Pedroni Eros Paul Scherrer Institute - Villigen PSI, Switzerland

Transcription

1 Pedroni Eros Paul Scherrer Institute - Villigen PSI, Switzerland Center for proton therapy - PSI GANTRY DESIGN AND RECENT EXPERIENCE AT PSI 2nd Workshop on Hadron Beam Therapy of Cancer ERICE May 23 th, 2011 on behalf of the Gantry 2 group

2 Layout of the presentation Gantry 1 experience Facility expansion at PSI - new SC cyclotron The new Gantry 2 of PSI Layout Beam optics and beam size Energy variations with the beam Sweeper magnets calibration Advanced beam scanning techniques Possible future clinical use of Gantry 2 Page 2

3 Early 90's GANTRY 1 USED FOR PATIENT TREATMENTS SINCE 1996 GOAL in 1989 BASIC FEASIBILITY OF PENCIL BEAM SCANNING Page 3

4 The long term experience of using GANTRY 1 Designed in 1991 for protons On the basis of the scanning experience with pion therapy with inverse planning based on CT data Gantry 1 Page 4

5 System characteristics of Gantry 1 Upstream scanning applied only in the dispersion plane of the beam Magnetic scanning started before the last bending magnet Eccentric mounting of patient table on the gantry front wheel (with counter-rotation) Gantry radius reduce to 2 m Still the smallest proton gantry in the world α rotation β rotation φ rotation Page 5

6 If i could do it again Eccentric mounting as with Gantry 1 (R = 2 m) But with rotation only on one side (0 180 ) as with the new Gantry 2 Permanent floor underneath the patient table "Treatment cell" moving with the gantry 2 m compact eccentric gantry A new Gantry 3? Page 6

7 Pencil beam scanning Small pencil beam: 3 mm σ (7-8 mm FWHM) Cartesian scanning (infinite SSD) Discrete spot scanning step and shoot,method - on a 5 mm grid From 1996 until May 2008 the only scanning gantry in the world Elements of scanning: Dose Monitor+Kicker X Sweeper magnet 5 ms / step fast Y Range shifter 30 ms average Z Patient table 1 cm/s slow 100 us reaction time Page 7

8 Patient handling with Gantry 1 Use of a CT outside of the treatment area Position next patient while treating the previous one Transfer of positioned patients with automated transporter Very useful for treating children with anesthesia Patient positioning control with X-ray on Gantry 1 (optional) Page 8

9 Clinical use of GANTRY 1 In use since1996 Full fractionation ~30 fractions Treatments 8:00 16:00 Max 19 patients/day (2.5 per hour) 2.8 fields/fractions in average 1/3 of patients are children Under anesthesia 1/3 of treatments are IMPT Courtesy of B. Timmermann Weak points of Gantry 1 Table motion is part of scanning Not possible to use collimators Not possible to apply repainting We treat only non moving targets! Page 9

10 Advantages of using scanning (vs. scattering) Automated treatments all by software Minimal simple equipment The (pencil) beam "does it all" Same approach from small to large fields No individualized hardware (no fabrication of collimators compensators ) Apply dose fields in sequence without personnel entering the treatment room Simplify treatment - reduce treatment time and costs True 3-dimensional conformation Variable modulation of the range Minimal neutron background (for the patient) Lower risk of second cancer than scattering (depends on sophistication) Important for pediatric treatments Less activation of equipment in the nozzle (for the personnel) Less activation of the accelerator Better lateral fall-off as compared to scattering (?) No material in the beam path (except air and monitors) - Less MCS effects Page 10

11 Dose shaping within the target IMPT (intensity modulated PT) Delivery of IMPT and IGPT The term of comparison with IMRT (conventional therapy) Biological targeting (image guided proton therapy) Intentional non-homogeneous dose distributions Dose proportional to the tumor activity (biological signal) The topic for the future Courtesy of A. Lomax PSI Page 11

12 Major disadvantage: the organ motion sensitivity Interference of organ motion with the scanning sequence M.Phillips PSI 1992 Disturbance of the dose homogeneity within the target With Gantry 1 we treat only immobilized lesions attached to bony structures Tumors in the head, spinal chord and low pelvis We accept only movements < + -2 mm for treatments with full fractionation! Can we overcome this drawback? Possible remedies : Fast Scanning (Repainting) Gating Tracking? or a combination of those the possible points to be developed with Gantry 2 The point of scanning to be improved Page 12

