGantry Dipole Power / Weight

0

2

4

6

8

10

12

0 5 10 151 sigma Isocenter Beam Size

Dip

ole

Po

wer/

Weig

ht

PS Current ~ Gap (2.1 sigma)

Magnet weight ~ gap 2

Power ~ Current 2 ~ gap 2

gap ~ 1/final beam size ?

Smaller Emittance

Larger Emittance

For Larger Emittance Beams, the power and weight requirements

increase a factor of 3 when reducing the beam size from 6mm to 3mm

!!Even smaller optical beam size is necessary when considering

scatter from MATERIAL IN THE BEAM PATH (Instruments,

Windows (Acoustic Accident?), Gas, Range shifter/Ridge filter. Small Beam at Isocenter --> Larger magnet aperture --> bigger magnet --> More power --> bigger gantry --> Higher cost

Flanz 2009 - Erice

No time to discuss

• Hypofractionation (Conformality or

Radiobiology?)

• Organ Motion

– Gating

– Tracking (Position AND Range)

• Instrumentation (What is needed?)

• Imaging (What is needed?)

Flanz 2009 - Erice

Conclusions? Postulates?• With Scanning it MAY be possible to consider reducing the number of beam

directions and Gantries may be less important for a large fraction of sites.

• Scanning can bring with it a new evolution in treatment efficiency

• A new evolution in cost reduction (Gantries…) will be achieved with imaging and

adaptive Treatment Planning /Scanning, (in the future).

• Along with the above; A new evolution in optimization complexity and Treatment

complexity wrt organ motion is achieved with Scanning

• Clinical Requirements for scanning have not really been established in any

systematic way. Technical Implementation can be difficult and expensive depending

upon the clinical input

• Scanning Nomenclature is misleading and can cause implementation „mistakes‟ at

worst, or reduced applications, at best (Small beams).

• IMPT is not a good term for Proton Scanning unless it is IMproved Proton Therapy

• For Larger emittance Systems, beam size below 5mm starts to cost “a lot”: extra

magnet size, weight (e.g. gantry weight) &PS, reduced instrumentation, windows?.

Energy Change speed increases the costs of the Power Supplies

• Concept of „changing energies‟, power, water to be revisited!! (See Dejan‟s Talk

Trap to Avoid

• Beam parameters will DRIVE beam

delivery modality.

– “Tell me what the machine can do so I can

simulate the modality.”

• OR Vice- Versa

– “Do simulations to help optimize treatments

and tell me what the machine should do”

Flanz 2009 - Erice

44

Flanz 2009 - Erice

44

The Francis H. Burr Proton Therapy Center

Thank You !

beam delivery systems: scattering, scanning, w/wo - erice 2009

Documents