ESTRO 36 Abstract Book

S190 ESTRO 36 _______________________________________________________________________________________________

and the radiotherapy management system, as the two systems are not linked.

Proffered Papers: Physics Dosimetry

OC-0357 Treatment planning dosimetry accuracy in 192Ir HDR brachytherapy of lip carcinoma P. Papagiannis 1 , V. Peppa 1 , T. Major 2 1 National and Kapodistrian University of Athens, Medical

Physics Lab.- Medical School, Athens, Greece 2 National Institute of Oncology, Department of Radiotherapy, Budapest, Hungary

Purpose or Objective Advanced dose calculation algorithms h ave become clinically available for 192 Ir HDR brachytherapy y to account for the effects disregarded by TG43 based dosimetry algorithms (heterogeneities, applicators and patient specific scatter conditions). The aim of this work is to study the effect of improved dosimetric accuracy in HDR brachytherapy of squamous carcinoma of the lip. Material and Methods Three anonimized patient cases were studied (treatments using the 192 Ir microSelectron-HDR v2 source, 27 Gy planning aim delivered in 3 Gy fractions b.d.). The plans were imported to OncentraBrachy v4.5 and dosimetry was repeated using both the TG43 and the Advanced Collapsed Cone Engine (ACE) TPS algorithms. The same TRAK was used with both algorithms for the same patient case. ACE calculations were performed using the high accuracy option taking into account individual voxel densities and assuming the elemental composition of water, average skin and cortical bone for the PTV and soft tissue, the skin, and the mandible, respectively. The spatial resolution of TPS dosimetry results was 1 mm, isotropic. Corresponding reference data were obtained from patient specific Monte Carlo (MC) simulations using the MCNP6 code with input files prepared from the parsing of dicom RT data with the BrachyGuide software tool. The TPS HU calibration was imported to BrachyGuide to ensure identical density input to ACE and MC. Dose was approximated by collision Kerma and kerma to medium in medium was scored using the F6 tally. BrachyGuide was also used for the comparison of the three RT dose files for each patient case (TG43, ACE, and MC). Results TG43 clearly overestimates results for all cases as shown in the left side of Figure 1 for an indicative case. This cannot be attributed solely to the difference between patient scatter conditions and TG43 assumptions since large differences are also observed close to the source dwell positions. The corresponding comparison between ACE and MC (right side of figure 1) shows agreement within MC type A uncertainty up to 5 cm from the implant. While ACE improves dosimetric accuracy, considerable differences are still observed close to the source dwell positions.

Results In total, 10 patients were treated with the long treatment schedule, and 10 with the short treatment schedule, resulting in 300 plan selections. Margin sets of 25 mm, 15 mm, 0 mm were created for 6 patients, and margin sets of 15 mm, 0 mm, -15 mm for 13 patients. One patient had a set of only two margins available (0 mm, 15 mm), due to insufficient time at treatment planning. Overall, the -15 mm, 0 mm, 15 mm and 25 mm plans were selected in 2%, 45%, 39% and 14% of fractions, respectively. For distributions per patient, see figure 2. The largest available margin was always sufficient. Treatment was delayed a total of 7 times (of which 5 times in 1 patient) to obtain a more favorable anatomy in case of a very full rectum, usually caused by gas pockets. Evaluation of the post-treatment CBCT scans showed for 1 fraction the selected plan was no longer suitable due to a moving gas pocket. The weekly review showed that a plan with a smaller margin could have been selected in 20% of fractions, and a larger margin in 2% of fractions. No inconsistencies were found in selected plans between the imaging system and radiotherapy management system.

Conclusion A plan selection strategy for rectum cancer patients was successfully and safely implemented. Next we will quantify the dosimetric impact of plan selection to the dose of the organs at risk in this dataset.

Results from the comparison of median DVH parameters in the Table show large differences between TPS calculations and MC for high dose PTV volumes (V150 και V200) in accordance with the above findings. Large differences between TPS calculations and MC are also observed for OAR parameters. These differences however correspond