13 2000 EXPANSION OF THE PROTON FACILITY AT PSI SUPERCONDUCTING CYCLOTRON FAST DEGRADER LAMINATED BEAM LINES GOALS: Stable DC beam Modulation of the beam intensity at the ion source Very fast energy changes Page 13

14 Layout of the newly expanded proton facility COMET - dedicated superconducting cyclotron [ACCEL - design H. Blosser] Beam for Gantry 1 all the year through patients treatments restarted in February 2007 no shut-downs since August 07 Horizontal beam line for OPTIS 2 transfer from OPTIS 1 last year Next generation scanning gantry : Gantry 2 1. patient planned for 2012 Medical cyclotron Exp. Area (PIF) OPTIS 2 Gantry 2 Disconnected from PSI ring cyclotron in 2006 Gantry 1 Page 14

15 Facility specifications mainly for the new Gantry 2 Super-conducting cyclotron Very stable beam at the ion source Aiming at 2-3% at the µs scale Deflector plate in the first orbit Dynamic control of the beam intensity µs time scale Fast degrader (moving carbon wedges) Continuous choice of the beam energy Beam line with laminated magnets Providing fast changes of beam energy Page 15

16 Excellent beam current stability an example Change of paradigm of testing the beam monitors Use the inherent stability of the beam to check the fine tuning of the electronics (capacitive matching) Result: improved linearity of the dose < 0.5% down to a 0.1 Gy dose Ideal beam for scanning 100% duty factor Dose measurements for diff. Plugin configurations 600 µs 33 nf 1.01 m e a s u re d D o s e /m e a s u re d D o s 1 G y applied Dose [cgy] ISO, 22pf ISO, 47 pf ISO, 33 pf ISO, 39 pf 22 nf Page 16

17 2005 A NEW PSI GANTRY - GANTRY 2 MAIN GOAL DEVELOP FURTHER SCANNING AS A UNIVERSAL BEAM DELIVERY TECHNIQUE Page 17

18 GANTRY 2 TOPIC 1: INNOVATIVE GANTRY LAYOUT READINESS FOR IMAGE GUIDED PROTON THERAPY EFFICIENT PATIENT HANDLING Page 18

19 Design started from the patient table Beam Line Support Bearing axle From -30 to +180 Patient table Room with fixed floor 0 to +180 would have been a better choice (cheaper) Page 19

20 Layout of the Gantry 2 room: patient table, compact nozzle Small compact nozzle Same patient table as at RPTC in Munich (Schär Engineering) Easy access to the patient table on fixed floor Fixed walls and ceiling for mounting commercial equipment (Vision RT?) A system open on both sides - lateral and front - Optimal for using an in-room sliding CT Page 20

21 In-room sliding CT - within reach with the patient table The beginning of IGPT? Patient positioning (tumor in soft tissues region) Setup for treating moving targets - 4D CT (relation external gate - internal motion) Page 21

22 BEV X-ray - synchronized with proton beam delivery BEV imaging - equivalent to portal imaging with photons Very large field-of-view (26 cm x 16 cm) not masked by equipment or collimators in the beam path QA control of gating and tracking (scanning + pulsed X-rays) Fluoroscopy mode Beam guidance? Proton beam Sweepers UPSTREAM SCANNING Bend and scan Scan and bend X rays tube Yoke hole Bending magnet Sweeper or Scatterer IGPT nozzle Patient Imager a) compact gantry b) long throw gantry Collimator Page 22

23 BEV - imager Check patient position at the isocenter Photograph of the retractable arm for holding the X-ray panel behind the patient on the side opposite to the nozzle (BEV X-ray). Page 23

24 GANTRY 2 TOPIC 2: PENCIL BEAM AND BEAM OPTICS GOALS: SMALL PENCIL BEAM FOR PRECISION THERAPY PARALLELISM OF SCANNING Page 24

25 GANTRY 2 beam optics M A 3 2 Q 6 W P T A 3 X Q 4 S Q 5 Q 3 Q 7 W U 3.2 m P M 1 Q C M2 S Q 1 Q 2 A 1 SPECIFICATIONS Rotational symmetric (large) phase space +- 3mm mrad % dp/p Complete achromatism Scanning-invariant beam focus Orthogonal (x-,y-) focal planes in T and U Focus to focus with 1:1 imaging from the coupling point of the gantry to the iso-center 2D- parallelism of scanning (in T and U) (on Gantry 1 only U) Page 25

26 GANTRY 2 beam optics (TRANSPORT and RAYTRACE) Q C A 1 A 2 W T W U H S Double parallel scanning A 3 Point to point focus Achromatic Q 1 Q 2 Q 3 Q 4 Q 5 Q 6 Q 7 Double parallel scanning TRANSPORT beam envelopes through the Gantry 2 Page 26

