An Overview of Next Generation Gantries

David Robin

Lawrence Berkeley National Laboratory

With input from

LBNL : S. Caspi, H. Felice, R. Hafalia,  A. Sessler, C. Sun, W. Wan, Postech : M. Yoon, CNAO (ULICE, PARTNER) : V. Lante, M.M. Necchi, M. Pullia, S. Savazzi, J. Osorio Moreno,  PSI : E. Pedroni, GSI : U. Weinrich, HIT: T. Haberer BNL: D. Trbojevic

Outline

•  Some  Gantry  Proper.es  and  Considera.ons    •  Present  State-­‐of-­‐the-­‐Art  

•  Future  Direc.ons  

1.  More  Compact  Gantries  using  Superconduc.ng  Magnets  

2.  Mobile  Isocenter  Gantries  3.  Fixed  Field  Alterna.ng  Gradient  Gantries  

•  Possible  to  access  the  full  4  π  solid  angle  with  a  combina.on  of  gantry  and  pa.ent  rota.on.  

•  A  gantry  is  a  beam  line  that  directs  and  focuses  the  beam  onto  the    pa.ent  at  whatever  angle  is  required  for  the  treatment  plan  op.miza.on.  

What is a Gantry

Some Existing Proton Gantries

PSI Gantry 2

Heidelberg Ion Therapy Carbon Ion Gantry

Only  Carbon  Gantry  Worldwide  

Treatment  with  gantry  possible  by  end  of  2011  

Gantries are Large

Weight    Ø  Proton  gantries  weight  about  100  tons  Ø  HIT  carbon  gantry  weighs  600  tons    

~12  m  

~8  m

 ~13  m

 

~22  m  

HIT  (Carbon)  

 PSI-­‐2  (Proton)  

1/10  of  the  Eiffel  Tower  

Proton Therapy Center

Gantries  are  larger  than  the  accelerator  

 Heidelberg  Ion  Therapy  (HIT)  Facility  

       

This  is  even  more  striking  for  carbon  facili=es  

Carbon  Ion  beams  are  harder  to  bend    

•  Difficulty  to  bend  measured  by  magne.c  rigidity,  Bρ [T  m]  

•  Bending  radius,  ρ,  is  :             ρ = Bρ/BMag  

•  For  the  same  penetra.on  depth    –  Magne.c  Rigidity is  nearly  3  .mes  higher  for  Carbon  Ion  Beams  

     

Why  Carbon  Gantries  Are  Bigger  

BMag  

ρ  

Magne.c  Rigidity    

   

0   2   4   6   8  

Proton  

Carbon  

Bρ [T  m]    

Magne.c  rigidity  for  3  –  30  cm  penetra.on  depth    

Example  :  HIT  90°  Dipole              Field  Strength  is  1.81  Tesla              For  Bρ  is  6.64  Tm  the  bending  radius  is  3.67  m  

Potential For Significant Improvement

•  Gantries  are  expensive  –  Large  structures  –  Requiring  big  heavily  shielded  (~5m  thick)  rooms  –  Substan.al  opera.ng  costs  (up  to  0.8  MW  for  HIT  Gantry)    

•  Advantages  of  gantries  are  such  that  many  new  proton  facili.es  are  being  built  with  mul.ple  gantries  

•  Gantry  development  has  poten.al  for  achieving  significant  gains  in  cost  and  performance  

 

Some Gantry Properties and Considerations

1.  Active Scanning 2.  Large Good-Field Region 3.  Dose Uniformity 4.  Parallel (or nearly) Scanning 5.  Fast Depth Scanning 6.  Fixed Versus Mobile Isocenter

1.  Ac.ve  Scanning  

•  Ac.ve  spot  or  pencil  scanning    -­‐    the  desired  dose  is  delivered  by  scanning  a  small  (few  millimeter)  beam  in  all  3-­‐dimensions  

 

 Will  only  consider  ac=ve  scanning  gantries  

Transverse  (outward)  

Lateral  plane  (Dispersive  and  Transverse  Dimensions)    

2. Good Field Region

Dispersive  

Dispersiv

e  

Transverse  

Good  field  region  is  the  lateral  region  that  can  be  scanned  without  moving  the  gantry  or  pa.ent    

Scanning  Magnets  

Field  Patching  

Move  couch  

Move  couch  

Large  Field  è  Scan  in  one  go  

Small  Field    è  Scan  and  move  couch  (PSI  gantry  1)  

•  Field  patching  is  slower    •  Large  field  may  require  larger  magnet  apertures  and  higher  costs  

What is an optimal size for the good field region?

