Recent developments in cyclotrons for proton therapy at IBA

Recent developments in cyclotrons for proton therapy at IBA Yves Jongen. Founder & CRO IBA sa We Protect, Enhance and Save Lives.

A typical PT center 30-55 millions for equipment 45-100 millions for the center 80-100 m 35 m 2

The accelerator is a very small part of a PT system A Proton therapy system is much more than an accelerator It is most often a complex, multi-room system, filling a Hospital building. The treatment rooms are larger than the cyclotron vault The total investment is around 100 M, of which 45 M for the equipment The cyclotron represents only 7 M of this! The investment to develop the cyclotron was less than 4 M, out of more than 60 M spent on developing IBA PT system 3

Proton Therapy end of 2007 70,000 40 60,000 35 Patien ts treated 50,000 40,000 30,000 20,000 PT center under operation 30 25 20 15 10 Operating facilities 10,000 5 0 0 1950 1960 1970 1980 1990 2000 2010 4 Courtesy Janet Sisterson & PTCOG

13 IBA PT customers in the world ProCure 2 Chicago MGH, Boston Beijing, China Wanjie, China ProCure 1 Oklahoma City MPRI, Indiana Hampton Univ., Virginia U.Penn, Philadelphia WPE, Essen NCC, Kashiwa UFPTI, Jacksonville, FL NCC, Ilsan Orsay (France) 5

IBA has currently the largest installed base in PT PT Installed base shares - PROTON - (1994-2008) in ROOM S Still River 4% MHI 13% Varian 7% SHI 3% Hitachi 13% IBA 60% 6

Cyclotrons for Proton therapy? In 1991, when IBA entered in PT, the consensus was that the best accelerator for PT was a synchrotron IBA introduced a very effective cyclotron design, and today the majority of PT centers use the cyclotron technology (not only IBA but Varian, Still Rivers) Over these 15 years, users came to appreciate the advantages of cyclotrons: Simplicity Reliability Lower cost and size But, most importantly, the ability to modulate rapidly and accurately the proton beam current 7

Proton beam current regulation 2.5 2 1.5 signal (V) 1 0.5 0-0.06-0.04-0.02 0 0.02 0.04 0.06-0.5-1 t (sec) 8

Change of energy? Cyclotrons are simpler at fixed energy Energy change by graphite/beryllium degrader at waist after cyclotron exit, followed by divergence slits and energy analyzer This very effectively decouples the accelerator from the patient Unlike the synchrotron, the emittance is identical in X and Y. This makes gantry optics much easier in scanning mode Yes, neutrons are produced, but ESS is well shielded and the average beam current in PT is low > limited activation How fast? 5 mm step in energy in 100 msec at PSI (vs. 2 sec for IBA or 4 sec. for a synchrotron). 9

The IBA ESS 10

More Expertise The energy selection system 11

More Experience UFPTI, Jacksonville, USA Construction start date: Mar 2004 PT equipment installation start: Mar 2005 1st Patient : Aug 2006! today : 130 patients/day treated in 3 Gantry rooms up to 250 fields/day 12

The UPHS Particle Therapy Centre, Philadelphia 13 The largest Particle Therapy centre to date! 4 Gantry Rooms 1 Fixed Beam Room (2 beams) + 1 Experimental Room Beam since July 2008 First patient treatment in Autumn 2009

Procure center #1, Oklahoma city, USA 14 First center of the Procure network 2 Gantry Rooms 1 Fixed & Inclined Beam Room Beam since July 2008 First patient treatment in Autumn 2009

Hampton University Proton Therapy Institute 4 Gantry Rooms 1 Fixed Beam Room All equipment installed Beam accelerated in the cyclotron 15

Westdeutsche Protonentherapiezentrum, Essen First Particle Therapy centre based on a Public Private Partnership (PPP) model 3 Gantry Treatment Rooms 1 Double Fixed Beam Room with Eye Treatment line Beam since September 2008 First patient treatment in Autumn 2009 16

New cyclotron and gantry for CPO in Orsay New equipment for an existing PT center New cyclotron, ESS and one new gantry room Transition to be made without interrupting treatments!!!!! 2 existing Fixed Beam Rooms All equipment installed, cyclotron beam extracted, optics tuning ongoing 17

