Models and Mechanisms Connecting Physics
and
d Bi
Biology
l
att Multiple
M lti l Scales
S l in
i the
th Biological
Bi l i l
Hierarchy
Help Me!
Robert D. Stewart, Ph.D.
Associate Professor of Radiation Oncology
University of Washington School of Medicine
Department of Radiation Oncology
1959 NE Pacific Street
Seattle, WA 98195-6043
206-598-7951 office
206-598-6218 fax
trawets@uw.edu
Presented at 2014 AAPM Meeting, Austin, TX
Session: TU-A-BRE-3 (last number indicates order within session)
Date and Time: Tuesday July 22, 2014 from 7:30 to 9:30 am
Location: Ballroom E
Abstract #23482
© University of Washington Department of Radiation Oncology
© University of Washington Department of Radiation Oncology
Learning Objectives
 Review models and mechanisms connecting radiation
biology at the molecular and cellular levels to
radiation biology at the tumor and tissue level
• Focus on effects of particle linear energy transfer (LET)
 Is
I the
h RBE for
f DNA d
damage useful
f l ffor predicting
di i cell
ll
survival?
 Is the RBE for cell survival useful for predicting the
RBE for clinical endpoints?
endpoints?
Slide 2
© University of Washington Department of Radiation Oncology
Slide 3
Spatial Pattern of Energy Deposits on the
Molecular and Cellular Levels (“Track Structure”)
10--15 μm
10
Image adapted from Muroya Y, Plante I, Azzam EI, Meesungnoen J, Katsumura Y, Jay-Gerin JP. High-LET ion radiolysis of water: visualization of the
formation and evolution of ion tracks and relevance to the radiation-induced bystander effect. Radiat Res. 165(4), 485-491 (2006).
Slide 4
© University of Washington Department of Radiation Oncology
Tracks formed by ions in water (70 keV/μm)
S
Structure
off the tracks produced
by particles with the same LET
are not q
quite the same and can
produce different biological effects
However if we “zoom out” to the
macroscale (> 0.1 to 1 mm),
mm), the
tracks of even very high LET
particles look quite similar
RBE effects must arise from
the cellular and subcellular
features of tracks – even for
clinical endpoints!
Image adapted from Muroya Y, Plante I, Azzam EI, Meesungnoen J, Katsumura Y, Jay-Gerin JP. High-LET ion radiolysis of water: visualization of the
formation and evolution of ion tracks and relevance to the radiation-induced bystander effect. Radiat Res. 165(4), 485-491 (2006).
Slide 5
© University of Washington Department of Radiation Oncology
Initial Damage to a Critical Molecule
Absorbed Dose
Chemical
s Repair
O2 fixation
Acute
Ionization
Excitation
Radiation
10-18 to 10-10 s
Cluster of damaged
nucleotides
(“lesions
(“
lesions”)
”) formed by
passage of charge
particle by DNA
10-3
hypoxia
10-6 s
~ 2 nm
Spatial pattern energy
deposits determines local
complexity
l i off the
h cluster
l
Overall, a 1 Gy dose damages about 1 in 106 nucleotides.
Slide 6
© University of Washington Department of Radiation Oncology
A Paradigm to Connect DNA Damage to Local
Tumor Control
U
Unrepairable
i bl
Damage
DNA Damage
Lethal after
biological processing
No
Damage
Tumor Cell
Survival
> 1 survive
Failure
Incorrectly
repaired Damage
Black Lines: potential molecular and
cellular pathways (mechanisms)
(
)
Control
Reproductive
Death
None
survive
Black Dashed Lines: transition from cell to tissuetissue-level biology
Slide 7
© University of Washington Department of Radiation Oncology
Reproductive Death?
Reproductive death is a general term that encompasses all
modes of cell death, including cells that remain metabolically
active and intact but unable to divide…
apoptosis
cell
necrosis
mitotic
catastrophe
Reproductive
Death
Delayed
y cell death
(days or weeks later)
later)
Permanent or qquasiquasi-ppermanent ((> 1010-14 days)
y)
loss of reproductive potential
Slide 8
© University of Washington Department of Radiation Oncology
Clusters of DNA lesions
Groups
p of several DNA lesions within one or two turns of the DNA are
termed a cluster or multiply damaged site (MDS)
(MDS)**
Organic base
Sugar-phosphate backbone
Undamaged DNA segment (20 bp)
(a) Base damage in opposed strands
(c) complex SSB with adjacent base damage
(b) complex SSB with opposed base damage
(d) Complex DSB
Most critical category of DNA Damage
* Clustered lesions are also referred to as locally multiply damaged sites (LMDS)
Slide 9
© University of Washington Department of Radiation Oncology
Are all DSB Lethal? What about SSB?