27 Advantages of coupling the gantry at a small beam focus At a small beam focus the spot image is insensitive to scattering Decouple vacuum (gantry from facility) Collimate the beam at the coupling point Forced beam centering on axis Measure the beam characteristics on-line before the gantry On-line independently of Gantry rotation and scanning Manipulate the beam shape imaged on the scanning beam Large Gaussian beam shape? Virtual dynamic collimation at coupling point? Disadvantage A focus coupling needs a longer beam line coupling with a parallel beam Page 27

28 Less skin dose Simplify treatment planning Advantage of using parallel scanning No inverse squared distance effects Simplify dosimetry and QA control Dose value and dose distribution in the patient and in the phantoms are the same Easier field patching (expansion of the scan range for treating long targets) Can be done without optimization within treatment planning Just exchange magnetic scan position with patient table offset Compensators (simulated scattering) No dose errors due to inverse R-square effects Collimators No tapered faces necessary Page 28

29 The difficult piece: the last 90 bending magnet A big beast - 34 tons 15 cm gap To accommodate a scan area of 20 x 12 cm Lamination leaks but in the end Vacuum problems well solved! Photograph of the mounting of the 90 bending magnet. Integrated vacuum chamber embracing the poles of the magnet 180 tons mechanical isocenter within mm Page 29

30 Compact optimized nozzle Vacuum up to the patient Sharp pencil beam - 3 mm sigma Two monitors and a strip monitor 2 mm strips (TERA collaboration) Removable pre-absorber IN and OUT of beam For ranges below 4 cm Telescopic motion of the nozzle To reduce air gap (keep patient at isocenter) Option to add collimator (compensator) 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) Page 30

31 Results: pencil beam size (on axis) Minimize material in the nozzle for keeping To have a sharp lateral fall-off the beam size between 3 and 4 mm sigma at all energies ( MeV) Region MeV 0.6 Adding piece by piece the materials in the nozzle Page 31

32 GANTRY 2 TOPIC 3 : ENERGY VARIATIONS WITH THE BEAM LINE COMPENSATED INTENSITY-ENERGY LOSSES VERY FAST ENERGY CHANGES Page 32

33 Compensation of beam losses in the range MeV na Constant ratio SC cyclotron Degrader Beam analysis Gantry 2 Coupling point na Rate Monitor Collimators for Intensity suppression khz Goal : reserve the use of the deflector plate for intensity modulated dose painting Energy [MeV] Page 33

34 Optimization for maximum transmission at 100 MeV Beam envelopes at 100 MeV. Horizontal (red) and vertical (blue) envelopes Green line = transmitted beam intensity Dots beam sizes calculated with TURTLE Page 34

35 Intentional beam defocusing at higher energies Beam envelopes at 200 MeV. yellow we four beam line sections. Beam defocusing at collimators (black bars) Same intensity losses as at 100 MeV Page 35

36 Taking into account hysteresis effects ( MeV) Energy loop Red: up-down Blue: down-up Uncorrected 1 mm Corrected Page 36

37 With very fast energy changes. 80 ms But - we observe Slow change of lateral beam position after a big energy change in the range MeV 1-3 mm with exponential decay with decay time of ~2 s < 0.3 mm position change after small energy changes within the SOBP We plan to correct these shift with sweeper offsets as a function of the time after the last energy change Strategy: Fixed targets - range precision of < 0.5 mm while working with full ramp MeV Moving targets - repainting - precision ~ 1mm repaint only the SOBP up and down Scintillator block the beam of Gantry 2 seen with a TV camera Page 37

38 GANTRY 2 TOPIC 4 : SWEEPER MAGNETS COMMISSIONING ISSUES NON LINEARITIES OF THE SWEEPERS Page 38

39 Need of a very precise mapping of the sweeper's action 210 MeV MeV 210 MEV 150 MEV T (cm) 0 T (cm) U (cm) 80 MeV 8 80 MEV T (cm) U (cm) U (cm) Measured (red) and calculated (blue) spot maps of the (linear input) action of the sweeper magnets on the scanned beam position. The non linearities are due to a curvature of the effective boundary of the magnetic field of the 90 bending magnet which is changing with energy (changing sign) Page 39

40 Invariant beam focusing needs some "tricks" First tests in MeV QMF1: 0% QMF1: -10% QMF1: -20% T U QMFC in series Anticipated by Raytrace calculations Tapering of the poles surface of the U sweeper Quadrupole corrector QMFC switched in series with the U-Sweeper Results: acceptable scanning-invariant focus Page 40

