Technological innovations - CERN Documents... · E. Pedroni Center for Proton Radiation Therapy -...

E. Pedroni Center for Proton Radiation Therapy - Paul Scherrer Institute - Proton Therapy Winter School 13 -01-2010 1

PSI13.01.2010

Technological innovations

Eros Pedroni

Paul Scherrer InstituteSWITZERLAND

D. Meer, C. Bula,S, Safai, S. Zenklusen

R. Kobler, S.Lin, …

E. Pedroni Center for Proton Radiation Therapy - Paul Scherrer Institute - Proton Therapy Winter School 13-01-2010 2

1. Gantry layout 2. Fast scanning

3. Organ motion solutions 4. Competition or complement? ion therapy

5. Proton Radiography

6. Questions

Summary


1. Gantry layout


Medical cyclotron

Gantry 2

Degrader

Collimator

•

Combined and optimized use

of–

Cyclotron –

degrader –

beam line

–

gantry

–

nozzle –

patient handling

–

All contributing together to the best possible (dynamic) beam delivery

Integration of the whole system for optimized beam delivery


Easy access to the patient

•

Gantry rotation limited to -30°

to + 180°

(instead of 360°)•

Flexibility of beam delivery by rotating the table in the horizontal plane

•

Analogy with longitude and latitude in the world-geography

•

Expected advantages:•

Fixed walls

for mounting supervision

equipment -

like Vision-RT•

Permanent fixed floor

for a better

access to the patient table•

Nobody (patient or personnel) falling in the gantry pit

PSIStill River


In-room positioning with sliding-CT

•

Within reach of the patient table–

Sliding CT of Siemens

–

Use of time-resolved images before (and after) treatment

•

Adapt dose field to the organ situation of the day (the body regions with soft tissues)

•

Setup of respiration gating

•

The first proton gantry of the world coupled with a sliding CT

•

Adaptive system to new diagnostic equipment

–

CT-PET? MRI?


BEV X-rays

•

Retractable support with a flat panel•

X-ray tube on beam axis

–

Shining through a hole in the yoke

of the 90°

bending magnet•

Simultaneous use of x-rays during proton beam delivery

For controlling the position of moving targets during beam delivery

Image guided Proton Therapy?


Option to use BEV X-ray -

simultaneous to proton beam?

•

Equivalent to portal imaging

with KV photon•

Large field-of-view area (26 cm x 16 cm) –

not masked by equipment or collimators

in the beam path

•

For QA control of gating and tracking (scanning + pulsed X-rays)•

Fluoroscopy?

•

Neutron damage to panel?

a) compact gantry b) long throw gantry

SweepersX rays tube

Proton beam

Bendingmagnet

nozzle

Yoke hole

PatientImager

Sweeper

or Scatterer

Collimator

Upstream scanningPSI

Downstream scanning

IBA Hitachi

Varian…


•

Vacuum “up to the patient”–

Sharp pencil beam -

3 mm sigma

•

Two monitors and a strip monitor–

2 mm strips (delivered by TERA collaboration)

•

Removable pre-absorber–

IN and OUT of beam (motorized)

–

For ranges below 4 ? -10 ? cm•

Telescopic motion of the nozzle–

To reduce air gap

(keep patient at isocenter)

•

Option to add collimator and compensators–

To shield OAR on top of scanning

–

To simulate passive scattering with a scanning beam•

Collision protection to treat patients remotely (multiple fields in one go)–

Field extension with patient table

Optimized nozzle design


Nozzle: preliminary results –

small size of pencil beam

•

Use of minimal material in the nozzle for keeping the beam size between 3 and 4 mm sigma at “all”

energies

•

The quality of the beam depends from the design of the beamline

and of the nozzle

Measured while removing piece by piecethe materials in the nozzle

mm


2. Advancing the scanning technology


Very fast changes of the beam energy

•

Continuous choice of the beam energy–

Setting all elements of the whole beam line within a single command

•

Constant beam transmission from COMET to the gantry–

“Compensation”

of degrader losses from 100 to 200 MeV

•

Fast energy changes–

Cyclotron (fixed energy)

–

Fast degrader ahead of the gantry

–

The beam line follows the energy variations in the degrader

•

Shown–

80 ms

dead time

for range steps of 5 mm

Dynamic beam energy

A factor of 10 better than other systems


Double parallel scanning

T U

•

Fast

parallel lateral scanning–

T sweeper

2 cm/ms

–

U sweeper

0.5 cm/ms•

Scan area of 12 cm by 20 cm–

Motion of patient table

for treating larger field sizes Table used as sweeper-offset

•

Experience with Gantry 1•

Apparent source at the infinity–

Simplify treatment planning

–

Easy field patching with table–

Avoid errors from

compensators

–

Simplify verification

dosimetry


100 MeV

120 MeV

140 MeV

160 MeV

180 MeV

200 MeV

Well focused beam for all energies…

•

Parallelism–

Max deviation ~4 mrad

(at edge of field)