After 1 Gy dose of low LET radiation, a typical human cell sustains
45 + 10 DSB Gy-1 cell-1 and 1000 + 200 SSB Gy-1 cell-1. If all DSB are
lethal, the fraction of cells that will survive a 2 Gy dose is
S = exp( −45 DSB Gy −1 ⋅ 2 Gy)  10 −40 (10-31 ,10-48 )
Onlyy those cells that do not sustain a radiation
radiation--induced DSB survive
(Poisson distribution of DSB among irradiated cells))
For comparisons, many published studies indicate a surviving
fraction of 0.1 (repair
(
compromised)) to 0.9 (repair
(
proficient)) cells
after a 2 Gy dose of radiation. Only way to reconcile observations is
< 2% of initial DSB formed
in a cell are lethal
and/or
< 0.1% of initial SSB formed
in a cell are lethal
Cells are really good at repairing DNA damage, even DSB!
See also the classic review: DT Goodhead. Initial events in the cellular effects of ionizing radiations: clustered damage in DNA. IJRB 65(1): 7-17 (1994).
Slide 10
© University of Washington Department of Radiation Oncology
RBE for DSB induction
Σ = number
b off DSB G
Gy-1 cell
ll-1
= slope of line
Dp = dose of proton
DSB
B Gbp-1
Dγ = dose
d
off reference
f
radiation
di ti
human skin fibroblasts
23 keV
keV//μm
0.2 keV
keV//μm
Want:
Number DSB = ΣγDγ = ΣpDp
RBE DSB ≡
Dγ
Dp
=
Σp
Σγ
≅ 1.5
Dose (Gy)
RBE is an example of an
“isoeffect calculation”
Measured data from Frankenberg D, Brede HJ, Schrewe UJ, Steinmetz C, Frankenberg-Schwager M, Kasten G, Pralle E. Induction of DNA double-strand breaks by
1H and 4He ions in primary human skin fibroblasts in the LET range of 8 to 124 keV/microm. Radiat Res. 151
151(5), 540-549 (1999).
Slide 11
© University of Washington Department of Radiation Oncology
Trends in RBEDSB with proton LET
2.8
Filled Yellow Symbols: Track
T k Structure
St t
Simulation (Nikjoo et al. 1997, 2001, 2002))
2.4
2.2
Simulation (Friedland et al. 2003))
2.0
Lines: Monte Carlo Damage Simulation*
Simulation*
1.8
RB
BE
Filled Red Symbols:
y
Track Structure
(MCDS)
DSB
2.6
1.7 + 0.1 (5.9%)
1.6
1.4
1.2
DSB are only category of DNA
damage that increases with
increasing particle LET (additional
evidence SSB less critical form of DNA
damage than DSB)
1.1 + 0.04 (3.6%)
1.0
SSB
0.8
0.6
Base Damage
0.4
0.2
0
5
10
15
20
25
30
35
40
LET (keV/μm)
* Semenenko and Stewart 2004, 2006, Stewart et al. 2011
45
50
55
60
65
Slide 12
© University of Washington Department of Radiation Oncology
Why are DSB so effective at killing cells?
(Breakage and Rejoining Theory)
Break-ends
DSB
DSB
DSB
Break-ends associated with one DSB incorrectly
Breakrejoined to breakbreak-end associated with a different DSB
Proximity Effects: pairs of DSB formed in close spatial and temporal proximity are
more likely to rejoin incorrectly than pairs of DSB separated in time and\
and\or space
(dose rate and LET effects)
effects)
R.K. Sachs and D.J. Brenner, Chromosome Aberrations Produced by Ionizing Radiation: Quantitative Studies
http://web.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC&rid=mono_002
Slide 13
© University of Washington Department of Radiation Oncology
Lethal and NonNon-Lethal Aberrations
Correct rejoining
Symmetric (reciprocal
(reciprocal)) translocation
Dicentric
Centromere
Acentric fragment
Dicentrics and acentric fragments are usually lethal in the reproductive sense because
segregation of chromosomes at mitosis is disturbed. In contrast, correct DSB
j
g and symmetric
y
(
(reciprocal)
p
) translocations are consistent with continued cell
rejoining
division
R.K. Sachs and D.J. Brenner, Chromosome Aberrations Produced by Ionizing Radiation: Quantitative Studies
http://web.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC&rid=mono_002.TOC&depth=10
Slide 14
© University of Washington Department of Radiation Oncology
Is there a 1:1 relationship?
AC 1522 normal human fibroblasts
irradiated by x-rays
S = e-Y
S = fraction that survive
Y = avg number of lethal
aberrations per cell
Source: Cornforth and Bedford, Rad. Res., 111, p 385-405 (1987). See also Figure 3.4 in Hall (p. 37)
Slide 15
© University of Washington Department of Radiation Oncology
Linear-Quadratic (LQ) Model for Cell Survival
Pairs of DSB formed by
same track interact
(intra-track interactions))

D 
Y ( D) = α D + β D = α D 1 +

 α /β 
2
P i off DSB fformed
Pairs
db
by
different tracks interact
(inter-track effect))
Cell survival after dose D
S ( D) = e
−Y ( D )
=e
−α D − β D 2
Only those cells without a lethal aberration in their DNA retain the
ability to divide and produce viable progeny (“reproductive
(“
survival”).
”).
