Monte Carlo-based dose-rate tables for the amersham CDCS.J and 3M model 6500 ¹³⁷CS tubes

Purpose: (1) To present reference-quality dose-rate distributions for the Amersham CDCS.J-type ¹³⁷Cs intracavitary source (hitherto unavailable in the literature) and updated tables for the 3M model 6500/6D6C source. (2) To assess the accuracy of the widely used 1D pathlength (Sievert integral) algorithm for lightly filtered ¹³⁷Cs tube sources. Methods and Materials: A Monte Carlo photon-transport code is used to calculate the dose-rate distributions about the 3M source and the CDCS.J source based on radiographic examination of the sources and the vendors' specifications. Dose-rate distributions are provided in the form of Cartesian 'away-and-along' lookup tables. Using a general form of the Sievert integral, calculated dose-rate distributions were compared to the Monte Carlo benchmark calculations treating the filtration coefficients as best-fit parameters as well as approximating them by linear energy absorption coefficients. In addition, the errors introduced by approximating the active source core by uniform cylinders or line sources was evaluated. Results: The Model CDCS.J dose distribution differs from that of the 3M model 6500 source by -5.9% to +14.4% (root-mean-square [RMS] average: 2.6%). The RMS accuracy of the Sievert algorithm is 2.4% to 2.8% (error range of -1.4% to 7.6%) when filtration coefficients for steel and ceramic media are approximated by linear energy absorption coefficients. If the filtration coefficients are treated as parameters of best fit, selected to minimize the discrepancies between 1D pathlength and Monte Carlo calculations, the RMS error is reduced to 0.8% (error range of -1.8% to 4.1%). The optimal values of stainless steel and low-density ceramic or glass filtration coefficients are approximately independent of the source geometry. Conclusions: The widely used Sievert integral algorithm accurately characterizes the dose distribution around stainless-steel clad low-density matrix ¹³⁷Cs sources, particularly if design-independent best-fit values of the filtration coefficients are used. Although both families of source designs studied produce similar dose distributions, source-design specific dose distributions should be used for clinical treatment planning and dose-algorithm validation.