1. Reinhard W. Schulte Loma Linda University Medical Center
Applications of Silicon Detectors in Proton Radiobiology and Radiation Therapy
Reinhard W. Schulte
Loma Linda University Medical Center

2. Outline Introduction to proton beam therapy
Applications of silicon detectors
Proton radiography and computed tomography
Particle tracking silicon microscope
Nanodosimetry

3. A Man - A Vision
In 1946 Harvard physicist Robert Wilson ( ) suggested*:
Protons can be used clinically
Accelerators are available
Maximum radiation dose can be placed into the tumor
Proton therapy provides sparing of normal tissues
Modulator wheels can spread narrow Bragg peak
The use of protons for radiation therapy was first suggested by Robert Wilson in 1946.
Wilson recognized that accelerators were under construction at that time, which were capable of generating proton beams of sufficient energy to provide a range in tissue comparable to body dimensions.
Wilson realized that the large mass of the proton would minimize lateral scattering and that the energy deposition pattern would allow to place the maximum dose, the Bragg peak, within the tumor, thus providing maximal sparing of normal tissues.
In his 1946 paper, Wilson also proposed that rotating modulator wheels could spread out the Bragg peak to cover larger tumors.
*Wilson, R.R. (1946), “Radiological use of fast protons,” Radiology 47, 487.

4. Short History of Proton Beam Therapy
1946 R. Wilson suggests use of protons
1954 First treatment of pituitary glands in Berkeley, USA
1956 Treatment of pituitary tumors in Berkeley, USA
1958 First use of protons as a neurosurgical tool in Sweden
1967 First large-field proton treatments in Sweden
1974 Large-field fractionated proton treatments
program begins at HCL, Cambridge, MA
1990 First hospital-based proton treatment center
opens at Loma Linda University Medical
Center
Berkeley was the birth place of particle therapy including proton beam therapy.
Two years after Wilson’s paper, the 184-inch Cyclotron at the Lawrence Berkeley Laboratory became available for physics and radiobiological investigations in preparation for human use.
The first therapeutical use of proton beams in humans was for pituitary hormone suppression in the treatment of patients with metastatic breast cancer in 1954 performed by John Lawrence, a medical doctor and brother of Ernest Lawrence the inventor of the cyclotron, and Cornelius Tobias a Berkeley physicist.
The choice of the pituitary as the first target was straightforward: It is a well-localized gland which is closely surrounded by radiosensitive neural structures but also by bony landmarks which made it locatable on x-ray films.
From treatment of normal glands for hormonal suppression there was only a little step to treat hormon producing tumors, so called adenomas, which was also done at Berkeley.
Inspired by the Berkeley work, the famous Swedish chemist and Nobel laureate Svedberg encouraged the young biophysicist Borje Larsson and the famous neurosurgeon Leksell to use the newly built synchrocyclotron in Uppsala as a neurosurgical tool.
Larrson and the Swedish radiation oncologist Sten Graffmann also developed first concepts for large field radiotherapy with protons and treated about 60 tumor patients during the 1960s.
Meanwhile, on the other side of the Atlantic during the early 1960s, the neurosurgeon Dr. William Sweet at the MGH in Boston encouraged the young neurosurgeon Dr. Ray Kjellberg and the Harvard physicist Andrew Koehler to use the proton Bragg peak of the Harvard cyclotron as a neurosurgical tool for the treatment of pituitary tumors. They developed an impressive treatment technique which would be used for the next 30 years.
Another turning point came in the early 1970s when Dr. Ian Constable, ophthalmic surgeon together with Andrew Koehler, by now director of the HCL, developed a treatment program for malignant melanomas of the eye. The extraordinary success of this program lead to the worldwide acceptance and use of this technique.
Also at MGH in Boston, Dr. Herman Suit, a well-known radiation oncologist, Michael Goitein, a physicist, and their colleagues initiated a large-field fractionated radiation treatment program for base of skull and brain tumors, for which they developed useful tools for three-dimensional treatment planning.
The great success of the Boston group lead to a rapidly growing interest worldwide in the development of clinical proton beam therapy and to the opening of the first hospital-based center for proton therapy established by Dr. James M. Slater at Loma Linda University Medical Center in 1990.