Figure 3.5 in EJ Hall and AJGiaccia, Radiobiology for the Radiologist, 6th Ed, Lippincott Williams & Wilkins (2006)
Slide 16
© University of Washington Department of Radiation Oncology
Connecting DSB to Local Tumor Control
Grey Dashed Lines: low probability molecular and cellular pathways (mechanisms)
(
)
Double Strand
Break (DSB)
No DSB
“sub--lethal damage”
“sub
Tumor Cell
Survival
> 1 survive
No lethal
aberrations
All or most DSB
lethal
Lethal after biological
processing
Chromosome
Aberrations
Black Lines: high probability
molecular and cellular pathways
(mechanisms))
Reproductive
Death
None
Are trends
in
the
RBE
for
DSB
induction
qualitatively
and
Failure
Control
survive
quantitatively
tit ti l similar
i il to
t the
th RBE for
f cell
ll survival?
i l?
Black Dashed Lines: transition from cell to tissuetissue-level biology
Slide 17
© University of Washington Department of Radiation Oncology
Low and High Dose RBE (cell
(
survival))
Surviving Fraction
100
200/250 kVp x-rays (1.3-1.5 keV/μm)
28.6 MeV 4He2+ (24.9 keV/μm)
25 MeV 4He2+ (26.3 keV/μm)
5.2 MeV 4He2+ (89 keV/μm)
RBE for
f a specific
ifi dose
d
((cell
ll survival
i l
level))
RBE =
10-1
≅
10-2
Dγ
Dp
9.2 Gy
= 3.3 (1% survival)
28G
2.8
Gy
Low Dose RBE: -lnS
lnS ≅ (αD)γ = (αD)p
αp
low dose RBESF =
=
Dp αγ
Dγ
10-3
Dp
“RBEmax”
Dγ
10-4
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
Absorbed Dose (Gy)
αD dominates
(D << α/β))
High Dose RBE: -lnS
lnS ≅ (βD2)γ = (βD2)p
βp
high dose RBESF =
=
Dp
βγ
Dγ
βD2 dominates
(D >> α/β))
“RBEmin”
Slide 18
© University of Washington Department of Radiation Oncology
Is RBEDSB predictive of RBESF?
14
1
10
1
+
1.9 MeV H
1 +
1.15 MeV H
0.76 MeV 1H+
4
2+
3.8 MeV He
+
DSB
4
2+
DSB
SF
SF
-1
9
βp
high dose RBESF =
βγ dose
3X jump in low
8
250 kVp x-ray
(1.46 mm Cu))
RBE
Surviving
g Fraction
100
Fit13LQ survival
i (lHmodel
t measured
d data
d t
RBE
) d l to
RBE low
( He and
)
and
high dose RBE.
12 compute
Low Dose RBE
11
αp
High Dose RBE
low
dose
RBE
=
SF
10
αγ
10-2
7
6
RBE (24
(24→
→32 keV
keV//μm)?
5
Compare
to MCDS*
MCDS* estimate of RBEDSB
4
as function
of LET.
10-3
3
2
1
RBE relative to 250 kVp x-rays
0
10-44
0
1
2
3
4
5
6
7
8
9
Absorbed Dose (Gy)
10
11
12
0
10
20
30
40
50
60
70
80
90
100 110 120
LET (keV/μm)
Little if any evidence RBESF (1H+ and 4He2+) is related to
RBEDSB, or so it seems…
Measured data from Prise et al. IJRB, 58, p 261-277 (1990)
* Semenenko and Stewart 2004, 2006, Stewart et al. 2011
Slide 19
© University of Washington Department of Radiation Oncology
A Mechanistic Model for α and α/β
The Repair
Repair--Misrepair
Misrepair--Rixation (RMF) model (Carlson
(
et al. 2008))
predicts, in the limit when the D is small compared to α/β
α/β,, that
Intra-track chromsomal
Intraaberrations
i
α = θΣ + κ zFΣ
2
Unrepairable and
misrepaired
β=
κ
2
Σ
2
IInter-track
Interk
aberrations
α 2
= (θ / κ ) + 2zF
β Σ
θ, κ are adjustable cell- or tissue-specific parameters related to biological
processing
i off DNA d
damage (i
(independent
d
d off LET andd O2 concentration)
i )
Σ is the number of DSB Gy-1 Gbp-1 (or per cell); estimate using the MCDS (strong
function of LET and O2 concentration)
zF is the frequency-mean specific energy (in Gy) delivered to the cell nucleus
(strong function of LET but independent of O2 concentration) – estimate with the
MCDS or other Monte Carlo code(s)
D.J. Carlson, R.D. Stewart, V.A. Semenenko and G.A. Sandison, Combined use of Monte Carlo DNA damage simulations and deterministic repair models
to examine putative mechanisms of cell killing. Rad. Res. 169, 447-459 (2008)
Slide 20
© University of Washington Department of Radiation Oncology
How is RBEDSB related to RBESF?
