scanning beams - ptcog
TRANSCRIPT
2
Outline
Why Scanning Beams?
PBS and Clinical Planning. What matters?
-Minimum Energy/Beam Degraders/Spot Size
- Planning Techniques: SFUD, IMPT, DET
- Robust Optimization and Evaluation
- Interplay
Site Specific Implementation; Technical Protocols.
Summary.
3
Proton Therapy
Technologies Passive Scattering PBS
Techniques SOBP SFUD IMPT
3DCRT/IMRT IMRT/VMAT
4
A proton pencil
Beam Spot…...
A few pencil beams
together…. Some more…
A full set, with a homogenous dose
conformed distally & proximally.
Pencil Beam Scanning – PBS
Image by E.Pedroni
PSI, Switzerland
Magnetically scan p beam (X,Y) and control depth switching Energy (Z)
A target is divided into many layers, a layer is divided into many spots, Spots are scanned one by one; layers by layers, from distal to proximal.
5
Why Pencil Beam Scanning?
PBS: Lowest Integral dose(2 to 3 times vs XRT). PBS 3D HIGH DOSE Conformality (1- 4Fs )~ IMXT (4-9Fs).
PBS can treat large fields, comparable with XRT.
PBS penetrates at larger depths if no beam modifiers present.
PBS produces less neutrons contamination outside of the patient vs
PS has no beam modifiers are required(less nuclear interactions).
6
Pencil Beam Scanning Challenges
It is difficult to generate very narrow pencil beams which result in optimal dose lateral fall off.
PBS Minimum Energy (70 to 100 MeV).
PBS has a time structure, can be more sensitive to motion than PS.
The main difference from XRT: Range Uncertainty.
7
Outline
Why Scanning Beams?
PBS and Clinical Planning:
- Minimum Energy/Beam Degraders/Spot Size
- Planning Techniques: SFUD, IMPT, DET
- Robust Optimization and Evaluation
- Interplay
Site Specific Implementation; Technical Protocols.
Summary.
8
PBS and Clinical Planning
What matters?
- Minimum Energy
- Planning Techniques
- Shallow Target/ Beam Degraders.
Spot Size vs. Air Gap:
Small => High Conformality/Low HI
Large => Less sensitive to motion
- Plan Robustness:
Optimization
Evaluation
Motion Management/Interplay
Site specific implementation
9
Beam Degrader/Minimum Energy
Beam Degrader Options:
• Machine Specific: Range Shifter on the snout.
Problems: Size of Air GAP & Collision, Beam data commissioning
workload.
Solutions: Telescopic Arms Bolus, Collision Detection.
10
Beam Degrader/Minimum Energy
Beam Degrader Options:
• Patient Specific: Bolus, Thick table top, Bridges, etc.
Problems: Treatment Table Weight limits, Patient Positioning
&Collision, Impact on Imaging.
Solutions: Use of Universal Removable Devices, Automated
Collision Detection, homogeneous & less dense material.
PENN/ Qfix
11
Spot Size Integrity: Bolus vs. Range Shifter
120 140 160 180 200
5
10
15
20
25
30
35
Beam Energy (MeV)
Sp
ot
siz
e (
mm
)
X direction
No RS
2cm Bolus
8cm Bolus
RS
120 140 160 180 200
5
10
15
20
25
30
35
Beam Energy (MeV)
Y direction
No RS
2cm Bolus
8cm Bolus
RS
S. Both, J.Shen et al, IJROBP 2014 – in press
12
Beam Degraders Collision Detection
-CAD /MATLAB ray casting algorithm.
-Incorporated during the proton treatment planning phase=>improved QA
processes and clinical efficiency.
Green – Body contour points; Red – gantry polygon
W. Zou, S Both et al. Med Phys J. 2012
A
B
13
PBS Treatment Planning
A single field uniform dose (SFUD) technique, combines one
or more fields optimized individually (SFO) to deliver a
relative uniform dose across the target volume.
Note: SFUD is the PBS equivalent to open field treatment.
An intensity modulated proton therapy technique (IMPT),
combines fields with heterogeneous dose distributions
optimized together(MFO) to deliver a relatively
homogeneous dose across the target.
Note: IMPT is the PBS equivalent to patch fields passive
scattering and IMRT.
14
SFO/SFUD vs. MFO/IMPT Techniques
SFUD RT
IMPT LT
LT
RT
15
PBS Optimization
Treatment plans are optimized using inverse planning techniques which
allow for variation in intensity and energy of each pencil beam (increases
the degree of freedoms vs IMXT=>better dose shaping)
3D Forward planning PBS: Inverse planning
?
?
16
PBS Optimization
Pencil beam weights are optimized for each beam direction
using inverse planning techniques => beam weight maps for
different energy layers.
17
Degeneracy of solution in inverse planning
IMPT SFUD
In general, based on the input the optimization problem may pose
many equivalent solutions. It is difficult to decide whether a result
produced by a planning algorithm can be further improved (e.g.
adding more beams, reducing #of spots, etc.)