C230 median plane view 18

C230 in numbers 230MeV, 500nA proton beam for therapy Resistive but high field magnet: 2.9T peak field, 1.1m extraction radius, 4 spiral poles, elliptical gap, 800A, 524 ka-turns, 9mm pole gap at outer radius Internal hot filament PIG source RF system: 106MHz 100kW, harmonic mode 4, dee voltage from 60 kv at the center to 120 kv at extraction Electrostatic deflector extraction 19

More Expertise The CYCLONE 230 cyclotron 20

The cyclotron opens at median plane for service 21

Inside the cyclotron 22

The ion source and central region 23

Electrostatic deflector 24

Recent improvements on the C230 15 C230 cyclotrons have been built, but we keep adding improvements. Recent developments include: Correction of slight tilts in the orbit plane Design improvements in the RF cavities New deflector design Improved beam current regulation 25

Recent improvements on the C230 15 C230 cyclotrons have been built, but we keep adding improvements. Recent developments include: Correction of slight tilts in the orbit plane Design improvements in the RF cavities New deflector design Improved beam current regulation 26

RF cavity redesign Problem: Due to the elliptical pole shape, the counter-dee gap decreases with radius. Consequence: Beam losses on counter-dees at large radius. Solution: Maximize the counter-dee gap. Method: Redesign the RF cavities to increase the counter-dee gap from 10mm to 12mm. 27

RF cavity redesign New cavity design 28

RF cavity redesign New cavity design Redesigned area 29

RF cavity redesign New cavity design 12mm 230MeV orbit 30

RF cavity tuning redesign Problem: The present cavity tuning by a variable capacitor in the median plane needs periodic replacement. It is difficult to share the larger RF current drawn by this capacitor equally in the upper and lower cavity. Consequence: Capacitor failures: loss of reliability Lack of up-down symmetry: leaks of RF in the cyclotron through the accelerating gaps Solution: Tune the cavities with inductive tuners in the valleys sliding on RF contacts 31

New cavity tuner design 32

Electrostatic deflector optimization Problem: The old septum intercepted a significant amount of beam. Consequence: Activation, limited extraction efficiency Solution: Reduce septum beam apparent thickness. Method: Analytical study and beam tracking. Then build it and try it! 33

Electrostatic deflector optimization Beam tracking simulation of new deflector. Pole edge Extracted beam Circulating beam Deflector 34

Electrostatic deflector optimization Beam tracking simulation at JINR Comparison between old and new septum : Old New Losses on septum entrance (0.1mm) Losses inside deflector on septum Losses inside deflector on HV plate 28%* 8% 0% 9% 1% 1% 35 * plus circulating beam

Electrostatic deflector optimization 36

Experimental results Radial track using integral radial probe Beam current on external beam stop 37 (Raw data, not corrected for RF noise and radial probe efficiency.)

Experimental results Radial track using integral radial probe Deflector Adj. dowel pins New RF cavity New deflector Adj. dowel pins New RF cavity Integral radial probe beam current External beam stop current 38

Proton beam current regulation optimization Present situation The proton beam current is slaved to an external time function by measuring the extracted beam with an ion chamber, doing a digital regulation by varying the arc current in the ion source. The current loop regulates the beam current with an accuracy better than 2%, up to a bandwidth of 2.5 KHz Problem: We have a dark current. Even when the arc is turned off, a proton beam current of 30 to 100 picoampere is extracted from the cyclotron 39

Proton beam current regulation optimization Consequence: The dark current results in small inaccuracies in beam delivery. In pencil beam scanning, it can result in small amounts of beam being delivered outside the treatment field Solution: Use a reduction of the dee voltage to suppress the proton beam when it is not needed. Possibly, use the dee voltage variation exclusively to regulate the beam current 40

3 kv dee voltage variation is enough Beam current vs Dee Voltage 450 400 350 Beam current (na) 300 250 200 150 100 50 0 39 39.5 40 40.5 41 41.5 42 42.5 43 Dee Voltage (kv) 22/11/2007 11:30 41

Modulating the dee voltage by 3 kv at 62 khz! 42

Regulating 500 µsec pulses 43

Data analysis on IC cyclo, 4 na peak Regulation triangle 100 Hz 4 na 4.5 4.0 Beam current on IC cyclo (na) 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Series1 Linear fit -0.5 0 2 4 6 8 10 12 14 16 18 20 Time (msec) 44

Thank you 45