Move  couch  

•  Dose uniformity (~2.5%) à precise beam position and minimal beam distortion (when scanning or rotating)

•  If not designed well magnetic elements between scanning magnets and patient can cause distortion (i.e. large T terms)

3.  Acceptably Small Beam Distortion While Scanning

x  

y  

y  

x  

How  much  distor=on  is  acceptable?  

Beam Distortion  

Δx  =  kxθx  +  T122θ2x  +  T144θ2

y  +  …  Δy  =  kyθy  +  T324θxθy  +  …  

Scanning  

θ

4.  Parallel  versus  Angular  Scanning    

•  To  minimize  normal  .ssue  dose  à  Parallel  scanning  is  preferable    

•  Source  to  Axis  Distance  (SAD)  

•  But  some  small  angle  might  be  acceptable    

Penalty  of  Angular  Scanning  

•  Higher  rela.ve  skin  dose   40%  increase  in  skin  dose  for  2  meter  SAD  

What  is  the  Minimal  Acceptable  SAD?    

ULICE  Study  

Scanning Magnets Location

•  Large aperture magnet •  Emphasis on magnet

field quality •  Higher operating costs

•  Larger gantry radius •  Larger room size

UPSTREAM  

DOWNSTREAM  

Upstream or Downstream of Final Bend?  

•  Depth  Change  :  5  mm  (ΔL)    •  Field  Change  ~5%  of  Peak  Field  change  in  field  (ΔB/BM)    •  Time  (ΔT)    -  Synchrotron  ~4  seconds    -  Cyclotron  ~0.1  seconds    -  Cyclinac  <  0.01  seconds  

Higher  field  magnets  require  faster  ramping    rates  just  to  maintain  the  same  depth  scanning  rate      

Depth  Scanning  Speed  

!L!t[mm/s] "

16.65 L[mm]( )3/4

BM [T]!B!t[T /s]

Non  Rela.vis.c  Approx  

Gantry Functional Specifications

•  As part of the Union of Light Ion Centers in Europe (ULICE), physicians and medical physicists from CNAO and other European facilities completed a survey

•  Some of the survey results are shown on the next page are useful in understanding the carbon ion gantry functional specifications

Gantry  funcBonal  specificaBons  

Minimum good field size   15 x 15 cm2  

Maximum number of fields per session  

4  

Penetration depth (range)   3 – 30 cm (corresponding energy: p = 60 - 220MeV;  C ion = 120 – 430 MeV/u)  

Voxel dose accuracy   ±1%  

Dose uniformity   ±2.5%  

Voxels characterization   3 x 3 x 3 mm3  

Voxels out of range   1%  

Field position accuracy   ±0.5 mm  

SAD   4 m  

Maximum treatment time   30 min  Minimum required space around isocenter  

60 cm  

Achieved beam directions   4π

Present State of the Art

•  Fast Conformal Scanning with Volumetric Repainting •  Active and Parallel scanning •  Short (few minute) treatment time per field (including

repaintings) •  Large Lateral Good-Field Regions > 10cm x 10cm •  Fixed Isocenter •  Field position accuracy ~1mm •  Full 4π angular coverage

PSI Gantry 2 and the HIT Gantry are the state of the art

•  Derivatives of the same design - Pavlovic Type

Fixed  Isocenter  

Fixed  Isocenter  -­‐  Pa=ent  does  not  move  ver=cally  when  gantry  rotates  

Point  is  at  the  same  posi.on  for  each  gantry  angle  

Pavlovic Type Gantry Design

   

M. Pavlovic, et al, Nucl. Inst. and Meth A 545 (2005) 412-426

Radially very compact by having the last magnet a 90 degree magnet. All the other magnets – quadrupoles, dipoles, and steering magnets, do not take up additional radial space.

The radial size is determined by only three things – (1) the space between the patient and the last magnet, (2) the bending radius of the 90 degree dipole, and (3) the outer diameter of the magnet.