18
PBS Techniques 2.5D uses poli-energetic SOBP pencil
beams( different weights by different
colors) individually adapted distally
and proximally to TV=> uniform dose
per field.
3D uses poli-energetic pencil beams,
non-uniformly distributed and
adapted distally and proximally to TV
=> non-uniform dose per field.
DET uses pencil beams distributed
individually only on the distal TV
edge => high non uniform dose per
field.
A. Lomax, 1999 PMB
2.5D
SFUD
3D
IMPT
DET
PB
19
19
PBS Optimization: PTV ?
For protons in addition to lateral margins, a margin in depth has to be left to allow for range uncertainty, therefore each beam orientation would need a different PTV, beam specific PTV(bsPTV).
Generally, in practice, the dose distribution is determined based on a Optimization Volume(OV) derived from CTV adding lateral and range margins:
• DM = (0.035 x CTVdist) + 1 mm
• PM = (0.035 x CTVprox) + 1mm
• LM based on setup, motion, penumbra.
3.5%- uncertainty in the CT# and their conversion to relative proton linear stopping power
1 mm - added to correct for range uncertainty
Note: it works for PS, SFUD-not necessarily for patched fields and IMPT.
CTV DM PM
LM
PTV
OV Patient
WED
20
PBS Plan Robustness
Plan robustness measures the differences in quality between
planned and delivered dose in the presence of uncertainties( e.g.
setup and range uncertainties)
o Robust Plan Optimization
Account for uncertainties in the optimization function
o Robust Plan Evaluation
“Worst case scenario”: Standard robustness test
For example: systematically simulate setup error by shifting
isocenter in 6 directions, introduce HU # variation
M. Engelsman, M, Schwartz, L.Dong, SRO 2013
21
PBS Beam Angle Optimization
Choosing the right beam
orientation
Choosing the best
geometry of irradiation
Shortest and most
homogeneous path
Ex: Pelvis (3 to 9 o'clock , CW)
Limited numbers of beams,
avoid serial OARs
Coplanar vs. non-coplanar
beams.
22
SFUD Optimization: Beam Specific PTV( bsPTV)
LM
CTV BEAM
Ray Tracing/ WED
DM
Physical
Margins
corrected for
SETUP&Range
based on local
heterogenities
P.Park, L.Dong et al, IJROBP 2012
•Avoids geometrical miss of the CTV through lateral expansion.
•DM, PM margins calculated from the target in beam direction for each Ray
•Extra margins based on local heterogeneities, using a density correction kernel
23
SFUD Optimisation: bsPTV
Dose distributions conforming dose to CTV using PTV(>30%)&bsPTV(<5%).
CTV CTV
CTV CTV
P.Park, L.Dong,et al IJROBP 2012
24
PBS Robust Optimization
Probabilistic Optimization:
The range of each pencil beam is
a random variable, the quantity to
be optimized is the residual cost
over the possible setup errors and
range uncertainties weighted
based on their probability.
No overshoot 5mm overshoot
Unkelbach et al, Med Phys 2009
25
Multi-Criteria Optimization (MCO)
In MCO:
- a database of plans each emphasizing
different treatment planning objectives,
is pre-computed to approximate the
Pareto surface.
- robustness can be integrated by
adding robustified objectives and
constraints to the MCO problem.
• Minimal set of absolute constraints
– D(GTV) > 50 Gy(RBE)
• Specify competing objectives –
– “minimize max OAR” vs “maximize min
GTV dose”
H.Kooy, MGH
Chen et al, PMB 2012
26
HN(PO vs. PO+AP): Robust Evaluation
27
HN :PBS vs. RA
LPO+RPO RA AP+LPO/RPO)
Init
Verify
Init Init
Verify Verify
28
HN Target Robustness:CTV54Gy(RBE)/CTV60Gy(RBE)
0 20 40 60 80 100 1200
20
40
60
80
100
Ratio o
f T
ota
l S
tructu
re V
olu
me [%
]
Ratio of Total Dose [%]
Verification study for CTV 54Gy(RBE), PT EE
Initial plan
Shift plans
Verification plans
0 20 40 60 80 100 120 1400
20
40
60
80
100
Ratio o
f T
ota
l S
tructu
re V
olu
me [%
]
Ratio of Total Dose [%]
Verification study for CTV 54Gy(RBE) with ANT field, PT EE
Initial plan
Shift plans
Verification plans
0 20 40 60 80 100 1200
20
40
60
80
100
Ratio
of T
ota
l Str
uct
ure
Volu
me [%
]
Ratio of Total Dose [%]
Verification study for CTV 60Gy(RBE), PT EE
Initial plan
Shift plans
Verification plans
0 20 40 60 80 100 120 1400
20
40
60
80
100
Ratio
of T
ota
l Str
uct
ure
Volu
me [%
]
Ratio of Total Dose [%]
Verification study for CTV 60Gy(RBE) with ANT field, PT EE
Initial plan
Shift plans
Verification plans
2PO
2PO 2PO+AP
2PO+AP
29
HN OAR Robustness: Cord/Oral Cavity
0 1000 2000 3000 4000 5000 6000 70000
20
40
60
80
100
Ratio o
f T
ota
l S
tructu
re V
olu
me [%
]
Total Dose [cGy]
Verification study for Cord, PT EE
Initial plan
Shift plans
Verification plans
0 1000 2000 3000 4000 5000 6000 7000 80000
20
40
60
80
100
Ratio o
f T
ota
l S
tructu
re V
olu
me [%
]
Total Dose [cGy]
Verification study for Cord with ANT field, PT EE
Initial plan
Shift plans
Verification plans
2PO
•The shift plans may not always reveal potential problems and
therefore verification volumetric imaging should be used.
