paul scherrer institut future directions and current challenges of proton therapy tony lomax,...

PAUL SCHERRER INSTITUT

Future directions and current challenges of proton therapy

Tony Lomax, Oxford, 2008


Tony Lomax, Centre for Proton Radiotherapy, Paul Scherrer Institute, Switzerland

CT after 5 weeksPlanning CT




Overview of presentation

1. State-of-the-art proton delivery

2. Current challenges

3. New directions in proton therapy

4. Summary




State-of-the-art proton delivery

Passive scattering

Range-shifter wheel

CollimatorCompensator

Target

Patient

Scatterer




Spot scanning

Pedroni et al, Med Phys. 22:37-53, 1995

Target

Magnetic scanner

‘Range shifter’ plate

Patient

Proton pencil beam





The present and future of proton therapy?

Passive scattering (patched fields)

Collimators and compensators required for each field

Intensity Modulated Proton Therapy (IMPT)

Fully automated delivery (scanning)





Sarcoma – 12 year old boy

Delivered single field plan 9 field IMRT plan

Factor 6 lower integral dose for protons





Integral dose and secondary cancer risk


K Haeller, PSI

N Tarbell, ASTRO, 2008

Study comparing 503 proton patients to 1591 photon patients (treated 1974-2001) showed a

50% reduction in secondary tumours in

proton patients








4. Summary




The advantage of protons is that they stop.

Range uncertainty

The disadvantage of protons is that we don’t always know where…

10% range error

Current challenges: range uncertainty




S. Mori, G. Chen, MGH, Boston


Tumour shrinkage

Initial Planning CTGTV 115 cc

5 weeks laterGTV 39 cc




Planning CT CT after 5 weeks

Beam stops at distal edge Beam overshoot

S. Mori, G. Chen, MGH, Boston

Tumour shrinkage





Patient weight changes3 field IMPT plan to an 8 year old

boy

Note, sparing of spinal cord in middle of PTV

Planning CT New CT

During treatment, 1.5kg weight gain was observed

Max range differences:SC 0.8cm

CTV 1.5cmFrancesca Albertini and Alessandra Bolsi (PSI)






CT artefacts

Many patients referred for RT post-operatively and with

metal (titanium) stabilisation

How accurately can we calculate proton ranges in

such CT data sets?




Months

0

.2

.4

.6

.8

1

0 10 20 30 40 50 60

Without implant (n = 13)

Implants (n =13)

3-yr, 5-yr 100%

3-yr 69%5-yr 46%

Chordomas and chondrosarcomas of the spinal axis

Rutz et al 2007, IJROBP, 67, 512-520





Martin von Siebenthal, Phillipe Cattin, Gabor Szekely, Tony Lomax, ETH, Zurich and PSI, Villigen

What is the effect of organ motion on proton therapy?

4D-CT derived from 4D-MRI

Current challenges: organ motion

Organ motion




Organ motion and passive scattering


Images courtesy of Thomas Bortfeld, MGH, Boston

Parallel opposed photons

Single field photons

Single field protons




Martin von Siebenthal, Phillipe Cattin, Gabor Szekely, Tony Lomax, ETH, Zurich and PSI, Villigen

A scanned beam in a moving

patient.

4D-CT derived from 4D-MRI


Organ motion and scanning




Nominal (static) doseCalculated with ‘real’ motion

from 4D-MRI of volunteer

Organ motion and the ‘interplay’ effect





Dose (%)

Vol

um

e (%

)

70 80 90 100 110 1200

20

40

60

80

100

Dose (%)V

olu

me

(%)

70 80 90 100 110 1200

20

40

60

80

100

Motion patient 1

Amplitude ~ 11mm

Motion patient 2

Amplitude ~ 8mm

Organ motion and the ‘interplay’ effect


Scanning is particularly sensitive to organ motion




Adenoid cystic carcinoma with node involvement

Current challenges: treatment gantries

Tricky to do without a gantry…





Avoiding density heterogeneities

% points with dDRR > 1cm

% d

ose

poin

ts w

ith M

C-R

C d

ose

<+

\-2%

0 20 40 60 8050

60

70

80

90

100


% d

ose

poin

ts w

ith M

C-R

C d

ose

<+

\-2%


% d

ose

poin

ts w

ith M

C-R

C d

ose

<+

\-2%


% d

ose

poin

ts w

ith M

C-R

C d

ose

<+

\-2%


% d

ose

poin

ts w

ith M

C-R

C d

ose

<+

\-2%

0 20 40 60 8050

60

70

80

90

100

Heterogeneity index

Each point is a different field calculated for the same skull base case

A comparison of MC and

analytical dose calculations as a

function of density

heterogeneities




0

5000

10000

15000

20000

25000

-20 -15 -10 -5 0 5 10 15 20

Dose difference

Fre

qu

ency

(n

um

ber

of

Vo

xels

) Applied plan

'Standard' plan

Field

Applied plan ‘Standard’ plan

Gantry angle

Table angle

Density heterogeneit

y index

Gantry angle

Table angle

Density heterogeneit

y index

1 -45 -90 20.4 -90 -120 28.2

2 -10 0 12.9 -90 -60 30.2

3 -120 -120 12.7 60 0 26.2

Avoiding density heterogeneities






3.5m diameter

PSI gantry 1

7m diameter

PSI gantry 2

12m diameter

Loma Linda

15m diameter

Heidelberg

Could this be the main limit to

the spread of particle

therapy?








