A new radioimmunotherapy (RIT) treatment planning system employing a semi‐automated segmentation toolkit and fast Hartley transform (FHT) for 3D convolution

Abstract

Efficient RIT depends on a priori knowledge of the radiation absorbed dose distribution. Straightforward application of published table of S values, which are for standard‐size homogeneous activity sources in a standard‐size body geometry, is not sufficient for accurate RIT dosimetry calculations. A new, fast, and functional treatment planning approach is developed that considers heterogeneous activity distributions. Calculation of the three‐dimensional absorbed dose distribution requires convolution of a cumulated activity distribution with a point‐source kernel; both represented by large matrices. To reduce the computation time required, a novel implementation of convolution using FHT is realized. Absorbed dose calculation time was reduced from 80 hours to 5 minutes. Volume measurement error has a direct impact on the absorbed dose calculation when mean activity information for an organ is used. Since identification of the tumor and other organs of interest is time‐consuming, a semi‐automated segmentation toolkit is developed. It is now possible to determine the volume of a structure in half the time with the same accuracy. The developed system was used to calculate and display as isodose contours the absorbed dose distribution in a patient's tumor and for a bone marrow sample. [For copies contact the author or University Microfilms, Inc., Ann Arbor, MI 48106 (Phone: 1‐800‐521‐0600).]