•We can learn from robust evaluation how to come up with the
clinically relevant cost functions for robust optimization.
2PO+AP
30
PBS&Interplay Effect Prostate Motion
PBS delivers a plan spots by spots; layers by layers.
Each layer is delivered almost instantaneously.
The switch (beam energy tuning) between layers takes about 1 to 5 s.
Prostate motion during beam energy tuning causes interplay effect.
31
PBS: Interplay & Dose Distribution
Worst Case Scenario Patient – Worst Single Fraction
1 2 3 4 5 6 71
2
3
4
5
6
7
81
2
34
56
78
9
101112
C-C Drift (mm)
A-P
Dri
ft (
mm
)
-2 0 2 4 6 8-2
0
2
4
6
8
12
34
56
7
8
910
1112
C-C Drift (mm)
A-P
Dri
ft (
mm
)
Tang, Both et al, IJROBP 2013
32
Outline
Why Scanning Beams?
PBS and Clinical Planning:
-Minimum Energy/Beam Degraders/Spot Size
- Planning Techniques: SFUD, IMPT, DET
- Robust Optimization and Evaluation
- Interplay
Site Specific Implementation: Technical Protocols.
Summary.
33
Site Specific Implementation: Technical Protocols
Site specific technical protocols have to be developed within the
limitations of existing technology for robust proton treatment clinical
implementation as proton dose distribution may change with:
different machine parameters.
treatment planning algorithms.
patient's anatomy changes in the beam (weight loss, air pockets, ports,
etc) or misalignments of the proton beam relative to the patient.
the way we manage uncertainties.
34
Site Specific Implementation: Technical Protocols
To determine meaningful technical protocols it is important to:
work with the physicians and the treatment team to acquire and
analyze prospective patient data within the treatment environment.
move consistently from simple to complex treatment sites( prostate,
brain, pelvis (GI, GU, GYN), HN, CSI, etc) enrolled and analyzed 10 to 12
on treatment patients’ CT data sets including weekly scans.
recognize potential problems, address them as necessary.
perform dose accumulation studies for moving targets
review the literature and one’s institutional data within the limitations
imposed by differences in technology, patient population, etc.
35
Lower GI: 2POs
SFUD vs.Rapid Arc
36
PBS CSI
Field Geometry
Two lateral PBS fields are used to treat the brain
One or more posterior fields are used to treat the spine
Fields overlap for 5-8cm
A shallow dose gradient is created between the fields in order to
create a safe, smeared field match
H..Lin, S. Both et al..IJROBP,I n press,
37
Target volumes for planning
PBSTV_Compo
PBSTV_Compo
Field specific target volume
PBSTV_compo
oBrain
oSpine
Optimization target volume
PBSTV
oBrain
oSpine
up gradient volume
o 80brain_20up
o 60brain_40up
o 40brain_60up
o 20brain_80up
low gradient volume
o 80upp_20low
o 60upp_40low
o 40upp_60low
o 20upp_80low
H.Lin. S. Both et al, PTCOG 53
38
CSI film measurements on field matching
Sagittal dose profiles comparison for the (a) spinal-spinal and (b) craniospinal junctions. The
blue lines indicated the location to draw the dose profiles.
H.Lin. S. Both et al, PTCOG 53
39
Patient related uncertainties…More complex
Image by B.White, UPENN
40
Dose Accumulation on IACT
Red contour: ITV
CT image shown: IACT C.Zeng, S.Both et al. PTCOG53
Dose accumulation over 8 phases(Left) vs. nominal dose distribution(Right)
Left Right
41
Large Spots
Small spots: σ ~ 5 mm; Big spots: σ ~ 10 mm
42
Summary Dose Conformality is best for PBS techniques and worst for
passive scattering techniques(except patching). High dose region
conformality will improve relative to IMRT as PBS matures.
Dose Heterogeneity in general is best for PBS techniques relative
to passive scattering and IMRT.
Plan Robustness will improve as robust optimization becomes
available, however the problem of motion requires additional forms
of management(minimize motion, tracking, rescanning).
Technical protocols have to be developed for each clinical site
implementation as not like in IMRT, the patient and machine are
decoupled and therefore geometry does not equals dose.
Communication between the planning team and clinical team is
crucial during treatment for a safe and accurate proton treatment.
43
Thank you
Acknowledgements:
Penn Radiation Oncology & Collaborators
44