Abstract Book

S1062

ESTRO 37

Material and Methods Treatment planning was performed with the TPS Pinnacle 14 (Philips) using FF and FFF beams (6MV) for 10 patients who received a local irradiation (L-RT) of the whole breast (50Gy) with a simultaneous integrated boost of 60Gy, and for 10 patients who received a loco-regional (LR) RT with target volumes consisting of the supraclavicular area, the upper axilla and the internal mammary chain. For an identical PTV coverage (at least 95% of PTV covered by 95% of the prescribed dose), the dosimetric comparison between FF and FFF VMAT plans was undertaken for organs at risk (OAR), such as heart, ipsilateral lung, contralateral breast and lung. The mean dose (Dmean) and the V50Gy of the skin were also evaluated. Arcs' arrangements in both cases were kept identical and the delivery treatment time on the Elekta Agility 160 was calculated. Results The mean PTV coverage was 96.5% and 97.3% for FF and FFF respectively. By using FFF beams, the mean heart dose was reduced from a mean value of 28% (0- 75% range) for L-RT and from 16% (0-44%) for LR-RT. The Dmean at the ipsilateral lung was identical for L-RT but slightly decreased of 4% (0-15%) for LR-RT, when the V5Gy was 7 % (0-17%) reduced with FFF-beam for both L and LR RT. The V50Gy of the skin was also lower and decreased from 150% (0-900%) and from 95% (0-400%) for L and LR-RT respectively. The OARs' sparing (especially of the skin) observed with FFF was mainly due to a combination of factors: softer X spectra, sharp fall-off of the penumbra, less electrons' contamination, lower MLC transmission, lower patient's scattering, which induced less dose to the organs surrounding the PTV. The Dmean at the contralateral breast was identical for FF and FFF: 1.1±0.5Gy and 1.8±0.6Gy for L and LR RT respectively. The V5Gy at the contralateral lung was negligible in all cases (<0.1Gy). As expected, the delivery time per arc with FFF beam was also reduced from 17±1s to 12±1s for L-RT and from 18±1s to 13±1s for LR- RT. Conclusion As expected, FFF-VMAT beams reduce the delivery time for a DIBH treatment which makes it more comfortable. PTV coverage is achieved with a dose reduction to OARs, especially the heart and ipsilateral lung, while preserving the contralateral organs during breath hold. EP-1954 Use of dose painting technique in single fraction celiac plexus radiosurgery M. Ben Ayun 1 , L. Hamer 1 , D. Alezra 1 , D. Sergey 1 , T. Lev 1 , W. Ilana 1 , S. Zvi 1 , R.L. Yaakov 1 1 Chaim Sheba Medical Center Institute of Oncology, Oncology, Tel Nashomer, Israel Purpose or Objective Celiac plexus radiosurgery is a new palliative treatment for severe pain experienced in pancreatic cancer, and is the subject of an ongoing clinical trial (NCT02356406). A single-fraction of 25 Gy is targeted to the celiac plexus. Close approximation of the plexus to the duodenum (small bowel) creates a severe dosimetric challenge. Here we compare two planning techniques: 1) physician prescribed dose-painting based upon distance from the duodenum (Four PTVs ranging from 25Gy to 10Gy) and 2) prescribing a uniform dose of 25Gy to the entire PTV with appropriate cost-functions to achieve bowel constraints. Material and Methods All patients underwent both 3D and 4D CT simulations with abdominal compression. Fifteen patients were planned and treated with a single fraction to celiac plexus. Individual bowel loops were contoured. The celiac plexus was contoured from T12 to L2 (GTV), and a 1cm expansion added to create the overall PTV.

Dose-painting technique: Volumes were created for the purpose of dose-painting included expand-bowel_0.5cm, expand-bowel_1cm, expand_GTV_1cm. A range of PTVs (25Gy, 20 Gy, 15Gy and 10Gy) were generated based upon Boolean operations: for instance, PTV25 = [GTV SUBTRACT expand bowel 1cm], PTV20 = [expand_GTV_1cm SUBTRACT (PTV25 OR expand- bowel_1cm)]. Uniform-dose technique: The entire PTV structure was prescribed 25Gy. Constraints for single fraction treatment (based upon Timmerman) included mean kidney dose 5.5Gy, max small bowel dose 11.2Gy. Calculations performed with AAA algorithm. Treatment plans were optimized using a volumetric intensity-modulated arc treatment (VMAT,Varian); dose was delivered with two/three arcs. Plan-specific QA check (using ion-chamber and ArcCheck) was done. Results Volume of the entire PTV structure was 300cc±80cc. For the dose-painting techniques volumes of PTV 25, 20, 15, and 10 were 30cc±15cc, 120cc±70, 150cc±100 and 30cc±12 respectively. The duodenum touched the celiac- plexus PTV in all cases. Dose coverage of the celiac-plexus (GTV) was significantly higher for the uniform-dose technique, D95 20Gy±4 vs 14Gy±4 (p<0.0012). Regarding normal tissue constraints: The small bowel volume that received above 11Gy was significantly higher for the uniform-dose plans 103cc±61 vs 4.2cc±3.2 (p<0.0086). The kidneys mean dose were 6Gy±0.6 vs 5.2Gy±0.4 (p<0.0005) respectively. The max dose to the spine was similar in both techniques: 10.6Gy±1.5 vs 10.4Gy±0.4 (p<0.778) respectively. The dose to 700cc of the liver was 3.97Gy±1.2 vs 3.5Gy±1.2 (p<0.0081)respectively. Clinical treatments were planned and delivered using the dose-painting technique. Pain relief was excellent and side-effects minimal. Conclusion For single-fraction radiosurgery to the celiac plexus, the dose-painting technique ensures rapid high-quality plans that meet normal-tissue constraints. Conversely, the uniform-dose technique ensured good GTV coverage but was unable to generate plans that respected normal tissue tolerance constraints, especially of the bowel. EP-1955 Evaluation of a knowledge-based model for VMAT planning of anal canal cancer S. Puccini 1 , D. Völzke 2 1 Strahlentherapie Bonn-Rhein-Sieg, Medical Physics, Bonn, Germany 2 Strahlentherapie Bonn-Rhein-Sieg, Radiation Oncology, Bonn, Germany Purpose or Objective To compare the performance of a knowledge-based optimisation process with the treatment planning process driven by experienced planners. Material and Methods A commercial system for knowledge-based radiotherapy planning, RapidPlan (RP), was employed to generate a DVH estimation model for the treatment of anal canal cancer. Sixty previously clinical accepted cases delivering 50,4 Gy in 28 fractions were used to train the model. The target volume was delineated according to the RTOG Consensus Panel Contouring Atlas for Anorectal Cancer, organs at risk included in the model were bladder, bowel, genitals and hip joints. Thirty new cases were included in the comparison. For each new patient an experienced planner (EP) individually chose the field geometry and the optimisation constraints. The planner was free to use as many dummy structures and as many iterations as needed to achieve a good plan. With the same field