With the RMFRMF-motivated formulas for α and β, the low and
high dose RBESF is
DSB Gy-1 Gbp-1
β=
α = θΣ + κ zF Σ
2
“RBEmin”
“RBEmax”
κ
2
βp
high dose RBESF =
= RBEDSB
βγ
low dose RBESF
Σ
2
DSB Gy-1 Gbp-1
reference radiation
αp
z F Σγ

=
≅ RBEdsb 1 + RBEdsb
αγ
θ /κ

D is “small
“small”
” compared to α/β
Intra-track DSB interactions increase with
Intraincreasing LET because of proximity effects
z F Σγ
θ /κ
∝
Σγ
θ /κ
⋅ LET

 ≥ RBEdsb

Slide 21
© University of Washington Department of Radiation Oncology
Is RBEDSB predictive of RBESF? Version 2.0
3.5
Si lt
Simultaneous
fit to
t data
d t ffor all
ll particles
ti l
and3.0energies using RMFRMF-motivated
1H+ -2
formulas (θ
( = 3.7×10
θ/κ = 142.3))
1.9 MeV 1H+
1.15 MeV 1H+
1 +
0.76 MeV H
3.8 MeV 4He2+
25
2.5
10-11
60Co
α = θΣ + κ zFΣ
γ-ray
β=
2
2.0
10-22
Σ2
2
low dose RBESF
RBEDSB
1.5
250 kVp x-ray
(1.46 mm Cu))
κ
4He2+
RBE
R
Survivin
ng Fraction
100
Low Dose RBESF
Used MCDS*
MCDS* to estimate ΣHigh
, mean
Dose RBE
1.0
specific
p
energy,
gy, and RBEDSB as function
Dotted vs aolid lines differ because of
of0.5LET.
SF
10-3
intra--track DSB interactions
intra
RBE relative to 60Co γ-rays
10-4
0.0
0
1
2
3
4
5
6
7
8
9
Absorbed Dose (Gy)
10
11
12
0
10
20
30
40
50
60
70
80
90
100 110 120
LET (keV/μm)
Reasonable fit to cell survival data for all energies.
Low dose RBESF > RBEDSB
Measured data from Prise et al. IJRB, 58, p 261-277 (1990)
* Semenenko and Stewart 2004, 2006, Stewart et al. 2011
© University of Washington Department of Radiation Oncology
Slide 22
Dosimetry of Short
Short--Range Particles is Tricky…
When the CSDA range of a charge particle is of the same
order of magnitude as the dimensions of the biological target,
dosimetry needs to be corrected for



Change in stopping power within target
Energy and path length straggling
Fi it particle
Finite
ti l range (“stoppers”),
(“ t
”) energy and
d angular
l
distribution of particles incident on target
For a monoenergetic particle incident on a 5 μm target
Slide 23
© University of Washington Department of Radiation Oncology
Fit with “Corrected” Dosimetry
1
1
0
1.9 MeV H
1.15 MeV 1H+
0.76 MeV 1H+
3.8 MeV 4He2+
10-2
γ-ray
250 kVp x-ray
(1.46 mm Cu))
4.7%
10.7%
10-3
10
1
+
2+
-1
10-22
31.4%
normoxic
+
10-3
7.4%
-4
1
4
10-1
60Co
+
1 9 MeV H
Non10-linear
N
Nonli
regression
i analysis
l i off1.9
survival
i l
1.15 MeV H
data with an adjustable DSF for0.76
each
MeV H
MeV
He
-2 and
particle type and energy (θ
( =3.623.810
θ/κ 10= 151
151.2)
2))
+
Survivin
ng Fraction
Survivin
ng Fraction
100
Same DSF and θ, θ/κ
θ/κ)) Used
MCDS to estimate RBEDSB
hypoxic
-4
0
1
2
3
4
5
6
7
8
9
Absorbed Dose (Gy)
10
11
12
10
0
2
4
6
8
10
12
14
16
18
20
22
Absorbed Dose (Gy)
Improved fit to measured data with small (quite
(
plausible)) changes
in the mean particle energy (<
(< 10 μm shift)
shift)
Measured data from Prise et al. IJRB, 58, p 261-277 (1990)
24
Slide 24
© University of Washington Department of Radiation Oncology
Is RBEDSB predictive of RBESF? Version 2.1
3.5
9.5
H
Hypoxic
i
proton
90
9.0
8.5
3.0
8.0
7.5
7.0
25
2.5
6.5
6.0
RBE
R
Yes, seems so
for V79 cells
1.5
5.5
RBE
R
2.0
5.0
45
4.5
4.0
Normoxic
proton
3.5
1
+
RBEDSB( H )
1.0
2.5
RBEDSB( He )
low dose RBESF
2.0
Low Dose RBESF
1.5
RBEDSB
High Dose RBESF
1.0
4
0.5
normoxic
3.0
2+
Low Dose RBESF
0.0
High Dose RBESF
hypoxic
0.5
0.0
0
10
20
30
40
50
60
70
80
LET (keV/μm)
90
100 110 120
0
10
20
30
40
50
60
70
80
90
100 110 120
LET (keV/μm)
For p
protons with an LET < 20 keV
keV//μm ((>
> 2 MeV),
MeV), RBE is about the
same in cells irradiated under normoxic and anoxic conditions (no
(
change in OER from 60Co γ-rays).
