ESTRO 2024 - Abstract Book

S4031

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

1 Leeds Teaching Hospitals NHS Trust, Medical Physics, Leeds, United Kingdom. 2 RaySearch Laboratories, Research, Stockholm, Sweden. 3 Leeds Teaching Hospitals NHS Trust, Clinical Oncology, Leeds, United Kingdom. 4 University of Leeds, Leeds Institute of Medical Research at St James’s, Leeds, United Kingdom

Purpose/Objective:

Transferring dose from a primary to a secondary scan is prone to uncertainties. Precise knowledge of mapped dose is essential when making informed decisions about organ at risk (OAR) dose limits. Most dose mapping methodologies do not take account of uncertainties in image registration [1,2]. The current work aims to rectify this by using estimates of OAR deformation errors, to inform OAR dose resampling and create a robust mapped organ dose (RMOD).

Material/Methods:

Figure 1 shows an overview of the RMOD process. In brief, this approach uses the difference between a global deformation (approach #2 below) and individual structure-only OAR deformation (#3). From the difference in these deformations, all OAR voxels are assigned an individual resampling uncertainty vector (Dv), from which an ovoid resampling kernel is generated. The position of the kernel centre is assigned as the voxel coordinate, mapped using the inverted individual OAR deformation (#3) back to the primary scan. The kernel is then used to resample doses from within the primary scan OAR contour. Resampled maximum doses based on this information represents a worst case for the mapped OAR dose at each voxel.