Dosimetric characterization of two radium sources for retrospective dosimetry studies

During the first part of the 20th century, (226)Ra was the most used radionuclide for brachytherapy. Retrospective accurate dosimetry, coupled with patient follow up, is important for advancing knowledge on long-term radiation effects. The purpose of this work was to dosimetrically characterize two (226)Ra sources, commonly used in Sweden during the first half of the 20th century, for retrospective dose-effect studies.

An 8 mg (226)Ra tube and a 10 mg (226)Ra needle, used at Radiumhemmet (Karolinska University Hospital, Stockholm, Sweden), from 1925 to the 1960s, were modeled in two independent Monte Carlo (MC) radiation transport codes: geant4 and mcnp5. Absorbed dose and collision kerma around the two sources were obtained, from which the TG-43 parameters were derived for the secular equilibrium state. Furthermore, results from this dosimetric formalism were compared with results from a MC simulation with a superficial mould constituted by five needles inside a glass casing, placed over a water phantom, trying to mimic a typical clinical setup. Calculated absorbed doses using the TG-43 formalism were also compared with previously reported measurements and calculations based on the Sievert integral. Finally, the dose rate at large distances from a (226)Ra point-like-source placed in the center of 1 m radius water sphere was calculated with geant4.

TG-43 parameters [including gL(r), F(r, θ), Λ, and sK] have been uploaded in spreadsheets as additional material, and the fitting parameters of a mathematical curve that provides the dose rate between 10 and 60 cm from the source have been provided. Results from TG-43 formalism are consistent within the treatment volume with those of a MC simulation of a typical clinical scenario. Comparisons with reported measurements made with thermoluminescent dosimeters show differences up to 13% along the transverse axis of the radium needle. It has been estimated that the uncertainty associated to the absorbed dose within the treatment volume is 10%-15%, whereas uncertainty of absorbed dose to distant organs is roughly 20%-25%.

The results provided here facilitate retrospective dosimetry studies of (226)Ra using modern treatment planning systems, which may be used to improve knowledge on long term radiation effects. It is surely important for the epidemiologic studies to be aware of the estimated uncertainty provided here before extracting their conclusions.