Abstract Book

ESTRO 37

S553

Results Optimized Halcyon plans which include the 10 MU MVCBCT compare well with optimized plans that do not include MV imaging. . Fig. 1 presents the dose clouds and DVH of the Halcyon and Tx-only optimized plans, showing very similar results for the PTV and OAR doses. This is likely because both the MVCBCT and the treatment fields use the same isocenter and therefore primarily deliver dose to the same regions, and because the Halcyon optimization process takes into account the MVCBCT imaging dose. Nevertheless, subtracting the Tx-only plan from the Halcyon plan, there is MV imaging dose outside the target (Fig. 2), which amounted to ~0.5 - ~2 Gy.

The user need to be mindful of the imaging field length as well when it comes to increase OAR doses.

PO-0994 A fast automated sanity check for online plan adaptation in MR-guided RT J. Kaas 1 , W. Van den Wollenberg 1 , A.J.A.J. Van de

Schoot 1 , F.W. Wittkämper 1 , T.M. Janssen 1 1 Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands

Purpose or Objective Clinical introduction of the MR-Linac (Elekta AB, Stockholm) includes online plan adaptation (i.e. simple dose shift) to correct for setup errors due to the fixed couch position. Pre-treatment dosimetric verification of the adapted plan will be unavailable with the patient on the couch. The online adaptation workflow is new and complex, so potentially prone to errors. Therefore, there is a clear clinical need for a fast QA tool to verify online adapted plans. The aim of this work was to develop a tool to perform an automated "sanity check" on online adapted plans to avoid gross errors. Material and Methods In the MR-Linac workflow, a reference plan is made offline, which can be dosimetrically verified pretreatment. Any adapted plan sufficiently similar to it will be considered safe to deliver. To determine the similarity between the reference plan and the adapted plan, we introduce the following metrics: total number of MU, largest segment area, the MU-weighted sum of segment areas and the center of mass of the fluence distribution, determined from the centers of mass of the MLC apertures of each beam, weighted by MU and corrected for the FFF profile shape. The difference in fluence center of mass between the reference plan and the adapted plan is compared to the known applied online plan shift. To judge the stability of these similarity metrics, a test data set of three rectal cancer patients and three prostate cancer patients was created. Reference IMRT plans were created in Monaco (v5.19, Elekta AB, Stockholm), and adapted based on simulated setup errors of up to 30 mm in steps of 3 mm (prostate) or 10 mm (rectum) in all 6 cardinal directions. Each adaptation was performed in 2 ways: Segment Weight Optimization (SWO), where segments are only rigidly translated and weights (MU) reoptimized, and Segment Shape Optimization (SSO), where segments are also reshaped, resulting in 466 adaptations in total. Also, to test whether errors can be caught, a H&N patient with a large (in CC-direction) PTV was adapted based on setup errors that take part of the PTV out of reach of the beam (due to MLC limits), resulting in poor PTV coverage. Results Observed deviations were small (Table 1). In particular, the MU-weighted sum of segment areas is well conserved during adaptation, with only -2% to +3% difference between reference and adapted plan. The fluence center of mass shift closely tracked the setup error, with differences less than 2.5 mm for both SWO and SSO. For the intentionally bad H&N plan adaptations, larger deviations from 3 mm up to 9 mm in the center of mass of the fluence distribution were observed (Figure 1).

Fig.1: Halcyon (squares ) vs. Tx only ( triangles ), showing similar DVHs.

Fig. 2: MVCBCT dose outside the target region.

Conclusion This is the first assessment of the MVCBCT imaging dose outside the target volume of H&N cases for the first clinical Halcyon in our institution. Our results show that Halcyon’s optimization process fully accounts for the MV imaging dose in generating the treatment fields so as to satisfy the clinical prescription. Imaging dose evaluation should be performed for treatment sites where OARs are contralateral to the treatment field arrangement since these OARs may not be included in the plan optimization.