Page 42 - 2014 Printable Abstract Book
P. 42
linear dose-response function to be smoothed through an auto-regressive structure among the random
slope coefficients defined over closely spaced dose categories. Simulation studies and application to Life
Span Study data of atomic-bomb survivors show that this approach can produce smooth and flexible dose-
response estimations while reasonably handling the risk uncertainty with confidence intervals estimated
by the method of Markov Chain Monte Carlo. With few assumptions and modeling options to be made by
analyst, the method would be particularly useful in assessing the risk associated with radiation exposure
at low doses.
(S505) Lung and other respiratory cancers, smoking, and radiation risks in the LSS. Elizabeth Cahoon,
National Cancer Institute, Bethesda, MD
Lung cancer has the second highest incidence rate among solid cancers in Japan and accounts for
11% of all solid cancers observed in the Life Span Study (LSS), while other respiratory cancers (e.g.,
trachea, mediastinum) account for 0.5% of all solid cancers observed in LSS. During 1958-2009, 2,628 first
primary lung cancers, 191 laryngeal cancers, and 130 other respiratory cancers were diagnosed among
the 111,917 LSS members eligible for this study. The current data contain 11 additional years of follow-
up since the last LSS cancer incidence report. These analyses use improved smoking summary information,
updated radiation doses, and updated migration coefficients to account for migration of cohort members
into and out of the cancer registries’ catchment areas. We investigated the radiation effects on the
incidence of lung and other respiratory cancers adjusted for smoking. For lung cancer, this investigation
looks more closely at the nature of the complex departures from a simple multiplicative model.
Preliminary analyses suggest that the super multiplicative radiation/smoking interaction among light
smokers is weaker than previously reported due to the utilization of the revised smoking data.
S06 CLINICAL RADIATION RESEARCH USING PATIENT SPECIFIC MATHEMATICAL MODELS
The clinical challenges of delivering the right dose to the right patient at the right time is becoming
increasingly complicated with the use of heavy charged particles to treat cancer. This session will present
mathematical and image-guided approaches to these clinical challenges.
(S601) Analytical solutions of a repair, misrepair fixation model for a range of doses, energies and
1
2
particle types relevant to the clinic. Russell C. Rockne, PhD ; Robert D. Stewart, PhD ; and Kristin R.
2
Swanson, PhD, Northwestern University, Chicago, IL and University of Washington, Seattle, WA
1
1
Analytic solutions to the system of nonlinear ordinary differential equations used to model double
strand break (DSB) induction and repair in a repair-misrepair fixation (RMF) model will be presented. For
large doses relevant to high dose per fraction treatments (5-15 Gy per fraction “radiosurgery”), estimates
of the surviving fraction from the analytical solutions to the full model may be orders of magnitude larger
than estimates from the linearized (i.e., low dose) linear-quadratic (LQ) approximation to the RMF model.
Intra-fraction repair is also very significant for the typical doses and dose rates used in radiosurgery. The
large differences in model predictions between the linearized and analytical solution to the RMF model
play an important role in model parameter estimation and biological inferences from those parameters.
40 | P a g e
slope coefficients defined over closely spaced dose categories. Simulation studies and application to Life
Span Study data of atomic-bomb survivors show that this approach can produce smooth and flexible dose-
response estimations while reasonably handling the risk uncertainty with confidence intervals estimated
by the method of Markov Chain Monte Carlo. With few assumptions and modeling options to be made by
analyst, the method would be particularly useful in assessing the risk associated with radiation exposure
at low doses.
(S505) Lung and other respiratory cancers, smoking, and radiation risks in the LSS. Elizabeth Cahoon,
National Cancer Institute, Bethesda, MD
Lung cancer has the second highest incidence rate among solid cancers in Japan and accounts for
11% of all solid cancers observed in the Life Span Study (LSS), while other respiratory cancers (e.g.,
trachea, mediastinum) account for 0.5% of all solid cancers observed in LSS. During 1958-2009, 2,628 first
primary lung cancers, 191 laryngeal cancers, and 130 other respiratory cancers were diagnosed among
the 111,917 LSS members eligible for this study. The current data contain 11 additional years of follow-
up since the last LSS cancer incidence report. These analyses use improved smoking summary information,
updated radiation doses, and updated migration coefficients to account for migration of cohort members
into and out of the cancer registries’ catchment areas. We investigated the radiation effects on the
incidence of lung and other respiratory cancers adjusted for smoking. For lung cancer, this investigation
looks more closely at the nature of the complex departures from a simple multiplicative model.
Preliminary analyses suggest that the super multiplicative radiation/smoking interaction among light
smokers is weaker than previously reported due to the utilization of the revised smoking data.
S06 CLINICAL RADIATION RESEARCH USING PATIENT SPECIFIC MATHEMATICAL MODELS
The clinical challenges of delivering the right dose to the right patient at the right time is becoming
increasingly complicated with the use of heavy charged particles to treat cancer. This session will present
mathematical and image-guided approaches to these clinical challenges.
(S601) Analytical solutions of a repair, misrepair fixation model for a range of doses, energies and
1
2
particle types relevant to the clinic. Russell C. Rockne, PhD ; Robert D. Stewart, PhD ; and Kristin R.
2
Swanson, PhD, Northwestern University, Chicago, IL and University of Washington, Seattle, WA
1
1
Analytic solutions to the system of nonlinear ordinary differential equations used to model double
strand break (DSB) induction and repair in a repair-misrepair fixation (RMF) model will be presented. For
large doses relevant to high dose per fraction treatments (5-15 Gy per fraction “radiosurgery”), estimates
of the surviving fraction from the analytical solutions to the full model may be orders of magnitude larger
than estimates from the linearized (i.e., low dose) linear-quadratic (LQ) approximation to the RMF model.
Intra-fraction repair is also very significant for the typical doses and dose rates used in radiosurgery. The
large differences in model predictions between the linearized and analytical solution to the RMF model
play an important role in model parameter estimation and biological inferences from those parameters.
40 | P a g e
