ESTRO 37 Abstract book

S1238

ESTRO 37

Results For oesophageal treatments, ACE calculations demons- trated up to a 2.5% reduction in PTV V100 and D98 was reduced by up to 0.18Gy compared to TG43-U1 calculations (based on a prescription dose of 7Gy). D0.1cc to the lung was reduced by up to 0.48Gy and D0.1cc to surrounding bone was reduced by up to 0.35Gy. For surface mould treatments, ACE calculations demonstrated up to an 11.2% reduction in PTV V100 and D98 reduced by up to 0.22Gy from TG43-U1 calculations (based on a prescription dose of 5.25Gy). D0.1cc to surrounding bone deviated from TG43-U1 by up to 0.3Gy. The deviation between TG43-U1 and ACE when calculating PTV V100, PTV D98 and Bone D0.1cc increased when using a larger mould. Conclusion The TG43-U1 dosimetry formalism for 192 Ir brachytherapy treatment planning cannot model a lack of full scatter conditions or account for the presence of heterogeneities. These assumptions lead to dose differences compared to ACE. During commissioning, it was shown that ACE demonstrated better agreement with MC calculations than TG43-U1. For all but one clinical case, TG43-U1 overestimated dose to both the target volume and organs at risk. The one case which did not exhibit this was a surface mould upper lip treatment, where the presence of the jaw and teeth under the target volume provided significant back- scattering. This led to an underestimation of PTV coverage by TG43-U1. Whilst this confirms the need for model-based dose calculation algorithms for these groups of patients, as discussed in TG-186, the disadvantage of extended calculation times and the requirement of additional contouring and staff training must be considered before implementing these new algorithms into clinical practice. EP-2240 HDR prostate brachytherapy database: preliminary dosimetric and radiobiological analysis. A. Balabanova 1 , B. Genova 1 , P. Stavrev 2 1 National Oncological Center Hospital, Radiotherapy department, Sofia, Bulgaria 2 Sofia University "St Kliment Ohridski", Faculty of Physics, Sofia, Bulgaria Purpose or Objective To create database comprising data on prostate patients treated by HDR brachytherapy in our institution. To analyze the shape of differential DVH and similarity between the different patients and to apply TCP and NTCP models in order to estimate the outcome of the treatment. Material and Methods During the last 6 years, 80 patients with prostate carcinoma were treated in our institution with HDR brachytherapy – monotherapy in three fractions of 10.5/11 Gy. Real-time US (ultrasound) based technique was used, with Oncentra Prostate planning software. In order to analyze the data from these patients a database was created. It comprises prostate DVHs, and those of OAR, the observed tumor control as well as the PSA test. So far there are only 23 entries in the database. Of all the treated patients recurrence with clinical imaging is present in only one case, while there were two cases with biochemical failure by the Phoenix definition (nadir+2ng/ml). We have constructed an average differential dDVH and using the method described in [1] have estimated the theoretical TCP value for all of the patients in the database. For this purpose the total dose was recalculated using an a/b ratio of 3Gy. Results On plot a/ of the Figure dDVHs for each fraction of all patients and the average one are presented. The TCP calculated using the average dDVH is given above the plot. Plots b)-d) show the average dDVH and the DVHs

Dose coverage was improved in HIPO plans compared to IPSA: D90 was 100.6 vs. 96.2% (p=0.0008), COIN was also higher: 0.6 vs. 0.47 (p<0.001), while DHI was lower: 0.34 vs. 0.41 (p<0.001), respectively. The D 2ccm of rectum, sigmoid and bowels was lower with HIPO optimisation: 2.5 vs. 2.7 Gy (p=0.0059), 3.2 vs. 3.6 Gy (p<0.001) and 4.1 vs. 4.6 Gy (p=0.0039). In both HIPO and IPSA plans, higher needle number resulted improved dose coverage (p=0.0234 and 0.0034), homogeneity (p=0.0004 and p=0.0004) and COIN (p=0.002 and 0.0121). Volume of HR- CTV correlated with COIN (R 2 =0.45, p=0.0015 and R 2 =0.52, p<0.001) and D 2ccm of bladder (R 2 =0.49, p=0.0006 and R 2 =0.36, p=0.0081), rectum (R 2 =0.65, p<0.001 and R 2 =0.59, p<0.001) and sigmoid (R 2 =0.60, p<0.001 and R 2 =0.44, p=0.0058), while DHI showed correlation only with COIN (R 2 =0.31, p=0.0338 and R 2 =0.33, p=0.0245), respectively. Conclusion HIPO optimisation results more improved treatment plans in combined intracavitary-interstitial IGABT of cervical cancer than IPSA, except for dose homogeneity. The use of higher number of needles improves dose coverage, homogeneity and conformality significantly with both invers algorithms. EP-2239 Evaluation of the Advanced Collapsed Cone Engine (ACE) for Ir-192 brachytherapy treatment planning L. Eason 1 , J. Mason 1 , P. Bownes 1 1 Leeds Cancer Centre, Medical Physics and Engineering, Leeds, United Kingdom Purpose or Objective The purpose of this work is to evaluate the Oncentra Advanced Collapsed Cone Engine (ACE) dose calculation algorithm for high-dose-rate (HDR) 192 Ir brachytherapy oesophageal and surface mould treatment planning. These treatment sites present a deviation from full- scatter conditions and contain significant heterogeneities, thus are likely to exhibit dosimetric differences from conventional all-water TG43-U1 dose calculations. ACE dose distributions are assessed against current clinical protocols and may help to inform of potential changes in the future. Material and Methods ACE was commissioned following guidance from the TG- 186 working group, comparing the accuracy of the ACE algorithm against conventional TG43-U1 calculations and gold-standard Monte Carlo (MC) calculations. Following commissioning, dose distributions generated using the high accuracy setting of the ACE algorithm were compared retrospectively with TG43-U1 distributions for five oesophageal and five surface mould treatment plans. The dosimetric differences between each algorithm were visually assessed using dose difference maps and side-by- side comparisons (Figure 1). Differences in associated DVH statistics (e.g. PTV D90) were assessed quantitatively and presented in Box and Whisker plots.