ESTRO 37 Abstract book

ESTRO 37

S574

Material and Methods First, dosilab Company has designed a phantom based on four GaN dosimeters with different inserts for machine QA: 1 and 2 channels. Specific calculation methods based on GaN dosimeter responses are used to accurate determination of dwell times and dwell position (Fig. 1).

guidelines. Interests of this phantom are multiples; on the one hand machines quality controls can be done quickly and simply. Indeed, source dwell position and dose delivered can be made in the same acquisition. On the other hand, the 6-channels insert allows quantitative verification of patient treatment plans before irradiation (position and dose). PO-1023 Verifying brachytherapy applicator models using an imaging panel M. Van den Bosch 1 , G. Fonseca 1 , R. Voncken 1 , M. Bellezzo 1 , F. Verhaegen 1 1 MAASTRO Clinic, Department of Radiation Oncology, Maastricht, The Netherlands Purpose or Objective A common applicator to treat cervical cancer patients with High-Dose-Rate (HDR) brachytherapy is a ring applicator. The highly curved source channel may lead to snaking of the source cable. In case of snaking, the cable position deviates from the center of the channel resulting in a source positioning error. It is common practise to exclude the source positions close to the tip of the channel for planning. However, the snaking effect is not limited to the end of the channel. As a result, the true source position differs from the planned dwell position in the Treatment Planning System (TPS) when assuming the source trajectory is perfectly in the center of the source channel. In this study, we quantify the positioning error using a default TPS model (the source trajectory is at the center of the channel) and a modified source trajectory derived with imaging panel (IP) measurements. Material and Methods A ring applicator (Varian) was attached to a vertically placed IP to mimic the orientation of the applicator during a treatment. Above the applicator a needle was placed. In this needle the HDR source was temporarily positioned to generate a projection image of the applicator. Next, the source was sent to dwell positions 129.4cm to 121.3cm with detrimental steps of 3mm. The center of the source was estimated using a 2D Gaussian fit to the dose profile measured by the IP and projected on top of the applicator geometry to determine the true source trajectory. The final result is a 2D picture that shows both the applicator geometry as the true source trajectory. This picture is imported into the TPS. The source trajectory of the default TPS applicator model was adjusted according to the derived true source trajectory. The mean and maximum errors were determined for the default and modified TPS models. Results As shown in Fig 1 the true source trajectory is not at the center of the channel but touches the inner and outer channel edge. The mean and maximum error were 2.9 and 3.9mm, and 0.0 and 1.2mm for the default and the modified model, respectively. The remaining error of the modified model could partly be explained by snaking in the third dimension (out-of-plane). However, the default model improved when the dwell positions were shifted 3mm along the source trajectory towards the channel tip. The mean and maximum error dropped to 0.0 and 1.2mm, respectively. Fig 1c shows the error per dwell position.

For this study, we validated the software development and check the accuracy on the source dwell position and on the measured dose (ESTRO guidelines 2mm and 5% respectively). Ten measurements were acquired for a prescribe dose of 5Gy and by using different inserts. Then we introduced an error on the source dwell positions from 1 to 10 mm to test the robustness of the system on the measured dose. Finally, we developed a 6-channel insert (Fig. 1) to perform quantitative patient quality control measurements in comparison with treatment plans. So, measurements of repeatability and reproducibility have been carried out on HDR-BT phantom. Results After ten successive measurements, errors between the measured and planned source dwell positions are not significant and respect the ESTRO guidelines of 2mm (mean values of 0.21mm [range -0.47:0.45], 0.14mm [range -0.53;0.38] and 0.06mm [range -0.59;0.33] for inserts with 1, 2 and 6-channels respectively). Taking into account the dose delivered, mean differences between measured and planned dose are 0.24% [range -3.24; 2.69], -1.64% [range -7.42;1.47] and -1.2% [range - 2.93;2.72] for inserts with 1, 2 and 6-channels respectively. Results respect the ESTRO guidelines of 5%. Finally, if we consider the error detection threshold with the GaN phantom, we notice that a dose greater than 5% is visible with an error of 2mm for insert 1 channel and 3mm for inserts 2 and 6-channels (Fig. 2).

Conclusion The designed HDR-BT phantom and inserts have been evaluated in accordance with the ESTRO booklet 8