The accuracy of the line and point source approximations in IR-192 dosimetry

The dosimetric approximations used in computer-aided treatment planning of Ir-192 seed implants generally ignore individual seed dimensions and internal structure. Most commonly, each seed is approximated by an isotropic point source. Alternately, each ribbon assembly consisting of uniformly spaced seeds is replaced by an unfiltered line source. Using filtration corrections applicable to platinum- and steel-encapsulated seeds calculated by the Monte Carlo method, the dosimetric errors introduced by these models into two-and-three-dimensional dose distributions arising from multiple plane implants are analyzed. Our results demonstrate that when anisotropy correction factors of 0.96 and 0.99 are used for platinum- and steel-filtered seeds, respectively, the point source model is accurate within 2%. The accuracy of the line source approximation depends significantly upon the details of its implementation. If the linear density of the line is set equal to individual seed strength for an inter-seed spacing of 1 cm, and filtration correction factors of 0.94 and 0.97 are used for platinum- and steel-clad seeds, respectively, the accuracy of the line source model is 1.5% near the implant center. The method of dose-volume histograms is used to compare the predictions of the different models.