41 After a proper mapping of the sweepers 70 MEV 120 MEV 170 MEV 220 MEV Beam spots (2 cm steps) at the isocenter covering the scan region of 12 cm x 20 cm Page 41

42 GANTRY 2 Topic 5 : the main goal of Gantry 2 NEW ADVANCED BEAM DELIVERY TECHNIQUES TO PROMOTE SCANNING AS A UNIVERSAL BEAM DELIVERY TECHNIQUE Page 42

43 Aiming for highest scanning speed From spots For reducing organ motion errors Goal - Fast painting with volumetric repainting Painting lines < 10 ms per line (10cm + line change) Painting energy layers 200 ms per plane (20 lines x 5 mm) Change of energy (100 ms - 5mm range) Painting of volume 6 s per liter (20 energies at 5mm steps) Volumetric repainting capability (aim) repaintings / liter in 2 minutes To lines To contours Page 43

44 Conformation requires non-homogenous proton fluences U depth T 3d-Dose homogeneous Repainted for coping with organ motion Proton fluence of energy layer is non homogeneous Page 44

45 FPGA based delivery of tabulated energy layers Time driven beam delivery Beam path downloaded as tables of the sweepers U - T and beam intensity I Dose control with feed-back loop: Monitor 1 required dose -> vertical deflector plate A) If sweeper speed is the limit use variable intensity (IM) B) If beam dose rate is the limit use variable sweeper speed (SSM) C) or use both! Full range of dynamic dose control from zero IM to any dose SSM Delivered MUs Dose linearity of simple T-lines Required Dose 5ms - 10 cm lines - painted with IM 23 times Max T speed Variable intensity Page 45

46 GANTRY 2 POSSIBLE CLINICAL USE OF GANTRY 2 GOAL: PROMOTE SCANNING AS A UNIVERSAL BEAM DELIVERY TECHNIQUE to completely replace scattering Page 46

47 Often required Improve lateral fall-off for treating static targets Prostate sharp lateral fall-off at the boundary between tumor and sensitive structure Scanning alone better at high energy (edge enhancement with pencil beam) Scanning with added collimation better at low energy (beam size limitations) Scanning with varying beam energy superior to scattering? less material in the beam path Brain stem FWHM mm Idealized scattering (zero phase space) Realistic scanning Factor 1.4 gain Difference Gauss to error-function (1.7) Scanning with collimation better Scanning alone better Range mm Page 47

48 Very big tumors Combine the use of fast scanning with patient table displacements Needs remote control of the patient table (collision detection) Take advantage of the parallelism of the beam trivial patching - shift table - and apply intensity filter to spot pattern Medulloblastoma Dorsal irradiation Example: a paediatric case treated in 2004 with Gantry 1 Page 48

49 Moderately moving targets (0-5 mm) Discrete spot scanning applied with iso-layered repainting (max dose per spot visit) Treatments in the trunk (pancreas, cervix, colon), breast, lymph nodes, etc. Advantage of having fast energy changes -> Volumetric repainting FAST SCANNING IS MORE THAN JUST REPEATING COMPARE G1 spots, scaled G2 spots, scaled G2 spots, iso-layer G2 lines, scaled Pancreas Rectum Breast From Simulation of repainting strategies S. Zenklusen's thesis work PMB 55 (2010) Page 49

50 Largely moving targets Volume painting within a single breath hold? (repeated) Using Conformal line painting Speed of painting 0.2 Gy of a sphere of (17 layers) ½ liter in 7 secs If not working, Gating Lung liver 7s Page 50

51 Simulated scattering Dose shaping with compensators Uniform scanning Energy layer with homogeneous fluence Magnetic line scanning at max. speed Very high repainting number Collimator optional As a sub-mode of conformal scanning To simulate scattering on a scanning-gantry To replace completely scattering?? Retinoblastoma? Page 51

52 Feasibility demonstration of "simulated scattering" 8 cm diameter sphere Field shape = target projection minimal neutron background Variable modulation of the range Layer shrinking Parallel beam no compensator dose errors Highly repainted Distal layer: repainted 48 times in 30s From S. Zenklusen PhD - Medical Physics 2011 Page 52

53 Status of the Gantry 2 project Significant project delays due to the Priority given to OPTIS 2 in 2008 Difficulties with the vacuum of the 90 bending magnet PSI internal development of the software for the control of the patient table insufficient resources THANK YOU Page 53