Abstract Book

S273

ESTRO 37

of the segmentation was expressed as the number of criteria that were met. Results Comparison of (a) and (b) in Fig.1 shows that FMO with high beamlet resolution clearly outperformed low resolution. However, if low resolution FMO is followed by high resolution multi-criterial segmentation (c), the overall quality approaches that of FMO with high resolution, compare (c) with (b). Without multi-criterial segmentation, this restoration in plan quality does not occur to the same extent as with multi-criterial segmentation (compare (d) with (a) and (c)). The high resolution FMO plans (b) could be fully restored with multi-criterial segmentation, compare with (e). Segmented plans were slightly less conformal than FMO plans; the Conformation Number (CN, van ‘t Riet et al. 1997) of low resolution FMO plans (a) of 0.859 reduced to 0.843 for the segmented plans (c), p<0.001. Plans (b) and (e) showed CNs of 0.866 and 0.846, p<0.001. On the other hand, no significant differences in CN were observed between segmented plans with (c) and without multi- criterial fine-tuning (d).

MRIdian planning system was used to generate step-and- shoot IMRT plans with 6MV (MRIdian LINAC ) and 60 Co (MRIdian 60 CO ), both with 15 equidistant beams. These plans were compared to VMAT plans, generated in Eclipse™ (Varian Medical Systems, Palo Alto) using two 6MV coplanar arcs (360°). Mean DVHs of PTVs and relevant organs at risk (OARs), conformity index (CI), gradient index (GI) were compared. Results Median PTV size was 88cc for prostate- and 55cc for pancreatic cases. In general, the mean CI was highest with MRIdian LINAC plans (CI 0.91 and 0.84 for prostate and pancreatic cancer, respectively). This difference was significant in comparison to both VMAT and MRIdian 60 CO for prostate plans (CI 0.85 and 0.76; both p<0.05), and for pancreas only compared to MRIdian 60 CO (0.69; p<0.05). Similarly, the GI was lowest for MRIdian LINAC plans (3.52 and 3.16), marginally better than obtained with VMAT (3.86 and 3.67), while MRIdian 60 CO plans had significantly higher GI (5.72 and 4.97) for prostate and pancreatic plans (Figure 1). Mean DVH’s of PTV coverage and doses to relevant OARs for MRIdian LINAC and VMAT were largely comparable, and better than MRIdian 60 CO . MRIdian LINAC allowed for improved target coverage (PTV V95% 91.2%) in pancreas cancer, without exceeding prioritized high-dose constraints for adjacent OARs, compared to the other techniques (89.7% and 83.2%; Figure 2) .

Conclusion Fluence map optimization with fast, low resolution beamlet optimization resulted in criteria violations that could largely be restored with the proposed multi- criterial MLC segmentation technique. OC-0519 Radiotherapy plan quality using a double focused, double stacked multi-leaf collimator O. Bohoudi 1 , M.A. Palacios 1 , B.J. Slotman 1 , S. Senan 1 , A.M.E. Bruynzeel 1 , F.J. Lagerwaard 1 1 VUMC, Radiotherapy, Amsterdam, The Netherlands Purpose or Objective MR-guided radiation therapy is an innovative approach for stereotactic body radiation therapy. The concept of double focused, double stacked multi-leaf collimators (MLC) has recently been introduced in the MRIdian Linac system (ViewRay Inc, Mountain View, USA). For selecting the optimal delivery technique, evaluation of plan conformity and dose gradients is essential. Therefore, we compared stereotactic plans for prostate- and pancreatic cancer, generated with this MRIdian Linac system, the 60 Co–based MRIdian system, and volumetric modulated arc therapy (VMAT). Material and Methods The RayZR double focused MLC has 34 leaf pairs in the upper stack and 35 leaf pairs in the lower stack, with each leaf projecting a nominal width of 8.30 mm at 90 cm SAD. The double stack MLC uses an offset between the top and bottom stack to achieve an effective leaf width of 4.15 mm at 90 cm SAD. Plans for stereotactic treatment of localized prostate (N=5) and pancreatic cancer (N=5) were generated with prescription doses of 36.25 Gy and 50 Gy in 5 fractions, respectively. The

Conclusion MRIdian LINAC

plans obtained with a double stacked double focused MLC showed superior CI and GI compared to MRIdian 60 CO plans, and to a lesser extent to VMAT plans. This allows for a better sparing of OARs in stereotactic plans for prostate cancer, and an increase in target coverage while adhering to high-dose constraints for adjacent OARs in pancreatic cancer plans.