).
Slide 25
© University of Washington Department of Radiation Oncology
Impact of Uncertainties on “observed” RBE
1.7
RBE iis ratio
ti off d
doses th
thatt
produce same biological effect
1.6
Obse
erved RBE
E Range
1.5
1.4
RBE =
1.3
1.2
Dγ ± σ γ
Dp ± σ p
1.1
RBE 1.1 + 0.1 (blue
(
shaded region))
1.0
1.8% uncertainty in equivalent
physical dose: RBE = 1.1 + 0.05
0.9
0.8
3.8% uncertainty in equivalent
physical dose: RBE = 1.1 + 0.1
0.7
0.6
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15
1-σ uncertainty (%) in biologically equivalent
photon and proton absorbed dose
10% uncertainty in equivalent
physical
h i l dose:
d
RBE = 1.1
1 1 + 0.3
03
Slide 26
© University of Washington Department of Radiation Oncology
Do we just need more accurate dosimetry
and better dosedose-response models
(e.g., RMF model)?
)?
N d + 3.8%
Need
3 8% accuracy iin equivalent
i l t doses
d
for
f RBE = 1.1
1 1 + 0.1
01
Adapted from Figure 2 in Paganetti, Niemierko, Ancukiewicz
Ancukiewicz,, Gerweck,
Gerweck, Goitein,
Goitein, Loeffler,
Loeffler, and Suit, Relative Biological Effectiveness (RBE)
Values for Proton Beam Therapy, IJROBP 53(2) 407407-421 (2002).
Slide 27
© University of Washington Department of Radiation Oncology
Strong evidence for spatial (and
(
LET)) variations
in proton RBE despite uncertainties …
Open circles (dose < 4 Gy)
Filled circles (dose > 4 Gy)
Low energy proton “sting
“sting”” on
distal edge of a pristine Bragg peak
26.9 keV
keV//μm
0.02 mm
7.9 keV
keV//μm
0.4 mm
4.5 keV
keV//μm
1.2 mm
RBE 1.1 + 0.1
in vitro cell survival
Adapted from Figure 3 in Paganetti, Niemierko, Ancukiewicz
Ancukiewicz,, Gerweck,
Gerweck, Goitein,
Goitein, Loeffler,
Loeffler, and Suit, Relative Biological Effectiveness (RBE)
Values for Proton Beam Therapy, IJROBP 53(2) 407407-421 (2002).
Slide 28
© University of Washington Department of Radiation Oncology
RBE effects in a 163 MeV pencil beam
Tuned MCNP 6.1 model of 163.25 MeV pencil beam to match
measured depth
depth--dose profile (SCCA
(
proton facility, Seattle, WA).
).
2.4
110
100
2.2
measured
MCNP
90
2.0
80
18
1.8
70
1.6
60
RBE
Relattive Dose
Low Dose RBESF
(α/β = 1 Gy)
Gy)
50
1.4
40
1.2
30
1.0
20
RBE
e to protons
R ddue
ReRe
-ran
MCNP with
i h RBE
slowing down below
weighted
dose tally*
tally
*
RBEDSB
~ 13
MeV (> 3.5(F6)
keV/μm)
RBE > 1.1
2 mm
0.8
10
0.6
0
0
5
10
15
Depth in Water (cm)
20
25
15 0
15.0
15 5
15.5
16 0
16.0
16 5
16.5
17 0
17.0
17 5
17.5
18 0
18.0
18 5
18.5
19 0
19.0
19 5
19.5
Depth in Water (cm)
* Really easy way to generate dose-averaged RBE in MCNP. See supplemental slides.
20 0
20.0
Slide 29
© University of Washington Department of Radiation Oncology
Proton RBE for Healthy Organs and Tissues?
Chemical
10-3 s Repair
Absorbed Dose
O2 fixation
Ionization
Excitation
Radiation
Correct
Repair
1 Gy ~ 1 in 106
102 s
Acute
hypoxia
Enzymatic Repair
(BER, NER, NHEJ, …))
10-6 s
10-18 to 10-10 s
Chronic
hypoxia
(> 11--2 h)
DNA d
damage
Local Control
Early Effects
(erythema, …)
Late Effects
fibrosis …)
(fibrosis,
106 s
Loss of Function
and Remodeling
108 s
108 s
s
Heritable
Effects
Chronic
hypoxia
(> 44-10 h?)
Incorrect or
Incomplete Repair
104 s
105 s
105 s
Germline
107 s
Viable
Neoplastic
Transformation
105 s
Non--Viable
Non
Inflamatory
Responses
Self renewal and
Differentiation
2nd Cancer
Clonal
Expansion
103
Angiogenesis and
Vasculogenesis
104 s
Somatic
cells
Small-- and largeSmall
large-scale mutations
(point mutations and chromosomal aberrations))
Slide 30
© University of Washington Department of Radiation Oncology
UW Experience with Fast Neutrons
80-90% of absorbed dose to
80patient is from lower energy
protons (E
( avg ≅ 16 MeV))
Tolerance doses derived from
over 25+
25 years off clinical
li i l
experience
Used as a guide for tolerance
doses in carbon ion therapy
What can fast neutron therapy tell us about
RBE effects in a proton Bragg peak?