5. World Wide Proton Treatment Centers
The early proton treatments were undertaken on machines designed for physics research, such as the LBL bevatron and the Havard cyclotron. Someof these historical machines are still operating and treating patients but are now replaced by proton accelerators located in major hospitals.
There are currently 23 proton facilities operating worldwide with plans to construct another 21 facilities as shown in this slide.

6. LLUMC Proton Treatment Center
Hospital-based facility
Fixed beam line
MeV Synchrotron
Gantry beam line
The opening of the hospital based facility at Loma Linda in 1990 operating with state-of-the-art equipment and computing technology provided the first opportunity for protons to show their full clinical potential in radiotherapy.
The facility, which has treated almost 7000 patients by now, is equipped with a compact synchrotron providing protons of energies high enough to reach any anatomical site. The accelerated protons travel through a beam transport system, directed and focused by dipole and quadrupole magnets to either one of 4 treatment rooms, including 3 rooms with rotating proton gantries, or a fixed beam room with two horizontal beam lines.
In the treatment room, the beam delivery system aims the protons at the tumor site via a nozzle, which spreads, shapes and measures the dose of proton radiation.
The staff at the LL facility treats about 100 patients every day and performs beam calibrations, QA tests, and research at night and during weekends.

7. Main Interactions of Protons
Electronic (a)
ionization
excitation
Nuclear (b-d)
Multiple Coulomb scattering (b), small q
Elastic nuclear collision (c), large q
Nonelastic nuclear interaction (d) p’ p
nucleus
g, n e
(b)
(c)
(d)
(a) q
Protons interact with matter in several different ways.
They primarily lose their energy through numerous interactions with atomic electrons.
Electronics interactions resulting in energy deposition occur randomly which leads to statistical fluctuations of the number of collisions and the energy transferred in each collision, which is known as energy straggling or range straggling.
The energy deposition is relatively independent of the composition of tissue but rather sensitive to the density of the traversed medium.
Proton undergo multiple small-angle scattering events in forward direction by multiple Coulomb scattering on heavier nuclei, which leads to an angular divergence of the beam.
Larger scattering angles are seen in elastic collisions with nuclei of equivalent or greater mass, which often leads to a large energy loss and removal of the primary proton.
Nonelastic nuclear interactions occur at higher proton energies and produce secondary particles such as protons, neutrons, b and g rays and heavy recoils. These secondaries usually stop in the vicinity of the interaction and have a relatively high biological effectiveness. Primary protons are lost in nonelastic nuclear interactions.

8. Why Protons are advantageous
Relatively low entrance dose
(plateau)
Maximum dose at depth
(Bragg peak)
Rapid distal dose fall-off
Energy modulation
(Spread-out Bragg peak)
RBE close to unity
The depth dose distribution of monoenergetic or energy modulated protons, when compared to that of high energy photons, has characteristic advantages for therapeutic applications.
A monoenergetic proton beam has an entrance region of slowly increasing dose, the plateau, which is followed by an ever more rapid increase of dose leading to a sharp peak, called the Bragg peak, named after the physicist William Bragg. Beyond the Bragg peak the dose rapidly goes down to zero.
For therapy purpose, multiple Bragg peaks of different energies are superimposed to create a region of relatively uniform high dose, called the SOBP, which is suitable to cover larger treatment volumes.

9. Why Silicon Detectors
Combined measurement of position, angle and energy or LET of single particles
High spatial resolution (microns)
Wide dynamic energy range
radiation hardness
compatibility with physiological conditions of cells
Silicon detectors have a range of useful applications in the proton radiation science. During the last decade detector systems have been developed mainly for high energy physics which are characterized by low noise readout electronics, high spatial resolution, and a wide dynamic range of energy measurement capabilities.
In certain applications useful in our field, silicon based detectors offer significant advantages over gas-based detectors, particularly with regard to spatial resolution and simultaneous measurement of position, energy, and angle of individual particles.
Another advantage is that silicon detectors work under conditions compatible with the environment of living cells.
In what follows, I will show how these advantages can be utilized in applications in radiation medicine and radiobiology of protons.

