Strength estimation of a moving 125Iodine source during implantation in brachytherapy: application to linked sources

摘要

The speed of a source varies during brachytherapy implantations because it is moved manually. In the method proposed in this study, the strength of a source moving at an unknown and varying speed is estimated by performing short time measurements. The measurement should be completed while the source is in the region where the efficiency of the detector is constant. This was detailed in a previous work [9]. Additionally, the conditions for the counting time and the distance between the detector and the guiding needle were investigated in order to determine the source strength, regardless of the source speed. Four 125I seeds (STM1251, Bard Inc., Murray Hill, NJ, USA) were used in this study. One of them was used in the form of a ‘loose source' (i.e. single source). The other three sources were linked (with a distance between them of 10.5 mm) and separated by spacers (5.5 mm Standard Source Link, and 5.0 mm Extension Source Link, Bard Inc.). The strength of the loose source was 8.48 U. The linked sources had strengths of 8.38 U for Source #1, 8.43 U for Source #2, and 8.63 U for Source #3. The difference between the strengths was within 2%. Sources #1, #2 and #3 approached the detector in that order. The detector is described in the next subsection. The experimental set-up is shown in Fig. 1. The source was moved into the needle (#918201, Bard Inc.) by an applicator and a push rod (200-TPV, Mick Radio-Nuclear Instruments Inc., Mount Vernon, NY). The position and speed of the source was controlled with an electric actuator (EZS3D025-A, Oriental Motor Inc., Tokyo, Japan), which was connected to the push rod. In this experiment, the source was not implanted into a patient but only moved in the needle. The detector used was a plastic scintillator (G-tech Inc., Saitama, Japan). This was comprised of the scintillator (EJ200, Eljen Technology Inc., Sweetwater, TX), with dimensions of 80 mm × 50 mm × 20 mm, and the photomultiplier tube (H7416, Hamamatsu Photonics Inc., Shizuoka, Japan). At first, the dependence of the detector response on the source position along the needle was measured using the static sources. The static source measurement was performed for both the loose source and the linked sources (i.e. #1, #2 and #3). The measurements were also performed for both geometries, with and without the grid. Measurements of 1-s duration were conducted 10 times. Among the measured values in each run for the moving source, the maximum was utilized in further analyses, assuming that the source reached the position with the highest detector response when the maximum reading in each run was obtained. In order to investigate the accuracy of the result for the moving source, its ratio against the result for the static source (i.e. moving:static) was computed. This was based on the assumption that the correct result was obtained with the conventional static source measurement. A moving-to-static source ratio close to unity means that the measurement for the moving source was performed correctly. The dependence of the detector response on distance for the static loose source is shown in Fig. 2. The source position is an arbitrary value, which approximately corresponds to the distance from the end of the applicator exit to the source in Fig. 1a. The error bars indicate the standard deviation for 10 measurements. The counting rate of the background signal was 22 ± 2 cps, and this was subtracted from the measured response for the static source shown in Fig. 2. The reading for the detector-to-needle distance of 5 mm reached a maximum of 42 309 ± 724 cps at 46.5 mm. For the distance of 20 mm, the maximum was 16 314 ± 336 cps. Further measurements were performed for the distance of 5 mm to obtain higher counting rates with better statistics. An example of the raw data is shown in Fig. 4 for the source moving at 100 mm s–1 with the counting time of 20 ms in the geometry without the grid. The measurements for the moving sources were performed as five runs. The average and standard deviation of the maximum in each of the five runs are indicated in Table 1. The ratio against the result for the static source (i.e. moving:static) is shown in Fig. 5. The results of the moving source showed better agreement with those of the static source when the grid was not used. Assuming that the source speed in the implantation ranged up to 200 mm s–1, a counting time of 10 ms was applied. In this case, the counting rate in Fig. 3b was converted to the counting time of 10 ms for the static source. As shown in Fig. 5b, the moving-to-static ratio in 10 ms counting time at 200 mm s–1 was between 1.02 and 1.08 for the three sources.