Comparison of calibration procedures for 192Ir high-dose-rate brachytherapy sources.

PURPOSE: To compare the efficacy of different calibration procedures for 192Ir high-dose-rate (HDR) brachytherapy sources and to determine their suitability in clinical practice. In addition the manufacturer's calibration is compared with our experimental measurements so that the accuracy of the source strength on the manufacturer certificate which is supplied with each new 192Ir source can be accessed. METHODS AND MATERIALS: We compared three types of calibration system: well-type chambers (HDR-1000 and SDS), cylindrical phantom, and plate phantom. The total number of measurements we obtained was 365. The number of sources used for the calibration procedure comparison was 20 and the number used for comparison with the manufacturer's calibration was 46. This study was made during the period 1989-1997. Also, Physikalisch-Technische Bundesanstalt (PTB) calibrated one of our sources using their PTB protocol so that the results could be compared with our own. RESULTS: The sensitivity of each system on scattering from the room walls was studied. It was found that different minimum lateral distances from the walls were required for the different systems tested: 15 cm and 25 cm for the well-type chambers, 75 cm for the cylindrical phantom, and 13 cm for the plate phantom. The minimum thickness required to reach phantom scattering saturation for the plate phantom setup is 24 cm. The influence of the applicator material used in the calibration setup was found to be 1.7% for the stainless steel dosimetry applicator compared to the plastic 5F applicator. The accuracy of source positioning within the applicator can lead to dosimetric errors of +/-1.2% for the radial distance of 8.0 cm used with both solid phantoms. The change in the response for both well-type chambers was only 0.1% for changes in the source position within +/-7.5 mm around the response peak. Good agreement was found between all dosimetry systems included in our study. Taking the HDR-1000 well-type chamber results as a reference, we observed percentage root mean square (RMS) values of 0.11% for the SDS well-type chamber, 0.44% for the cylindrical, and 0.60% for the plate phantom setup. A comparison of our results using the cylindrical phantom with those of the manufacturer showed a percentage RMS value of 3.3% with a percentage fractional error range of -13.0% to +6.0%. The comparison of our calibration results with those of PTB gave deviations less than 0.4% for all systems. CONCLUSIONS: Our results have shown that with careful use of all calibration system protocols an accurate determination of source strength can be obtained. However, the manufacturer's calibration is not accurate enough on its own, and it should be mandatory for clinics to always measure the source strength of newly delivered 192Ir brachytherapy sources. The influence of the applicator material, metal or plastic, should always be taken into account.

Application of the Monte Carlo integration (MCI) method for calculation of the anisotropy of 192Ir brachytherapy sources.

Source anisotropy is a very important factor in the brachytherapy quality assurance of high-dose rate (HDR) 192Ir afterloading stepping sources. If anisotropy is not taken into account then doses received by a brachytherapy patient in certain directions can be in error by a clinically significant amount. Experimental measurements of anisotropy are very labour intensive. We have shown that within acceptable limits of accuracy, Monte Carlo integration (MCI) of a modified Sievert integral (3D generalization) can provide the necessary data within a much shorter time scale than can experiments. Hence MCI can be used for routine quality assurance schedules whenever a new design of HDR or PDR 192Ir is used for brachytherapy afterloading. Our MCI calculation results are compared with published experimental data and Monte Carlo simulation data for microSelectron and VariSource 192Ir sources. We have shown not only that MCI offers advantages over alternative numerical integration methods, but also that treating filtration coefficients as radial distance-dependent functions improves Sievert integral accuracy at low energies. This paper also provides anisotropy data for three new 192Ir sources, one for the microSelectron-HDR and two for the microSelectron-PDR, for which data are currently not available. The information we have obtained in this study can be incorporated into clinical practice.