Achieved by connecting a shortcorrection quadrupole

in series

with the U sweeper


Vertical deflector plate (COMET)

•

Dynamic use of the modulation of the beam intensity–

Deflector plate and vertical collimators in the first beam turn after the ion source

–

Time delay to extracted beam in the order of 100 us•

Examples–

Delivery of line segments with changing voltages on the deflector plate

–

“Pulsing beam”


Flexible control system

•

Steering file for combined delivery of–

Spots

•

Spot scanning as the default (starting) mode–

Lines

•

For maximum repainting number and simulated scattering –

Contours?

•

For optimizing repainting and lateral fall-off (difference Gaussian to error-function) •

Including the dose control for the passive scattering system of Optis 2


Tabulated dose delivery with FPGA

•

Combined tabulated control of –

U-sweeper

–

T-sweeper–

Beam intensity

•

As a function of timeT

U

Example 2

-

Dose box with continuous scanning 494 energy layers

(85 ms per layer)

(6 x 8 cm) in less than 1 minute

Example 1

–

U T meander path


Dose control of time driven lines scan

•

Time driven devices U -

T -

I–

Dose control with feed-back loop Input: difference of Monitor 1

to required dose rate Output vertical deflector plate voltage

•

Paths with variable speed and/or variable intensity

0.2 0.7 10

500

1000

1500

2000

Required Dose

Deliv

ered

MUs

Dose linearity of simple T-lines

10 ms

23 timesMax T speedVariable intensity

1 time *0.5Constant intensityVariable T speed

10 cm in20-200 ms

10 cm in 5 ms

5 ms is the limit …

systematic errors !!!


Conformal volumetric repainting

•

Painting of lines

(contours?)–

With maximal possible velocity ~ 2 cm / ms, 0.5 cm/ms

–

Dose shaping through Beam Intensity Modulation

(BIM)•

Painting of energy iso-layers –

< 200 ms per plane

(20 lines x 5 mm)

–

Change of energy (100 ms

-

5mm range)•

Repainting of iso-layers–

~ 6 s per liter

(20 energies at 5mm steps)

•

Volumetric repainting capability (aiming at)–

10 repaintings

/ liter in 1 or 2

minutes

–

Analogy with a TELEVISION tube


3. Coping with the organ motion problem (of scanning)


(Uncontrolled) motions less than 5 mm –

repainting

methods

• Scaled repainting•

Repeat same scan n times with dose divided by n

• Layered repainting•

Deliver partial scans with constant max. partial dose

•

Revisit only those spots which need more dose

• Volumetric repainting•

Repeat volumes

•

Requires fast energy changes

• Non volumetric repainting•

Repeat only iso-energy layers

beam

0 % missing

dose25 % missing

dose

75 % missing

dose50 % missing

dose

100 % missing

dose


How many repaintings

do we need? Analysis of repainting strategies -

thesis work of S. Zenklusen –

(on press)

Full parameter space:• Frequency [1/min]: 2 -

100

• Start phase: 0 -

2π• tmax

[ms]: 1, 2, 3, 4, 5, 7,10, 20, 1000 → 2700 dose calculations

→ on PSI Linux ClusterStandard error reported for each dose distribution as difference to ∞-

repainting within the targetMotion amplitude 5 mmVolume 0.5 liter

G1 –

volumetric –

scaled repainingG2 –

volumetric –

scaled repainting

G2 –

volumetric –

iso-layer repaintingG2 –

volumetric –

lines

Volumetric repainting is important for the motion mitigation of scanning beams

Breast, pancreas, cervix, …

15 repainting in 1.5 minutes


Respiration motion (> 5mm)

– Does gating

need repainting?

•

Only if residual motion error is small enough (we don’t known yet)

–

Which gating signal should we use?•

External signal (like thoracic strip ?)

•

3-d stereo viewing of external marks on patient skin?•

A whole energy layer delivered within a gate?

•

Needs fast iso-energy layer painting (50-200 ms per plane)–

Synchronize energy change to next gate

•

No need for fast energy changes•

But treatment time can be very high

Plan

1

Beam

Target at times ti

Iso-energylayers E at ti

E


Respiration motion-

is tracking

a viable solution?