10. Applications of Silicon Detectors
Proton Treatment planning
Proton radiography
Proton computed tomography (CT)
Proton Radiobiology
Particle microscope
Nanodosimetry

11. Proton Treatment Planning
Acquisition of imaging data (CT, MRI)
Conversion of CT values into stopping power
Delineation of regions of interest
Selection of proton beam directions
Design of each beam
Optimization of the plan
Proton therapy requires three-dimensional planning, which begins with the acquisition of imaging data.
For proton therapy, the position an density of each heterodense region on the beam path must be defined. Presently, CT is the only way to obtain these data, although MRI can assist greatly in the definition of target and normal tissue boundaries.
CT data are acquired with filtered kVp x-rays an represent linear attenuation coefficients of these x-rays. These data then have to be converted into proton stopping powers in order to calculate the proton energy and range required to reach the target.
The physicist or dosimetrist developing the plan typically selects 2-4 beams per target to achieve the desired dose distribution. Target volumes often have a close geometrical relationship to critical normal tissues, which suggests a few optimal beam angles to minimize normal tissue doses.
Having selected a beam direction, a field-defining aperture is designed taking into account the lateral beam falloff, target motion, and set-up uncertainties. Each beam is designed to penetrate the body to a specified depth.
After dose calculation, individual beam parameter may still be changed until the plan is optimized.

12. Computed Tomography (CT)
Faithful reconstruction of patient’s anatomy
Stacked 2D maps of linear X-ray attenuation
Electron density relative to water can be derived
Calibration curve relates CT numbers to relative proton stopping power
Computed tomography (CT) images of the patient provide the basic data for the treatment planning program.
CT data are used to define the position of the target and of critical normal tissues and to quantify inhomogeneities in the path of each proton beam.
The CT images provide 2D photon attenuation data which can be used to derive electron density information relative to water. For proton therapy, these electron density values have to be converted to stopping power relative to water.
X-ray tube
Detector array

13. Processing of Imaging Data
SP = dE/dxtissue /dE/dxwater
H = 1000 mtissue /mwater
Relative proton stopping power (SP)
CT Hounsfield values (H)
Calibration curve H SP
A CT image represents a spatial distribution of photon attenuation coefficients relative to water, which after multiplication with a factor of 1000 are called Hounsfield values.
These Hounsfield (H) values have to be converted into relative stopping power (relative to water) in order to calculate a proton treatment plan.
This conversion is accomplished by means of a calibration curve.
Dose calculation
Isodose distribution

14. CT Calibration Curve Stoichiometric Method*
Step 1: Parameterization of H
Choose tissue substitutes
Obtain best-fitting parameters A, B, C
H = Nerel {A (ZPE)3.6 + B (Zcoh)1.9 + C}
The probably most accurate way of converting Hounsfield values into stopping powers is the stoichiometric method developed by Uwe Schneider at PSI.
In the first step of this method, one chooses some tissue substitutes with known chemical composition an density. CT measured Hounsfield values of these samples are then used to parameterize the response function of the CT scanner.
The parametric formula relating H to the Z values of the substitute takes into account the three principle processes of photon attenuation: photoelectric effect, coherent scattering, and Compton scattering.
Rel. electron density
Photo electric effect
Coherent scattering
Klein-Nishina cross section
*Schneider U. (1996), “The calibraion of CT Hounsfield units for radiotherapy treatment planning,” Phys. Med. Biol. 47, 487.

15. CT Calibration Curve Stoichiometric Method
Step 2: Define Calibration Curve
select different standard tissues with known composition (e.g., ICRP)
calculate H using parametric equation for each tissue
calculate SP using Bethe Bloch equation
fit linear segments through data points
Fat
In the second step of the stoichiometric method, the parameterized CT response curve is used for the computation of Hounsfield values for a set of standard tissues with known atomic composition. Such data are available from ICRP report 23.
Relative stopping power values of the standard tissues are calculated using the Bethe Block equation for a representative energy.
The resulting point pairs (H, SP) are fitted by linear segments, which results in the desired calibration curve.

16. Problems with the Current Method
Proton interaction  Photon interaction
Multi-segmental calibration curve required
No unique SP values for soft tissue Hounsfield range
Tissue substitutes  real tissues
Uncertainty requires larger range to cover tumor
Risk for sensitive structures
Even with this elaborate method there are a number of problems, which lead to range uncertainties in CT based proton treatment planning.
Proton interactions with matter are quite different from those of photons. While the energy deposition of photons has a strong dependence on atomic number (Z), the energy deposition of protons is rather independent of the chemical composition of the tissue and depends mainly on density.
This leads to some ambiguity of proton SP values over the soft tissue Hounsfield range, in particular when fat tissue is involved.
Furthermore, the chemical composition of tissue substitutes is often quite different from that of real tissues.
The uncertainty of proton range often forces the treatment planner to increase the planned range of protons ensure tumor coverage, which could mean an increased risk for sensitive structures.

