Late radiation damage in bone, bone marrow and brain vasculature, with particular emphasis upon fractionation models

Abstract

X-ray induced changes in rat and human bone and bone marrow vasculature and in rat brain vasculature were measured as a function of time after irradiation and absorbed dose. The absorbed dose in the organ varied from 5 to 25 Gy for single dose irradiations and from 19 to 58 Gy for fractionated irradiations. The number of fractions varied from 3 to 10 for the rat experiments and from 12 to 25 for the human studies. Blood flow changes were measured using an 125I antipyrine or 86RbCl extraction technique. The red blood cell (RBC) volume was examined by 51Cr labelled red cells. Furthermore, different fractionation models have been compared. Radiation induced reduction of bone and bone marrow blood flow were both time and dose dependent (8, 10, 11). Reduced blood flow 3 months after irradiation would seem to be an important factor in the subsequent atrophy of bones (8). With a single dose of 10 Gy the bone marrow blood flow, although initially reduced returned to the control level by 7 months after irradiation (10). In the irradiated bone the RBC volume was about same as that in the control side but in bone marrow the reduction was from 32 to 59 % (9, 10). The dose levels predicted by the nominal standard dose (NSD) formula produced approximately the same damage to the rat femur seven months after irradiation when the extraction of 86Rb chloride and the dry weight were concerned as the end points (9). However, the results suggest that the NSD formula underestimates the late radiation damage in bone marrow when a small number of large fractions are used (10). The present data for late changes in bone marrow do not permit an accurate assessment of the α/β ratio (probability of the component of irreparable events to the component of cumulative sublethal damage), but the results would suggest a value of between 2 and 4 Gy (10). In the irradiated brains of the rats the blood flow was on average 20.4 % higher compared to that in the control group. There was no significant difference in brain blood flow between different fractionation schemes (12). The value of 0.42 for the exponent of N corresponds to the average value for central nervous system (CNS) tolerance that is available in the literature. The model used may be sufficiently accurate for clinical work provided the treatment schemes used do not depart too radically from standard practice.