New Developments in Proton Treatment Planning Systems Daniel … · 2009-05-12 · Beam Profile...

Post on 17-Jun-2020

9 views 0 download

transcript

New Developments in New Developments in Proton Treatment Planning Systems

Daniel Yeung

Statement of Disclosure

Funded Research & Development:

Philips Medical Systems

IBA

Proton Planning Systemsg y

C i l S t A d i S tCommercial Systems Academic SystemsAnalytical/Semi-Analytical CMS - Xio MGH, HCL

Varian – Eclipse Proton PSIVarian – EyePlan ClatterbridgeOptivus - Odyssey Loma LindaDosigray OrsayDosigray Orsay

Others… Others…* Under Development Philips – Pinnacle* UF*Monte Carlo CMS*, Varian*,… MGH, DKFZ,…

Proton Pencil Beam Algorithmsg

Hong et al, MGH 1996

Szymanowski et al, Institut Curie 2001

Beam Model dosimetry in water inhomogeneity correction g y

• water equivalent thickness (wet) • HU stopping power

Proton Pencil Beam Algorithmg

Beam Model

CAX Depth Dose (DD) broad beam pristine peak sobp

Radial Spread (RS) multi coloumb scatter

• beamline: degrader, nozzle elements (wet only)• compensator: material dependent scattering powerp p g p• patient: wet only

gaussian kernelσ2

total = σ2line + σ2

comp + σ2patientσ total σ line σ comp σ patient

Proton Pencil Beam Algorithmg

Dose CalculationDose Calculation

Dose from a pencil beam

Convolve DD with RS for each pencil Sum dose from all pencils

Bortfeld Model of Pristine Peak

Analytical model of proton BP (up to ~ 200 MeV)• accounts for energy spread • empirical model of nuclear fragmentation (data fitting)empirical model of nuclear fragmentation (data fitting)• numeric depth dose calculation of fitted BP• assumption – range straggling ‘constant’ with depth

Pristine Peak - Analytical Model y

R=9 0 g/cm2R=5 86 g/cm2

R=21.86 g/cm2

R 9.0 g/cmR 5.86 g/cmV 8.0

V 8.1R=19.8 g/cm2

For R > 15 g/cm2,

Range straggling depth

E i B tf ld d l

Eclipse Model Bortfeld Model

Error in Bortfeld model

Pristine Peak – Bortfeld Model

R=19.8 g/cm2

R=28.4 g/cm2

Bortfeld Model Bortfeld - Enhanced

SOBP Calc vs MeasurementCalc vs Measurement

Courtesy R. Slopsema

Beam Profile Calc vs MeasurementCalc vs Measurement

R=15.1 M=10.4

d=5 mm

d=10 cm

Option B5 - R=15.1 g/cm^2, M=10.4 g/cm^2, Air gap = 11.7cm, SSD=220.1, aperture: 15cmx15cm

1

1.11.2

1.3

umbr

a [c

m]

Eclipse convolved - @9.9 cm Measurement (av) - @9.9 cmEclipse convolved - @0.5 cm Measurement (av) - @0.5 cmEclipse convolved - @14.1 cm Measurement (av) - @14.1 cm

0.5

0.6

0.70.80.9

180

% L

ater

al P

enu

d=15 cm 0.3

0.4

0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00Air Gap [cm]

20%

-8

Courtesy R. Slopsema

Proton Dose Calc in water is generally accurate!

…ok, what about Clinical Issues?

Physics Clinical

Match & Patch Fields

used to avoid OARs adjacent to target used to avoid OARs adjacent to target partition target into segments (sub-targets)

• sub-targets treated with ‘sub-beams’ • angle sub-beams to avoid OARs

combined with other fields for dose uniformity

Match Fields match fields abutting each other penumbra matching penumbra penumbra matching penumbra

Inf - LAT Sup - RAO

matchline

Patch Fields

th b t t ti l t t thru beam txt partial target residual txt with patch lateral penumbra (t beam) ‘matched’ with distal lateral penumbra (t-beam) matched with distal

falloff (p-beam) LPO beam (inferior) patched with SPO (superior) LPO beam (inferior) patched with SPO (superior)