Slide 31
© University of Washington Department of Radiation Oncology
Neutron TD5/5 Dose (Brainstem
(
and Lung))
20
55
(α/β
α/β))γ ≅ 1.6 Gy
18
50
TD
D for Lung (p
pneumonitis
s)
TD for Brrainstem (ne
ecrosis infarction)
60
45
40
35
Neutron RBE
(clinical endpoint)
30
25
20
15
10
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
5
(α/β
α/β))γ ≅ 10 Gy
16
14
12
10
8
Blue dotted lines RMF predicted
neutron TD with RBEDSB = 2
6
4
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
2
0
0
5
10
15
20
25
30
35
Number of Fractions
40
45
50
5
10
15
20
25
30
35
40
45
Number of Fractions
U l caveats
Usual
t apply
l about
b t accuracy off ttolerance
l
dose
d
estimates
ti t
50
Slide 32
© University of Washington Department of Radiation Oncology
Neutron TD5/5 Dose (Cauda
(
Equina and Cord))
65
60
55
(α/β
α/β))γ ≅ 3.0 Gy
TD for Spina
T
al Cord (my
yelitis necro
osis)
TD for Caud
T
da Equina (nerve dama
age)
70
55
50
45
40
35
30
25
20
15
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
10
5
0
50
(α/β
α/β))γ ≅ 2.2 Gy
45
40
35
30
25
20
15
10
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
5
0
5
10
15
20
25
30
35
Number of Fractions
40
45
50
5
10
15
20
25
30
35
Number of Fractions
40
45
50
Slide 33
© University of Washington Department of Radiation Oncology
Neutron TD5/5 Dose (Skin
(
and Ribs))
60
(α/β
α/β))γ ≅ 9.0 Gy
55
TD for R
Ribs (patho
ological frac
cture)
TD for Skin (ulcerration/necro
osis)
55
60
50
45
40
35
30
25
20
15
10
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
5
(α/β
α/β))γ ≅ 3.0 Gy
50
45
40
35
30
25
20
15
10
Neutron (Laramore 2007)
UW SBRT program (2013)
Emami et al. 1991
5
0
0
5
10
15
20
25
30
35
Number of Fractions
40
45
50
5
10
15
20
25
30
35
Number of Fractions
40
45
50
Slide 34
© University of Washington Department of Radiation Oncology
Fast Neutron RBE for Selected Tissue
3.50
Filled
Fill d circles:
i l RBE(n
RBE(n = 16)
16) = TDγ /TDn
Avg RBE = 2.6 + 0.3
3.25
27
7 tissues/endpoints
ssues/e dpo s
Red Dashed line: Monte Carlo (“first
(“
principles”)
”) simulation of neutron RBE
for DSB induction.
2.75
2.50
2.25
2 00
2.00
RBEDSB (MCNP+MCDS simulation))
1.75
1 50
1.50
Bladder
Femoral Head
Skin (te
elangiesctasia)
Skin (necrosis)
S
Brain
Brainstem
Spinal cord
Cauda equina
Brachial plexus
B
Eye lens I & II (cataract)
Eye retina I & II (blindness)
Ea
ar mid/external
Parotid I & II
Esophagus
Rectum
Liver
T-M joint Mandible
Rib Cage
Colon
Heart
Larrynx (necrosis)
La
arynx (edema)
Lung
Optic Nerve I & II
Chiasma
Stomach
Small intestine
S
Fast Neu
utron RBE
3.00
Blue Shaded Region: Estimate of RBE
for DSB induction derived from analysis
of in vitro cell survival data for 30
human tumor cell lines (Warenius
(
et al.
IJROBP 1994).
1994)).
)
Clinical RBE > RBEDSB, as predicted by
the RMF model
© University of Washington Department of Radiation Oncology
Summary and Conclusions
 Much of the uncertainty in proton RBE due to
• Uncertainty in dosimetry of the reference radiation and proton beam
• Need for mechanistic dose-response models to guide the
interpretation and analysis of measured data
 RBEDSB (RBEmin) and low dose RBESF (RBEmax) are
relevant biological endpoints for (1
(1) local tumor control
and (2) tolerance doses for healthy organs and tissues
 Ample (very
(
strong)) evidence that spatial\
spatial\LET variations
in proton RBE are real and clinically relevant
(exploitable))
• Sticking with a constant RBE = 1.1 is a missed opportunity to
enhance the therapeutic ratio!
Slide 35
© University of Washington Department of Radiation Oncology
Thank You!
S
Supplemental
l
t l Slides…
Slid
 Acknowledgements
 Example of an easy way to setup dosedose-weighted RBE calculations




in MCNP and MCNPX (DE
(
DF modified F6 tally))
Is dose
dose--averaged LET a could surrogate for RBEDSB and
and\\or the
RBESF?
