[ RadSafe ] Article: Fitting the linear-quadratic model to detailed data sets for different dose ranges

[John Jacobus]

>From *Physics in Medicine and Biology*
Volume 51, 7 June 2006

Fitting the linear-quadratic model to detailed data
sets for different dose ranges
    L M Garcia, J Leblanc, D Wilkins and G P Raaphorst
    2006 Phys. Med. Biol. 51 2813-2823
    Abstract:  
http://www.iop.org/EJ/abstract/-alert=1462/0031-9155/51/11/009
    Full text PDF: 
http://www.iop.org/EJ/article/-alert=1462/0031-9155/51/11/009/pmb6_11_009.pdf

    Survival curve behaviour and degree of
correspondence between the linear-quadratic (LQ) model
and experimental data in an extensive dose range for
high dose rates were analysed. Detailed clonogenic
assays with irradiation given in 0.5 Gy increments and
a total dose range varying from 10.5 to 16 Gy were
performed. The cell lines investigated were: CHOAA8
(Chinese hamster fibroblast cells), U373MG (human
glioblastoma cells), CP3 and DU145 (human prostate
carcinoma cell lines). The analyses were based on
χ2-statistics and Monte Carlo simulation of the
experiments. A decline of LQ fit quality at very low
doses (<2 Gy) is observed. This result can be
explained by the hypersensitive effect observed in
CHOAA8, U373MG and DU145 data and an adaptive-type
response in the CP3 cell line. A clear improvement of
the fit is discerned by removing the low dose data
points. The fit worsening at high doses also shows
that LQ cannot explain this region. This shows that
the LQ model fits better the middle dose region of the
survival curve. The analysis conducted in our study
reveals a dose dependency of the LQ fit in different
cell lines.
 ---------------------
As noted in the Discusson:

     "A drop in the fit quality was found when low
doses were included. This could be a result of the
linearity in the trend of the survival curve at low
doses that will affect the total fit in a range from 0
Gy to the final dose in the linear–quadratic region.
     "However, the very low dose region (&#8764;0–2
Gy) strongly influences the decline of the fit
quality. This outcome can be explained by the
hypersensitive effect observed in CHOAA8 and U373MG
data and an adaptive-type response in CP3 cell line.
Detailed study becomes necessary to assess a
hyper-radiosensitive/induced-radioresistant response
in both cell lines, especially, because this effect
has not been detected in human U373MG and hamster
ovarian 
CHOAA8 cells, but has been reported in other cell
lines (Marples et al 2004, Bartkowiak et al 2001).
     "The LQ behaviour characterizes the survival
curves in the middle dose region, which is
demonstrated by a clear improvement of goodness of fit
and Monte Carlo analyses. 
     "The fit worsening at high doses showed as
expected that the LQ model cannot explain this region.
LQ model predicts a constant increasing slope at high
doses in contrast to the constant slope observed in
experimental survival curves. This can be translated
in our study as a deterioration of goodness of fit at
those dose ranges.
     "Exactness of region delimitation is affected by
hypersensitivity, adaptive response at very low doses
and data uncertainties."

+++++++++++++++++++
"You get a lot more authority when the workforce doesn't think it's amateur hour on the top floor."
GEN. MICHAEL V. HAYDEN, President Bush's nominee for C.I.A. director.

-- John
John Jacobus, MS
Certified Health Physicist
e-mail:  crispy_bird at yahoo.com

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com