Pros and Cons of Pavlovic Design

Pros •  Efficient use of Magnets •  Radially compact •  Parallel Scanning

Cons •  Last bending magnet needs to have a large aperture to

accommodate the scanning •  Strong requirements on field quality

–  Heavy (~90 tons for the HIT gantry)

Example – PSI Gantry 2

•  Good Field Region 12 cm by 20 cm

•  Fast double parallel scanning –  Lines ~10 ms –  Lateral Plane (one energy) <200 ms (20 lines) –  Beam Intensity Modulation (BIM)

•  Quick Depth Change –  100 ms per 5 mm

 

T   U  

•  Repainting of iso-layers (Goal) •  ~  6  seconds  per  cubic  liter  (20  energies)  

•  Volumetric repainting capability •  10  repain.ngs  /  liter  in  1  (2)  minutes  

Lateral  Scanning    

Treatment  with  gantry  within  a  year  

Fast beam energy changes

Carbon wedges moving against each other in the beam

Fast energy changes with degrader and beam line (including GANTRY 2 magnets) Aiming  at  100  ms  dead  .me  for  range  steps  of  5  mm  

Achieved  80  ms    !!    

Future  Direc.ons  

1.  More  Compact  Gantries  Using  Superconduc.ng  Magnets  

 2.  Mobile  Isocenter  Gantry  for  Poten.al  Cost  Savings  

and/or  Mul.ple  Treatment  Rooms  

3.  Faster  and  Easier  Energy  Scanning  Using  FFAG  Magnets  

More Compact Gantries

Weight    •  Proton  100  tons  •  Carbon  600  tons    ~12  m  

~8  m

 ~13  m

 

~22  m  

HIT  (Carbon)  

 

PSI-­‐2  (Proton)  

What  is  possible?    For  instance  can  the  size  of  a  carbon  gantry  be  reduced  to  about  the  size  and  weight  of  an  exis.ng  proton  gantry?  

Future Direction 1: Reducing the size of the Final 90° Bend?

HIT 90◦ Weighs 90 tons (65% of all rotating beamline

components)

•  The  Final  90° dipole is one of the main drivers of the size, weight and cost of the gantry

Many Requirements

•  Bend  the  beam  by  90◦  •  Large  aperture  to  accommodate  scanned  

beam  •  Rapidly  ramp  the  field  •  Large  SAD:  Point-­‐to-­‐Parallel  focusing  

from  the  scanning  magnets  to  the  pa.ent  

•  Minimize  beam  distor.on  while  scanning  

Superconducting Magnets

•  Attraction –  Superconducting magnets can achieve many times

the magnetic field strength compared with normal conducting magnets

–  more compact lighter systems

•  Some major challenges such as –  Rotatable cryogenic systems –  Rapid field changes for scanning depth

Superconducting Magnets (Cont)

•  Several groups are looking at Superconducting Gantry Magnets (CEA, HIMAC, LBNL, and INFN Genova)

•  Talks –  Noda            “The  rota.ng  gantry  for  the  HIMAC  facility”    –  Kircher            “Cryogenic  gantry  at  CEA“  –  Caspi            “A  SC  dipole  magnet  for  a  carbon  therapy  gantry”

•  Posters –  Felice/Caspi      “Conceptual  design  of  a  curved  superconduc.ng                                                                  combined  func.on  magnet”  –  Fabbricatore    “Superconduc.ng  Magnets  for  Accelerator  Gantries  for                                                                      Hadron  Therapy  ”      

ETOILE    (Superconduc.ng  90  degree  Bend,  CEA)  

•  3.2  Tesla  Superconduc.ng  Dipole  •  2  meter  bending  radius  •  Significantly  lighter  than  HIT  dipole  •  See  talk  by  F.  Kircher  

(17tons  for  magnet)    

M.  Bajard,  EPAC08,  (2008  )1779-­‐1781  

Superconduc.ng  (5  T)  Combined  Func.on  Toroidal  Magnet  (Lawrence  Berkeley  Lab)  

Work  of  Caspi,  Robin,  Arbelaez,  Sessler,  Sun,  Robin,  Hafalia,  Yoon,  Wan,  and  Robin  

•  Double  .lted  solenoid  coil  winding  

More  in  talk  by  S.  Caspi   •  Compact  (1.3  m  radius)  •  Light  (~14  tons)  

•  Large  good  field  region  (26  cm  diameter)  

Scanning    Mag.  Pos.  