17. Proton Transmission Radiography - PTR
MWPC 2
MWPC 1 SC p
Energy detector
First suggested by Wilson (1946)
Images contain residual energy/range information of individual protons
Resolution limited by multiple Coulomb scattering
Spatial resolution of 1mm possible
As one way to measure range uncertainties

18. Proton Range Uncertainties
Alderson Head Phantom
Range Uncertainties
(measured with PTR)
> 5 mm
> 10 mm
> 15 mm
Schneider U. (1994), “Proton radiography as a tool for quality control in proton therapy,” Med Phys. 22, 353.

19. Proton Beam Computed Tomography
Proton CT for diagnosis
first studied for diagnostic use during the 1970s
dose advantage over x rays for similar resolution
not further developed after development of x-ray CT
Proton CT for treatment planning and delivery
renewed interest during the 1990s (2 Ph.D. theses)
fast data acquisition and proton gantries available
further R&D needed

20. Proton Beam Computed Tomography
Applications
Precise calculation of dose distributions
3D verification of dose patient treatment position
tumor delineation without need of contrast media

21. Proton Beam Computed Tomography
SSD 3
SSD 1
SSD 2
SSD 4 ED
Conceptual design
single particle resolution
3D track reconstruction
Si microstrip detectors
p cone beam geometry
multiple beam directions
energy loss measurement
analysis of scattering and nuclear interactions
p cone beam
Trigger logic
DAQ

22. Development of Proton Beam Computed Tomography
Experimental Study
two detector planes
water phantom on turntable
Theoretical Study
GEANT MC simulation
influence of MCS and range straggling
importance of angular measurements

23. Applications of Silicon Detectors
Proton Treatment Planning
Proton radiography
Proton computed tomography (CT)
Proton Radiobiology
Particle microscope
Nanodosimetry

24. Proton Radiobiology in Perspective
D = 1 Gy
10 mm
n = 112
10 MeV protons
n = 54
4 MeV protons
n = 416
50 MeV protons
dE/dx per mm
4.7 keV
134 ionizations
10 keV
276 ionizations
1.3 keV
36 ionizations
RBE*
1.4
2.0
1.1
* rel to 60 Co g rays

25. Study of Cellular Radiation Responses
in vitro (in glass ware):
single cell suspension seeded in culture flasks or Petri dishes
immortalized cell lines
exponential or stationary phase
in vivo (in a living organism):
tumor growth in animals
normal tissue response in animals (e.g., crypt cells)
response of microscopic animals (e.g., nematodes)

26. Study of Cell Survival in vitro
seeded cells are incubated for days
‘surviving cells’ form large colonies (> 50 cells)
surviving fraction is defined as
plating efficiency (PE) is defined as the fraction (%) of cells in an unirradiated culture that form colonies
Dose S
0.1 -
0.01 -
1 -

27. Applications of Silicon Detectors
Proton Treatment Planning
Proton radiography
Proton computed tomography (CT)
Proton Radiobiology
Particle microscope
Nanodosimetry

28. Particle Tracking Silicon Microscope
Conventional radiobiological experiment
random traversal of cells by a broad particle beam
only average number of hits per cell is known
Particle-tracking radiobiological experiment
number of particles per cell is exactly known
broad beam or microbeam setup
l = P(n) = ln/n! e-l

29. Particle Tracking Silicon Microscope
Conceptual design
biological targets located on detector surface
single-particle tracking
energy or LET measurement
ASIC and controller design adapted to application
dedicated data acquisition system

30. Low-Dose Cell Survival
Dose (Gy)
3.2 MeV protons
Low-dose studies with a proton microbeam
precise low-dose/fluence cell survival curves
hypersensitive region at low doses
more pronounced at higher proton energies (3.2 MeV vs. 1 MeV)
Schettino et al. Radiation Res. 156, , 2001

31. Adaptive Response & Bystander Effect
Low-dose studies with an alpha particle microbeam
only 10% of cells exposed
more cells inactivated than traversed (bystander effect)
previous exposure to low level of DNA damage increases resistance (adaptive response)
--- expected
-o- 6 hrs after priming g dose
-- 6 hrs after priming g dose
Sawant et al. Radiation Res. 156, , 2001

