At sufficiently large dosages, chronic alpha-radiation from skeletally deposited radium and mesothorium in humans results in osteoporosis, dense bone necrosis, spontaneous fractures, bone sarcomas, and carcinomas of the paranasal sinuses and mastoids. At high dose rates and high cumulative dosages (1000–50,000 rads) the occurrence of radiogenic malignancy appears to be roughly independent of dosage, but the tumor appearance time seems to be longer for the lower doses than for the larger doses. Possibly a low dose domain exists in which the tumor appearance time exceeds the life span which remains after exposure. At cumulative dosages below the order of 1000 skeletal average rads no clinically significant radiobiological injury has yet been observed in the M.I.T. series over a time span of 40–50 yr in more than 500 persons. It may be that in the low-dose domain the rate of radiation injury is slower than the body's recovery and repair rates. The dosage vs response relationships for bone sarcoma and head carcinomas appear to be curvilinear and cannot be fitted to any known linear nonthreshold model. Linear extrapolations into the occupational dosage domain of a few rads can be made only with reservations and explicitly stated assumptions, and with recognition that the true value probably lies between the linear extrapolation value and zero. In constructing dosage vs response relationships it is imperative to recall that many contemporary tabulations of data, including present CHR, BEIR, and UNSCEAR tables, do not distinguish between unselected epidemiologically suitable cases and symptom-selected epidemiologically unsuitable cases. Use of these gross mixtures of unselected and selected cases has introduced significant bias into several recently proposed dosage vs response relationships.
