THE SPATIAL ACCURACY OF CELLULAR DOSE ESTIMATES OBTAINED FROM 3D RECONSTRUCTED SERIAL TISSUE AUTORADIOGRAPHS

In order to better predict and understand the effects of radiopharmace uticals used for therapy, it is necessary to determine more accurately the radiation absorbed dose to cells in tissue. Using thin-section au toradiography, the spatial distribution of sources relative to the cel ls can be obtained from a single section with micrometre resolution. B y collecting and analysing serial sections, the 3D microscopic distrib ution of radionuclide relative to the cellular histology, and therefor e the dose rate distribution, can be established. In this paper, a met hod of 3D reconstruction of serial sections is proposed, and measureme nts are reported of (i) the accuracy and reproducibility of quantitati ve autoradiography and (ii) the spatial precision with which tissue fe atures from one section can be related to adjacent sections. Uncertain ties in the activity determination for the specimen result from activi ty losses during tissue processing (4-11%), and the variation of grain count per unit activity between batches of serial sections (6-25%). C orrelation of the section activity to grain count densities showed dev iations ranging from 6-34%. The spatial alignment uncertainties were a ssessed using nylon fibre fiduciary markers incorporated into the tiss ue block, and compared to those for alignment based on internal tissue landmarks. The standard deviation for the variation in nylon fibre fi duciary alignment was measured to be 41 mu m cm(-1), compared to 69 mu m cm(-1) when internal tissue histology landmarks were used. In addit ion, tissue shrinkage during histological processing of up to 10% was observed. The implications of these measured activity and spatial dist ribution uncertainties upon the estimate of cellular dose rate distrib ution depends upon the range of the radiation emissions. For long-rang e beta particles, uncertainties in both the activity and spatial distr ibution translate linearly to the uncertainty in dose rate of < 15%. F or short-range emitters (< 100 mu m), such as alpha particle sources, the magnitude of the uncertainty in serial section alignment is compar able with the particle track length. Under these circumstances, dosime tric errors are introduced in proportion to the serial section alignme nt inaccuracy.