32. Goals of the LLU/SCIPP Particle Tracking Microscope Project
Develop a versatile and inexpensive broad-beam and microbeam particle tracking system for
protons and alpha particles
wide range of energies (1 MeV - 70 MeV protons)
in vitro and in vivo radiobiological studies
research studies for radiation therapy and protection
support of DOE and NASA low-dose research programs

33. Applications of Silicon Detectors
Proton Treatment Planning
Proton radiography
Proton computed tomography (CT)
Proton Radiobiology
Particle microscope
Nanodosimetry

34. Nanodosimetry Collaboration
Loma Linda University Medical Center (1997)
Weizmann Institute of Science (1997)
UCSD (1998)
UCSC (2000) SCIPP

35. Nanodosimetry Concepts
DNA is the principle target in radiobiology
Radiation interaction with DNA is a stochastic event
Single damages (break or base oxidation) are easily repaired
Clustered damages are difficult or impossible to repair
Clustered
damage
irreparable
Single damage
reparable
charged
particle ~2 nm d
electron
50 base pair DNA segment

36. Conceptual Approaches
Track Structure Imaging
Single-Volume Sampling
This is the proton treatment plan of a 23-year-old male with recurrent hemangiopericytoma of the cervical spine, status post two attempts of resection. The treatment plan forsees to give a dose of GyE to the site of residual disease using only proton therapy. The spinal cord will receive approximately 40 GyE to the center and up to 60 GyE at the cord surface.

37. Mean Free Path versus Gas Pressure
n, targets per unit volume
s, interaction cross section
Assumptions:
same atomic composition
s is independent of density
Density Scaling:
1 Torr Propane (C3H8)
l = 1 / (n s) l
target
projectile
1 mm in 1 Torr propane  2.4 nm in unit density material
lgas = (rref / rgas) lref

38. Ion Counting Nanodosimetry
trigger E1
Ion counter SV
ions
gas E2
aperture
DAQ
vacuum
particle
SSD
Ionization volume filled with low-pressure gas
single particle detection
ion drift through aperture
wall-less sensitive volume
evacuated ion detection volume

39. The Ion Counting Nanodosimeter
Anode
50mm
Cathode E1 E2 e-
ion
1 Torr
Intermediate
vacuum
particle EM
to pump 2
to pump 1
High vacuum
Pulsed drift field
differential pumping system
electron multiplier
internal alpha source

40. Single Charge Counting
trigger E1
Ion counter SV
ions
gas E2
aperture
Particle
detector
DAQ
vacuum
Ionization volume filled with low-pressure gas
single particle detection
ion drift through aperture
wall-less sensitive volume
evacuated ion detection volume

41. The Ion Counting Nanodosimeter
Anode
SSD1
SSD2
Pulsed drift field
differential pumping system
electron multiplier
four SS detector planes for particle tracking and energy reconstruction
1 Torr E1
50mm e-
particle
ion
Cathode E2
Intermediate
vacuum
High vacuum EM
to pump 1
to pump 2

42. Nanodosimetric Spectra
ND spectra change with particle type and energy
average cluster size increases with increasing LET
protons
alpha
carbon

43. Applications New Standard of Radiation Quality in Mixed Fields
Radiation Treatment Planning: biological weighting factor
Radiation Protection: risk-related weighting factors
Manned Space Flight: Risk prediction (cancer & inherited diseases)

44. Acknowledgements LLUMC WIS UCSD - Radiobiology UCSC - SCIPP
Vladimir Bashkirov
George Coutrakon
Pete Koss
WIS
Amos Breskin
Rachel Chechik
Sergei Shchemelinin
Guy Garty
Itzik Orion
Bernd Grosswendt - PTB
UCSD - Radiobiology
John Ward
Jamie Milligan
Joe Aguilera
UCSC - SCIPP
Abe Seiden
Hartmut Sadrozinsky
Brian Keeney
Wilko Kroeger
Patrick Spradlin
The nanodosimetry project has been funded by the National Medical Technology Testbed (NMTB) and the US Army under the U.S. Department of the Army Medical Research Acquisition Activity, Cooperative Agreement # DAMD The views and conclusions contained in this presentation are those of the presenter and do not necessarily reflect the position or the policy of the U.S. Army or NMTB.