skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Maximum likelihood estimation for cytogenetic dose-response curves

Technical Report ·
DOI:https://doi.org/10.2172/5652085· OSTI ID:5652085

In vitro dose-response curves are used to describe the relation between the yield of dicentric chromosome aberrations and radiation dose for human lymphocytes. The dicentric yields follow the Poisson distribution, and the expected yield depends on both the magnitude and the temporal distribution of the dose for low LET radiation. A general dose-response model that describes this relation has been obtained by Kellerer and Rossi using the theory of dual radiation action. The yield of elementary lesions is kappa(..gamma..d + g(t, tau)d/sup 2/), where t is the time and d is dose. The coefficient of the d/sup 2/ term is determined by the recovery function and the temporal mode of irradiation. Two special cases of practical interest are split-dose and continuous exposure experiments, and the resulting models are intrinsically nonlinear in the parameters. A general purpose maximum likelihood estimation procedure is described and illustrated with numerical examples from both experimental designs. Poisson regression analysis is used for estimation, hypothesis testing, and regression diagnostics. Results are discussed in the context of exposure assessment procedures for both acute and chronic human radiation exposure.

Research Organization:
Union Carbide Corp., Oak Ridge, TN (USA). Nuclear Div.
DOE Contract Number:
W-7405-ENG-26
OSTI ID:
5652085
Report Number(s):
ORNL/CSD-123; ON: DE84001582
Country of Publication:
United States
Language:
English