ESTRO 2023 - Abstract Book

S922

Digital Posters

ESTRO 2023

Materials and Methods We investigated pCSI for the cranial junction and for the spinal junction. Treatments plans were generated with our in house research treatment planning system TRiP98. The biological effect was estimated by the Local-effect model IV. Figure 1 shows that recalculating a plan for constant RBE with varying RBE leads to a V105>30%. To regain robustness, two new strategies were developed: for parallel fields, opposing ramps in particle numbers were generated from a single biologically optimized treatment plan, utilizing that particle numbers are additive, while dose gradients are not under variable RBE. The second strategy considered direct robust optimization assuming symmetric shifts in anterior-posterior and lateral direction. Furthermore, asymmetric shifts, where one field was shifted cranially and the other caudally (and vice versa) were included. We investigated the impact of the following parameters: field overlap (1cm to 8cm) and spine target including or excluding the vertebral column. The plans were evaluated with two robustness analyses. A global robustness analysis studied 21 position and range uncertainty scenarios affecting each field in equal measure. A local robustness analysis contained nine uncertainty scenarios shifting the fields against each other.

Results The ramp in particle numbers delivered decent results for long overlapping regions (D95>93% for a ramp length of 8cm in all uncertainty scenarios), but fails for short ramp lengths. It cannot be used for the cranial joint since the lateral fields do not have the same LET spectra as the posterior spinal field. Robust optimization performed decent in the cranial junction with a D95>97% in each scenario for an overlap of 1cm, but worse for increasing overlap (D95>94% for 8cm respectivly). Robust optimization for the spinal junction including the vertebral column performed not as good as for the cranial junction, but improved with greater overlapping regions reaching D95>93% in all scenarios for 8cm overlap. If the vertebral column was excluded, robust optimization achieved D95>94%.

Conclusion Exsiting strategies for pCSI result in clinically relevent overdosage if variable RBE is assumed. We presented two new strategies for pCSI, which enable robust treatment planning for a variable proton RBE and open the possibility for a more benefical proton therpay in the future.

PO-1152 Initial and recurrent radiation-induced contrast enhancements following RT for brain metastases

E. Meixner 1 , J. Hörner-Rieber 1 , J.W. Lischalk 2 , T. Eichkorn 1 , A. Krämer 1 , E. Sandrini 1 , A. Paul 1 , P. Hoegen 1 , M. Deng 1 , T. Welzel 1 , S. Erdem 1 , J. Debus 1 , L. König 1 1 Heidelberg University Hospital, Department of Radiation Oncology, Heidelberg, Germany; 2 Perlmutter Cancer Center, New York University Langone Health, Department of Radiation Oncology, New York, USA Purpose or Objective The appearance of radiation-induced contrast enhancements (RICE) after radiotherapy for brain metastases can go along with severe neurological impairments and requires complex interdisciplinary management. Prior studies have reported the benefit of bevacizumab for first line treatment, but the radiological changes and risk of RICE recurrence have not been adequately explored. The aim of our analysis was to evaluate the course and recurrence of RICE and to identify any associated prognostic factors.

Materials and Methods