4. Summary




New directions in proton therapy

1. Possible improvements to passive scattering

2. Dealing with range uncertainties

3. Organ motion and scanning







3. Gantry design




Imaging for range

Ospedale San Rafaele, MilanFrancesca Albertini, PSI

kV-CT

MV-CT (Hi-Art)

MV-CTProton radiograph

Proton DRR

Proton radiography

Uwe Schneider, ZurichAlexander Tourovsky, PSI

Measured PET activation

Calculated PET activation

Activation PET

Katia Parodi, Thomas BortfeldMGH, Boston

Dealing with range uncertainties




Robust planning techniques

Tumour

Spinal cord

Lomax et al.:Med. Phys. 28(3): 317-324, 2001

Example paraspinal case

Nominal plan

10% overshoot





3 patched, intensitymodulated fields....

...give a homogenousdose without the useof fields that abutdistally against thespinal cord

Robust planning techniques

Lomax et al.:Med. Phys. 28(3): 317-324, 2001





10% overshoot plans

IMPT

Single field

IMPT

Nominal plans

Single field

Single field

IMPT plan

Nominal

Overshoot

Nominal

Overshoot

Relative dose

Rel

ativ

e vo

lum

e

Dtol

DVH analysis

Spinal cord

Robust planning techniquesDealing with range uncertainties




Range adapted proton therapy


Alessandra Bolsi, PSI




Work of Alessandra Bolsi.Dealing with range uncertainties






CTV

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100 110 120

Dose (%)

Vo

lum

e (

%) Nominal CTV

Weight increase CTV

Range adapted CTV

Nominal SC

Weight increase SC

Range adapted SC










3. Gantry design




Repaint scanned beam many times such that statistics dictate coverage and homogeneity of dose in target (c.f. fractionation)


Rescanning




Rescanning

4s period

Re-scanning in presence of Cos4 motion with 1cm amplitude

• Cylindrical target volume

• Re-scanned different times to same total dose

• Scan times calculated for realistic beam intensities and dead times between spots

• Analysis carried out for different periods of motion

Marco Schwarz, Silvan Zenklusen ATREP and PSI

Not always improving homogeneity with

number of re-scans!





RescanningThe ‘synchronicity’ effect

• Very preliminary results

• A ‘real’ effect for perfectly regular breathing?

• Could well be less of an issue when breathing is more irregular

• For regular breathing, could be avoided by selecting the re-scanning period to avoid effect or varying period scan-to-scan

• Probably not a big issue in reality

See presentation from Silvan Zenklusen, Saturday





Track motion of tumour using scanning system

based on some anatomical/physiological

signal

Tracking





Steven van de Water, PSI/TUDelft

Tumour Tracking

Plot of dose homogeneity as function of RMS

position error due to motion and

‘imperfect’ tracking

Cos4 motion with varying detection

delays and tracking

accuracies

~150ms delay

Vedam et al 2004





Steven van de Water, PSI/TUDelft

Tumour Tracking

Re-tracking – tracking the

tumour repeatedly within one fraction

E.g. 4 times

~150ms delay

Vedam et al 2004








3. Gantry design




Still rivers systems single room proton facility

Gantry design

DWA based proton tomotherapy

Probably the biggest step forward for particle therapy

would come from a significant reduction in gantry size

To stay competitive with other therapies for the next 25 years, treatment gantries will almost

certainly be necessary




Summary

• Although passive scattering is still the preferred choice for proton therapy, scanning and IMPT will become more widespread in the next years (c.f. MD Anderson)

• To what extent can scattering be improved through the use of automated field hardware (MLC’s etc)?

• Range uncertainty and organ motion (particularly for scanning) remain the main challenges to proton therapy and much interesting and exciting work is still to be done in organ management, range imaging and adaptive proton therapy

• The field is ripe for new input, ideas and innovations…

paul scherrer institut future directions and current challenges of proton therapy tony lomax,...

Documents

future of proton therapy

proton patients

proton therapy impt

proton radiotherapy

boston current challenges

range error current

paul scherrer institute

new directions