Lacrimal Gland CarcinomaPTV 50 4PTV 50.4

partition into sup + inf targetsb i t

sup

inf

brainstem

inf

PTV50 4 5 fi ld ith t h & t h

LPO-Inf (spare optics & BS)SPO-patch

PTV50.4: 5 fields with match & patch Target FieldsPTV LAO LPO50%

PTV(Inf) LPO-InfPTV(Sup) LAO-match SPO-patch

Lacrimal Gland CarcinomaPTV50 4 – 5 FieldsPTV50.4 5 Fields

Inf - LPO Sup – LAO match Sup – SPO patch

LPO LAOLPO LAO

Patch Field Selection

Patch Field Angle Selection optimal geometric coverage (G = 230 ) avoid inhomogeneity along path (G = 205 )

Patch line Patch line

Gantry 230 Gantry 205

Patch Field – Beam Angle Selectiong

Gantry 230° Gantry 205°

G=205°

G 230°G=230°

Distal Blockingg

selective pullback of range to spare OARs pullback achieved with added compensator potential pitfalls setup error or motion may nullify sparing ‘simple’ distal blocking may compromise

target coverage b t f h assess robustness of approach

Distal Blockinggadded compensator

distal blockTarget

p

OAR

range pullback

Distal Blockingg

underdose

Whole BS Partial BS

PTV

Whole BS Partial BS

BS

RPO FieldPartial BS

Clinical Tools

I t t Cli i l kflIntegrate Clinical workflow clinical database web-based applications mu model physics qa plan evaluation

quality assurance clinical efficiency & efficacy

Plan/Dvh Evaluation/

Dvh statistics for CTV, PTV, OARs

Courtesy R. Malyapa, C.McKenzie, Z. Li

, ,are extracted manually from plots

RT PACSTPS

Treatment Plans Plan, Dose , ROI Information

Clinical RT-PACSTPS Information Server

Active Forms

SharePointServerPhysics

• Plan• Contouring • DVH

Q/A Forms

Clinicians

• Prescription

Tx Delivery

Electronic Q/A Process:Electronic Q/A Process:Upon export of Upon export of txtx plan to RTplan to RT--PACS, clinical Q/A forms are generated. PACS, clinical Q/A forms are generated. Tx. DeliveryActive forms are sent via email to the personnel on the list . Active forms are sent via email to the personnel on the list . Contents of active forms is stored in a Clinical Information Database.Contents of active forms is stored in a Clinical Information Database.

Courtesy V. Frouhar

Electronic Q/A Process:Electronic Q/A Process:U f l RTU f l RT PACS i li i l Q/APACS i li i l Q/AUpon expert of treatment plan to RTUpon expert of treatment plan to RT--PACS, appropriate clinical Q/A PACS, appropriate clinical Q/A forms are generated. Active forms are sent via email to the personnel forms are generated. Active forms are sent via email to the personnel on the list . Contents of active forms is stored in a Clinical on the list . Contents of active forms is stored in a Clinical Information Database.Information Database.

Integration of Clinical Workflow

Patient Information

R&VHIS

External Connectivity

IGRT LinacTPSPhysics Calc. Clinical

D t M t Machine Data

TPSCT SimPET/MRQuality

Assurance

RXDelivery Setup

Data Management&

Workflow

Clinical I f ti

PhysicianPhysicist

Information DosimetristTherapist

Golden Beam Data

TPS commissioning time consuming share beam data among gantries (institutions)?s a e bea data a o g ga t es ( st tut o s)? golden beam data set accuracy requirements on modeling parameters accuracy requirements on modeling parameters

• pristine depth dose & sobp• effective source size• virtual sad• effective sad

Golden Beam Data

max error in penumbra of ±0.5mm in air, at isocenter

Courtesy R. Slopsema

Golden Beam Data

Summaryy

Analytical Proton Algorithms provide accurate dose Analytical Proton Algorithms provide accurate dose model

Proton specific clinical planning issues requires p p g qvigilance

Dose plan is a snap shot of dose distribution Dose delivered depends on uncertainties in range, setup, organ motion, etc.

Select beams and parameters to minimi e ncertainties Select beams and parameters to minimize uncertainties Tools to integrate clinical workflow are essential Golden beam data looks feasible Golden beam data looks feasible

Thanks!