The distribution of time intervals between irradiation by X-rays and onset of acute or chronic myeloid leukaemia has been investigated, in particular for the ankylosing spondylitics in COURT-BROWN and DOLL's study series. Two methods of analysis are given in outline, both of which allow (a) for several courses of irradiation on one patient, any of which might have induced leukaemia, and (b) for limited information on longer time intervals because there were many recent treatments. Intervals in sub-groups, e.g. for acute leukaemia only, and for “doubtful” cases only, are also given separately. The probability of onset is found to increase suddenly to a maximum in the fourth year after the inducing irradiation. The corresponding maximum incidence among all 3-year-old children is discussed in relation to the time intervals following irradiation of a foetus or a young baby. The analysis in general confirms that leukaemia can be caused by X-rays but throws doubt on the assumed linear dose-response relationships, which depend upon assuming a uniform distribution of latent periods. Accepted ideas on risk in terms of dose to the spinal marrow, and on mechanisms of induction, are briefly criticized.
©1961Health Physics Society