Nodule

Rib

Lung

•

Lateral corrections

as a function of the target displacement (easy)•

Energy corrections as a function of the target displacement (Range shifter)–

Density heterogeneities in the beam path (example ribs in front of a lung tumor)

•

Deformation model of the target (complex delivery and treatment planning)•

Energy corrections -> Slow painting -> Probably only one single painting

•

Residual error precision? Are 1-2 mm required?–

Is the residual error of detecting motion small enough to avoid repainting?

•

Most of the expertise in this direction is at GSI (C. Bert) Beam


Why not then “Multi-gating”? –

synchronized plans delivery •

Associate several time-gates

with corresponding gated-CT plans

–

Deliver multiple “time-instances-treatment plans”

synchronized with breathing•

Energy layer of a plan is delivered within its own gate (50-100 ms/plane)–

A better alternative to tracking?

•

No need for energy corrections -

No target motion-deformation model needed •

Repainted by definition –

Fast scanning -

Good duty factor of using of the beam

t1 t2

Plan

1

Plan

3

Plan

2

change to the next energy when all iso-energy plans are completed

Beam

Target at times ti

Iso-energylayers E at ti

E

t3


•

Small volume (<1/4 liter) –

Full volume treated within 4s with intensity modulated line scanning

–

Needs fast energy changes–

High dose rate

could be of advantage for reducing treatment time

•

Especially in the context of a hypo-fractionation•

A breath hold length of 5 s would then be sufficient–

Repainted, high efficiency, static planning, easier QA

Respiration motion (> 5mm)

Our preferred solution Whole painting within a breath hold cycle

Beam

Wholetarget

Breath hold


Part 4. Competition with carbon ions


The rationale of using carbon ions

•

Arguments in favor of carbon–

Correct -

The high LET of heavy ions could be of advantage for treating

•

Radio-resistant tumors•

Poorly oxygenated tumors

(OER oxygen enhancement ratio)

–

Correct –

Better precision of the beam -

Less MCS -

Sharper Bragg peaks –•

But with fractionation tails –

Large dose uncertainties due to the varying RBE

–

Simplistic -

Carbon is better because is “more effective”

(higher RBE) •

What matters is the RBE difference of healthy to tumor cells

–

the therapeutic ratio

–

Possibly wrong -

Carbon therapy “needs”

only few fractions –

is therefore less expensive•

Protons can also be delivered with low fractionation (small tumors –

radio-surgery)

•

At a high dose per fraction protons behave similar to high-LET (example OPTIS) •

Arguments against carbon–

High LET can be a contra-indication

(for example for pediatric treatments)

•

Absence of repair

of the healthy tissues

(tolerance within the target)•

Risk of late effects

(not convincing experience with neutrons and pions)

–

Assessment of the dose is difficult -

needs an own radiobiology research group


Carbon + proton

facility: a new standard for the hospitals?

70m

37m

Carbon

ions

Protons

The magnetic rigidity of carbon is 3 times higher than with protonsThe dimensions and costs of the facility

are > doubled size x 23

Huge gantry or no gantry at all

C+P? or hypofractionated

P?The answer will come in the future


5. Other DEVELOPMENT topics”

(QA related)


•

Image of the range

loss through the patient in-vivo

•

Poor image contrast due to MCS•

QA tool for range control

•

Predict proton stop–

(PSI development work in the 90s)

x,y x,y

Range

Therapy

Radiography

PR of Rando

phantom

E1E2

Proton radiography (tomography)

The key for a new paradigm of “working with the precision in range”Distal fall-off is potentially sharper then lateral fall-off (by a factor of 2)

Stopping the beam in front of OAR


CONCLUSIONS

•

Proton therapy must remain competitive with conventional photon therapy–

IMPT

and scanning beams are needed to compete with IMRT

–

PT will have to follow the development towards image-guided RT–

Scanning is most optimal

•

for biological targeting

(intentional non-homogeneous dose distributions) •

and adaptive beam delivery

•

The experience of using scanning with organ motion

is still missing–

Fast volumetric repainting and volume within a breath-hold

could be an “easy”

solution of this problem

•

Initiatives to reduce the size and costs of the equipment (new accelerators)–

are welcome … but difficult to achieve …

at equal performance

•

Is high-LET radiation a true clinical need?–

More clinical research with ions needed

–

Superconducting cyclotron and gantries could help ion therapy


THANK YOU

Technological innovations - CERN Documents... · E. Pedroni Center for Proton Radiation Therapy -...

Documents

Transcript of Technological innovations - CERN Documents... · E. Pedroni Center for Proton Radiation Therapy -...