Approximate
pp
formula linear and linear
linear--q
quadratic formulas to
estimate RBEDSB as a function of proton LET
Why does RBEDSB increase with increasing LET and the RBE for
SSB and
d base
b
damage
d
decrease
d
with
i h increasing
i
i LET?
Why are some DSB more lethal than others?
Email: trawets@uw.edu
NOTE: ““trawets
trawets”” = ““stewart
stewart”” spelled backwards.
Slide 36
© University of Washington Department of Radiation Oncology
Acknowledgements
E i UW physics
Entire
h i group but
b t especially
i ll D.
D Argento,
A
t
G. Sandison, C. Bloch, E. Ford, F. Yang; other UW
f
faculty:
J. S
Schwartz, G
G. Laramore, U. Parvathenii
Many Other (non-UW)
(
) Collaborators, especially V.
Semenenko D.
Semenenko,
D Carlson,
Carlson C
C. Kirkby
Kirkby, N
N. Cao
Cao, G
G.
Moffitt, S. Streitmatter, T. Jevremovic, Y. Hsaio,
Hsaio,
VK Y
V.K.
Yu, JJ.H.
H P
Park,
k M
M. Frese,
F
R.
R Rockne,
R k
K
K.
Swanson
Slide 37
Slide 38
© University of Washington Department of Radiation Oncology
Example of D × RBE tally in MCNP
FC1026 RBE-weighted proton (1H+) dose; DSB induction (aerobic)
F1026:H 3
FM1026 0.1602
DE1026 1.000E-03 2.000E-03 3.000E-03 4.000E-03 5.000E-03 6.470E-03
7.500E-03 1.000E-02 2.000E-02 3.000E-02 4.000E-02 5.000E-02
6 000E-02 7
6.000E-02
7.000E-02
000E-02 8
8.000E-02
000E-02 9
9.000E-02
000E-02 1
1.000E-01
000E-01 2
2.000E-01
000E-01
3.000E-01 5.000E-01 9.000E-01 1.000E+00 1.100E+00 1.300E+00
1.500E+00 2.000E+00 2.500E+00 3.000E+00 3.500E+00 4.000E+00
5.000E+00 6.000E+00 7.500E+00 1.000E+01 1.500E+01 2.500E+01
5.000E+01 7.500E+01 1.000E+02 1.500E+02 2.000E+02 2.500E+02
5.000E+02 1.000E+03
DF1026 3.375E+00 3.367E+00 3.368E+00 3.363E+00 3.359E+00 3.352E+00
3.348E+00 3.340E+00 3.317E+00 3.290E+00 3.264E+00 3.242E+00
3.216E+00 3.193E+00 3.168E+00 3.143E+00 3.122E+00 2.889E+00
2 687E+00 2
2.687E+00
2.370E+00
370E+00 1
1.986E+00
986E+00 1
1.916E+00
916E+00 1
1.860E+00
860E+00 1
1.760E+00
760E+00
1.685E+00 1.542E+00 1.451E+00 1.386E+00 1.336E+00 1.297E+00
1.244E+00 1.204E+00 1.164E+00 1.123E+00 1.083E+00 1.051E+00
1.026E+00 1.016E+00 1.012E+00 1.004E+00 1.004E+00 1.003E+00
1.001E+00
1.001E
00 9.995E-01
9.995E 01
KE of proton
RBEDSB from
MCDS 3.10A
Above tally will record D × RBEDSB. Divide by dose (separate tally) to get dose-averaged RBE.
See http://faculty.washington.edu/trawets/
http://faculty washington edu/trawets/ for additional
(downloadable) examples for protons and other particles
Slide 39
© University of Washington Department of Radiation Oncology
Analytic way to estimate an approximate
RBEDSB×Dose for proton beams?
Recall
D ( x ) S ( x ) LET∞ ( x )
=
=
Φ( x)
ρ
ρ
D ( x ) / Φ( x)
LET∞ ( x ) / ρ
∴
=
D ( xr ) / Φ ( xr ) LET∞ ( xr ) / ρ
If we know the LET and dose p
per unit fluence at a reference location
xr, LET at other locations along depth-dose curve computed from
D ( x ) / Φ( x)
LET∞ ( x ) = LET∞ ( xr )
D ( xr ) / Φ ( xr )
RBEDSB ( x ) D ( x ) = ( mLET∞ ( x ) + b ) D ( x )
m = 0.03193 μm/keV and b = 0.98274
Simple formulism to connect patient-specific QA measurements of doseaverage LET to RBEDSB and (hence) the RBESF (via RMF model)?
Slide 40
© University of Washington Department of Radiation Oncology
LETD instead of RBEDSB or RBESF?
6.0
MCDS estimate of RBEDSB
5.5
RBEDSB(LET) = m x LET + b
5.0
RBESF(α/β = 3 Gy)
RBESF(α/β = 10 Gy)
4.5
R
RBE
4.0
3.5
3.0
For a clinically
F
li i ll relevant*
relevant
l
t*range off
t*
proton energies (LET
(
< 25
keV/μm),
μ ), dosedose-averaged
g LET is a
good surrogate for RBEDSB and
the low dose RBESF.
m = 0.03193 RBE
RBE/(
/(keV
keV//μm)
b = 0.98274
2.5
2.0
1.5
1.0
0.5
0
5
10 15 20 25 30 35 40 45 50 55 60 65 70 75
For tumors and\
and\or tissues with a
low α\β, RBESF may be larger
than RBEDSB by 2525-30% larger at
25 keV
k V/
keV/
V/μm.
