Background: The measurement of the heterogeneity of radiolabeled monoclonal antibody uptake in tumor has an essential role in the calculation and interpretation of the absorbed dose of radiation. Large data arrays and long calculation times have been limiting factors in the calculation of three-dimensional dose-rate distributions used to study the relationship between uptake heterogeneity and dose.
Methods: Serial autoradiographs of tumor sections were digitized with approximately 100 microns resolution using a laser densitometer. The section images were aligned to form a registered tumor-image data set. The image data were corrected for film response versus activity density to create a three dimensional activity density distribution using features of a three-dimensional radiotherapy treatment planning system. Dose-rate distributions were formed by convolution with a beta dose kernel using fast Fourier transforms.
Results: Differential dose-rate-volume histograms (derived from the dose-rate distribution) were created to summarize the dose-rate nonuniformity throughout the tumor volume. Effects of section sampling interval, interpolation methods between section planes, and calculation resolution on the dose-rate-volume histograms were illustrated.
Conclusions: The several orders of magnitude improvement in calculational speed provided by the fast Fourier transform technique allowed an investigation of the effects of the calculational parameters. This investigation enabled tuning of both data acquisition and dose computation. These studies can lead to further enhancements in the calculational efficiency of three-dimensional dose-rate distributions. These improvements will allow the study of summing techniques to yield average total dose distributions.