ESTRO 2024 - Abstract Book

S3677

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

Treatment planning for lung cancer re-irradiation accounting for previously delivered dose

David Walton 1 , Christopher Thompson 2 , Dominic Lowe 3 , Christopher Pagett 2 , John Lilley 2 , Stina Svensson 4 , Kjell Eriksson 4 , Rasmus Bokrantz 4 , Jakob Ödén 4 , Louise Murray 3 , Mark Teo 3 , Ane Appelt 2,5 1 Barts Teaching Hospitals NHS Trust, Department of Radiotherapy, London, United Kingdom. 2 Leeds Cancer Centre, Leeds Teaching Hospitals NHS Trust, Department of Medical Physics, Leeds, United Kingdom. 3 Leeds Cancer Centre, Leeds Teaching Hospitals NHS Trust, Department of Clinical Oncology, Leeds, United Kingdom. 4 RaySearch Laboratories, Research Group, Stockholm, Sweden. 5 University of Leeds, Leeds Institute of Medical Research at St James's, Leeds, United Kingdom

Purpose/Objective:

Re-irradiation is clinically and technically challenging: treatment planning must account for previously delivered dose, while taking anatomical changes as well as radiobiological considerations into account. Most clinical workflows use cumbersome methods, including point dose summation and manual calculation of equieffective doses. For re-irradiation planning, directly utilising the previous dose distribution, mapped to the new CT in 3D, is attractive, but is not available in standard clinical treatment planning systems (TPS). Here, we report on EQD2 based planning for lung re-irradiation, including evaluation of plan robustness to dose mapping uncertainties.

Material/Methods:

Our novel EQD2-based planning approach has been described in detail previously. In brief, dose distributions from previous treatments are mapped onto the re-irradiation CT scan, using deformable image registration (DIR). Re-irradiation treatment plan optimisation uses organ at risk (OAR) objective functions which evaluate the cumulative dose in EQD2, voxel-by-voxel, whilst optimising. This retrospective planning study included six patients previously treated with stereotactic ablative radiotherapy (SABR) for lung cancer, using a variety of schedules, followed by conventionally fractionated re-irradiation. Patient cases were selected based on (potential) overlap between treatment volumes and to avoid major anatomical changes between treatment courses which would render DIR unreliable. Planning goals, including cumulative OAR constraints, and normal tissue α/β values were sourced from the literature (primarily Rulach et al.[1]). No additional tissue recovery was assumed. Re-irradiation plans were optimised using the novel EQD2-based planning approach. Additionally, a standard plan, using a manual planning pathway, was produced for each patient for comparison: for each OAR, the maximum dose D0.1cm3 was determined from the original plan, and subtracted (in EQD2) from the cumulative OAR constraint, to produce the OAR objective for the re-irradiation plan. VMAT plans for 60Gy in 30# were created in RayStation (research version 11A, RaySearch Labs, Stockholm), with individually adjusted arc angles (identical for EQD2-based and standard plans). For plan evaluation, original treatment dose distributions (mapped to re-irradiation CTs using DIR) were summed with re-irradiation dose distributions in EQD2, using OAR-specific α/β values. Plan review was blind to the planning technique and was performed by a senior consultant clinical oncologist, who evaluated the clinical acceptability of all plans, based on achieved target dose and mandatory OAR constraints, and scored the relative preference for the novel EQD2-based plans compared to the standard plans on a five-point Likert scale. Geometric uncertainty of the dose mapping procedure was evaluated using the mean distance to agreement (MDA) between mapped OAR contours. Radiobiological model uncertainties were simulated by varying α/β values within realistic ranges consistent with the literature. We re-sampled the background plan accounting for these