ESTRO 2022

Session Item

Optimisation and algorithms for ion beam treatment planning

Session Code: 7006

Session Type: Poster (digital)

Track: Physics

Novel optimization strategies to account for RBE variability in proton therapy

Christian Hahn, Germany

Presentation Number: PO-1729

Abstract

Abstract Title:

Novel optimization strategies to account for RBE variability in proton therapy

Authors:

Christian Hahn^1,2,3,4, Lena Heuchel¹, Jakob Ödén⁵, Jörg Wulff^6,7, Sandija Plaude^6,7, Beate Timmermann^4,6,7,8, Christian Bäumer^9,4,6,7, Erik Traneus⁵, Armin Lühr¹

¹Medical Physics and Radiotherapy, Department of Physics, TU Dortmund University, Dortmund, Germany; ²OncoRay – National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany; ³Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany; ⁴German Cancer Consortium (DKTK), and German Cancer Research Center (DKFZ), Heidelberg, Germany; ⁵RaySearch Laboratories AB, Research, Stockholm, Sweden; ⁶West German Proton Therapy Centre Essen, University Hospital Essen, Essen, Germany; ⁷West German Cancer Center (WTZ), University Hospital Essen, Essen, Germany; ⁸Department of Particle Therapy, University Hospital Essen, Essen, Germany; ⁹Department of Physics, TU Dortmund University, Dortmund, Germany

Show Affiliations

Purpose or Objective

Emerging clinical studies demand to consider relative biological effectiveness (RBE) variability in proton therapy treatment planning. Current clinical planning strategies include the use of an additional beam direction to reduce the number of stopping protons and elevated RBE within organs at risk (OAR) adjacent to the tumour. This study introduces and compares four plan optimization strategies that account for RBE variability without increasing integral dose to healthy tissue.

Material and Methods

A brain tumour patient was considered receiving 54 Gy(RBE), assuming a constant RBE (D1.1), to the clinical target volume (CTV) with pencil beam scanning. The clinically delivered dose-only robustly optimized three-field (3F) plan was used as reference. A dosimetrically competing two-field (2F) plan was clinically accepted.

We used the same two fields and dose objectives as the 2F plan to keep the absorbed integral dose to healthy tissue as low as possible. Novel optimization strategies were applied to reduce the variable RBE-weighted doses (DRBE) to the adjacent brainstem by penalizing either a) the track-end fraction, b) elevated dose-averaged linear energy transfer (LETd) in voxels above a dose threshold, c) dose emanating from protons with LET above an LET threshold or d) DRBE exceeding tolerance dose (Fig.1). Thus, four alternative treatment plans were created using a research version of RayStation (v8.99.30).

D1.1 and DRBE of all plans were evaluated based on plan quality indices (homogeneity, conformity, gradient), dose-volume histogram parameters for CTV and brainstem, integral dose to healthy tissue (median dose to isotropic ring structure from 1 to 3.5 cm around CTV) and normal tissue complication probabilities (NTCP) for brainstem necrosis.

Results

Adequate D1.1 coverage of the CTV was achieved by all optimization strategies with similar homogeneity (range: 0.91-0.92) and conformity (range: 0.88-0.95). The clinical 3F plan showed superior brainstem sparing compared to the conventional 2F plan for both near-max D1.1 and DRBE (both -1.8 Gy(RBE)) while increasing integral dose by around 2 Gy(RBE), respectively (Tab.1).

Compared to the 3F plan, the 2F plans with novel optimization objectives b)-d) achieved lower integral doses and gradient indices while sparing near-max DRBE in the brainstem to the same degree. This was achieved by a combined reduction of absorbed dose and LET in the brainstem. However, using 3F or 2F with additional novel objectives did not substantially reduce NTCP for brainstem necrosis (NTCP range: 0.7-1.2%) compared to the conventional 2F plan (NTCP: 1.2%). Strategy a) did not produce improved DRBE distributions compared to the conventional 2F plan.

Conclusion

With similar CTV coverage and plan quality, 2F plans with novel optimization strategies reduce RBE uncertainty and integral dose in OAR and may replace the clinical 3F plans for selected cases. Still, a third field may allow for better redistribution of stopping protons which may further reduce DRBE to OAR.