Tucker S L, Thames H D, Brown B W, Mason K A, Hunter N R, Withers H R
Department of Biomathematics, University of Texas M. D. Anderson Cancer Center, Houston 77030.
Int J Radiat Biol. 1991 Mar;59(3):777-95. doi: 10.1080/09553009114550681.
Several methods are described for analysing the results of in vivo colony assays using the statistical procedure called maximum-likelihood analysis. The methods differ in the way in which they take into account possible sources of variability in the data. The methods described here for analysing microcolony data are direct methods, in that they use the observed colony counts rather than transformed (e.g. Poisson-corrected) data. Each method can be used to estimate the average number of surviving cells per tissue structure (e.g. per jejunal crypt) in a single dose group, together with 95% confidence intervals, or to fit cell-survival models to data from a range of dose groups (e.g. to obtain estimates of D0 or of the linear-quadratic parameters alpha and beta). Experimental microcolony data from murine jejunum, colon, and hair follicles irradiated in anagen (proliferative) or telogen (resting) phase have been analysed. Estimates of D0 have been derived from single-dose data and estimates of alpha, beta, and the initial number of clonogenic cells per structure have been derived from fractionation data. For hair follicles, the half-time of repair of sublethal radiation injury has also been derived from fractionation data.
