Abstract Book

S1021

ESTRO 37

and automatically (AP) generated plans were created by means of the 'dual arc” feature. For the MP plans, additional non-anatomical structures needed to be delineated in order to interactively guide the optimization process. For AP plans, a progressive optimization algorithm is used to continually adjust initial targets/OARs objectives. Moreover, tuning structures and objectives are automatically added during optimization to increase the dose fall-off outside targets and improve the dose conformity and homogeneity. Various dose and dose-volume metrics (D98%, D95%, D50%, D2%, Dmean, V95% for target volumes; Dmean, D2% and various Vx% for OARs), as well as conformity (CI) indexes and healthy- tissue integral dose (ID) were evaluated. A Wilcoxon paired test was performed for plan comparison with statistical significance set at p<0.05. The time efficiency for plan optimization was estimated. Results Differences in all dose coverage metrics (in terms of V95%, D98%, D50%, D2% and Dmean) for both PTVs were not statistically significant (p<0.05). Differences in CI reached significance only for PTV2 (MP:1.59 vs. AP:1.48). Differences in DVHs were no significant in overall dose range for rectum, bladder and small bowel. For rectal doses: V 50Gy : 31.7 vs. 32.2Gy, V 60Gy : 21.1 vs. 22.5Gy; Dmean: 40.6 vs. 40.5Gy. For bladder doses: V 65Gy : 19.6 vs. 20.6 and Dmean: 44.3 vs. 43.8Gy. For small bowel: V 15Gy : 105.0 vs 119.1cc, Dmean: 13.3 vs. 12.8cc. AP plans provided a decrease in Integral Dose of 5.1%. The mean number of MUs was very similar between MP (537) and AP (546). Conclusion The Pinnacle Auto-Planning module achieved highly consistent treatment plans also in the cases of complex anatomical sites, meeting our institutional clinical constraints. No significant differences were found with respect to MP generated by experienced physicists. The effective working time was substantially reduced with Auto-Planning. EP-1886 Coplanar and non-coplanar VMAT facilitate OAR dose sparing in central lung SBRT J.B. Thomsen 1 , G.F. Persson 1 , M.C. Aznar 2 , A.K. Berthelsen 1 , L. Specht 1,3 , M. Josipovic 1,4 1 Rigshospitalet Copenhagen University Hospital, Department of Oncology- Section of Radiotherapy, Copenhagen, Denmark 2 University of Manchester, Manchester research cancer centre- Division of Cancers sciences, Manchester, United Kingdom 3 University of Copenhagen, Faculty of Health and Medical Sciences, Copenhagen, Denmark 4 University of Copenhagen, Niels Bohr Institute- Faculty of Science-, Copenhagen, Denmark Purpose or Objective Stereotactic body radiotherapy (SBRT) in centrally located lung tumors is associated with high toxicity risk. Data suggest that high maximum doses to the heart correlate with overall survival [IJROBP 2017, 99(2): 239- 240]. We hypothesize that volumetric modulated arc therapy (VMAT) will enable better organs at risk (OAR) sparing compared to 3D conformal (3DC) technique, and that non-coplanar geometry with VMAT will facilitate additional sparing without compromising the target dose. Material and Methods Ten consecutive patients treated with SBRT for single central tumor were included retrospectively. Tumor position characteristic is shown in table 1. A dose of 7 Gy x 8 was prescribed to the PTV periphery, aiming for ≥140% hot spot centrally in the tumor. CTV-GTV margins were 0-2 mm, CTV-PTV margins were based on geometric uncertainties and tumor motion measured using a 4D- CTplanning scan. 3DC clinical treatment plans were recalculated with a type C dose calculation algorithm and

furthermore two VMAT plans were made per patient for comparison: 1) with two coplanar 180º arcs (C-VMAT) and 2) a non-coplanar plan with single 180º arc at couch 0º and two 60º arcs with couch rotated 90º (NC-VMAT). GTV and PTV coverage were evaluated with maximum (max) and minimum (min) point dose, near maximum (n- max, lowest dose to the hottest 0.05cm 3 ) and near minimum (n-min, highest dose to the coldest 0.05cm 3 ) dose, and mean dose. OAR hard constraints[MA1] were spinal cord <33.6 Gy, trachea <48.8 Gy and contralateral main bronchus (CMB) <48.8 Gy. Soft constraints were ipsilateral main bronchus (IMB) <56.0 Gy, heart <41.6 Gy and esophagus <41.6 Gy. We reported max and n-max doses to these OAR and mean heart dose (MHD). Non-parametric statistics were used: Friedman’s two way analysis of variance and Wilcoxon signed rank test.

Results C-VMAT reduced max dose to the heart (p<0.03) compared to 3DC and NC-VMAT, while MHD remained unchanged (p=0.5). No significant differences were found for trachea (p=0.7). NC-VMAT was superior to 3DC for max IMB dose (p<0.01) and offered best sparing to spinal cord, esophagus and CMB (all p<0.01), which were all outside the high dose area (more details in Table 2). The PTV included or abutted the heart in 6 patients and IMB in 7 patients. PTV min, n-min, max and n-max doses did not differ significantly between the plans (p>0.08). Mean GTV and PTV doses were reduced with both VMAT plans (p<0.05), resulting in less steep dose gradient within the PTV. The steepness of the dose gradient close to the PTV edge in 3DC was also expressed in substantial median differences between the max and n-max doses to the heart ( 2.5 Gy ) and IMB (6.8 Gy ), and min and n-min GTV dose (4 Gy).

Conclusion C-VMAT reduced max heart dose compared with 3DC; NC- VMAT was not superior to C-VMAT for heart parameters.