Magnet    Structure  

Coil  

Iron  

•  Minimized  beam-­‐shape  distor.on  

Future Direction 2: Mobile Isocenter Gantry

     Fixed  Isocenter                                      Mobile  Isocenter              

Different  Gantry  typologies  1)  Fixed  isocenter  gantry:  Heidelberg  (2007)          

 2)  Mobile  isocenter  gantry:  PSI  -­‐  1  (1992)                    

 3)  Mobile  isocenter  gantry:  PIMMS  study  (1993):  Riesenrad  gantry          

Beam  delivery  system  rotates  around  the  pa.ent  

that  is  not  moved  

Pa.ent  and  magnets  rotate  around  the  central  axis  –  smallest  possible  radius  

Isocenter  moves  according  to  the  beam  direc.on  –  only  one  90°  

bending  magnet  

Mobile Isocenter Gantries

Potential Advantages •  More compact

•  Serve more than one patient room

•  Fewer rotating magnetic components

Not  a  New  Concept  

•  PSI  Gantry  1  (1993)  – More  compact  that  Gantry  2  

Example  1.  Moving  Isocenter  Gantry  serving  4  rooms  

(A.  Brahme,  Erice  2009)  

Example  2:  ULICE  Workpackage  6    Riesenrad  Like  Gantry  

•  Rota.ng  structure  (room  with  the  pa.ent)    significantly  lighter  •  90°  magnet  alignment  is  easier  in  case  of  magnet  rota.on  around  its  axis  rather  

than  in  case  of  complete  rota.on  of  the  line;  

V. Lante, M.M. Necchi, M. Pullia, S. Savazzi, J. Osorio Moreno  

ULICE  Workpackage  6  Riesenrad  Carbon  Ion  Gantry  

Reasons for the final choice (ULICE meeting 10th March 2011)

“The aim of the ULICE WP6 is to analyse the various aspects that influence the design of a carbon-ion gantry and to produce a design of the device. Both medical/functional and technological issues shall be considered to obtain a design that controls costs without limiting significantly the functionalities. The final target is the production of the conceptual design of an optimised gantry”

Geometry if conventional magnets are considered, the gantry layout has to be

innovative

mobile isocentre  

PSI1-like   Riesenrad-like  

mechanical structure cost ~ 105% wrt HIT  

mechanical structure cost ~ 70% wrt HIT  

concrete & excavation cost ~ 1,377 k€

concrete & excavation cost ~ 1,963 k€

Future  Direc.on  3:    Fixed  Field  Alterna.ng  Gradient  Gantry  

•  Small  aperture  high  gradient  dipoles  •  See  talk  by  D.  Trbojevic  

Dejan Trbojevic, Workshop on Hadron Beam Therapy of Cancer, Erice 2009

Future  Direc.on  3:    Fixed  Field  Alterna.ng  Gradient  Gantry  

A  Poten.al  Major  Advantage    •  Large  Energy/Momentum  Acceptance    -  May  not  need  to  change  field  while  scanning  depth  -  Poten.al  to  enable  very  fast  depth  scanning  

•  Ideally  like  to  have  scanning  magnets  ater  the  FFAG  -  FFAGs  become  very  interes.ng  if  small  (~2m  SAD)    -  If  not  FFAG  gantries  will  be  radially  very  large  

!L!t[mm/s] "

16.65 L[mm]( )3/4

BM [T]!B!t[T /s]

These  three  direc.ons  are  not  mutually  exclusive  

•  For  instance  one  can  conceive  of  FFAG  gantries  built  up  of  superconduc.ng  magnets  

•  Or  combining  superconduc.ng  magnets  in  a  Riesenrad  design  

Summary  

•  State  of  the  art  gantries  are  delivering  impressive  performance  

•  Several  exci.ng  direc.ons  being  explored  that  have  the  poten.al  for  achieving  significant  gains  in  cost  and  performance  

•  Some  requirements,  such  as  acceptable  minimum  SAD,  are  important  to  understand      

`  THANK YOU

Superconduc.ng  Magnet  (IRFU  Saclay)