LET (keV/μm)
* Caveat: + 2 mm of the Bragg peak really low energy protons
(< 1-5 MeV)) contribute in a substantial way to dose and RBE.
Slide 41
© University of Washington Department of Radiation Oncology
LQ fit to proton RBEDSB as function of LET
To better capture trends in the RBE for DSB induction as a
function of LET, a linearlinear-quadratic (LQ) fit is highly
recommended.
3.0
RBEDSB = a ⋅ LET + b ⋅ LET + c
2
2.6
RBE for DS
SB Induction
a = 6.771 × 10-3
b = 2.553
2 553 × 10-22
c = 9.969 × 10-1
2.8
2.4
2.2
20
2.0
1.8
1.6
MCDS
Linear fit to data < 10 keV/mm
LQ fit to data < 75 keV/μm
14
1.4
* Coefficients for fit may need to be adjusted
slightly to correct for uncertainties in proton
LET.
1.2
1.0
0
5
10 15 20 25 30 35 40 45 50 55 60 65 70 75
LET (keV/μm)
Slide 42
© University of Washington Department of Radiation Oncology
Ionization Density and Cluster Complexity
+
=
1.8 to 2.3 nm
Number of DNA lesions per cluster tends to
increase with increasing particle LET.
Slide 43
© University of Washington Department of Radiation Oncology
Why does RBEDSB ↑ and RBESSB and RBEBd ↓
A cluster categorized as a DSB cannot also be categorized
as a (simple
(
or complex)) SSB or as a (simple
(
or complex))
cluster of nucleotides with base damage – mutually
exclusive
l i categories
t
i off DNA damage
d
Cluster = 2 nucleotides with base damage
Cluster = complex SSB
Cluster = complex DSB
Chance a cluster will contain a pair of opposed strand breaks
within about 10 bp increases ((i.e., be a simple or complex DSB))
as LET and the number of DNA lesions per cluster increases.
Slide 44
© University of Washington Department of Radiation Oncology
Why are some DSB (more) lethal and others not?
Answer: we don
don’tt know for sure…
sure
Hypotheses:
(1) Some DSB are unrepairable
nrepairable (unrejoinable)
(
j i bl ) or more slowly
slo l rejoined than
others because of the local (spatial)
(
) complexity of DNA lesions
Simple DSB
Complex DSB
(2) DSB formed in close spatial and temporal proximity to other DSB are more
often mis
mis--rejoined to form chromosome aberrations than DSB separated in time
and/or space (breakage
(
and reunion theory))
(3) Combination of mechanisms (1
(1) and (2
(2 )
RMF model (Carlson
(
et al. 2008)) tends to emphasize mechanism 2
© University of Washington Department of Radiation Oncology
Critical MCDS and RMF Literature Citations







V.A. Semenenko and R.D. Stewart. A fast Monte Carlo algorithm to simulate the spectrum of
DNA damages formed by ionizing radiation. Radiat Res. 161(4), 451-457 (2004)
V.A. Semenenko and R.D. Stewart. Fast Monte Carlo simulation of DNA damage formed by
electrons and light ions. Phys. Med. Biol. 51(7), 1693-1706 (2006)
D.J. Carlson, R.D. Stewart, V.A. Semenenko and G.A. Sandison, Combined use of Monte
Carlo DNA damage simulations and deterministic repair models to examine putative
mechanisms of cell killing. Rad. Res. 169, 447-459 (2008)
R.D. Stewart, V.K. Yu, A.G. Georgakilas, C. Koumenise, J.H. Park, D.J. Carlson, Effects of
Radiation Quality and Oxygen on Clustered DNA Lesions and Cell Death, Radiat. Res. 176,
587-602 (2011). MCDS Version 3.10A developed and tested in this one.
M.C.
C Frese, V.K. Yu, R.D. S
Stewart, D.J. C
Carlson,
l
A Mechanism-Based
h i
d Approach
A
h to Predict
di
the Relative Biological Effectiveness of Protons and Carbon Ions in Radiation Therapy, Int. J.
Radiat. Oncol. Biol. Phys., 83, 442-450 (2012).
• C Kirkby
Kirkby, E Ghasroddashti
Ghasroddashti, Y Poirier
Poirier, M Tambasco
Tambasco, RD Stewart
Stewart, Monte Carlo Simulations
of Relative DNA Damage From KV CBCT Radiation. Phys. Med. Biol. 58, 5693-5704 (2013)
A.G. Georgakilas, P.O'Neill, R.D. Stewart, Induction and Repair of Clustered DNA Lesions:
What Do We Know So Far? Radiat.
Radiat Res.
Res 180,100
180 100-109
109 (2013)
Slide 45