ESTRO 2024 - Abstract Book

S3833

Physics - Image acquisition and processing

ESTRO 2024

To further improve MRI-only RT workflows accuracy and efficiency, attenuation or stopping power maps, for photon and proton-based treatments respectively, could be generated from MRI, to avoid sCT computation and perform the dose calculation directly on the synthetic images.

Keywords: Synthetic CT, MRI, commercial solutions

1371

Digital Poster

Impact of breathing signal-guided dose modulation on 4DCT image reconstruction

Lukas Wimmert 1 , Annette Schwarz 2 , Tobias Gauer 3 , Chrisitan Hofmann 2 , Jannis Dickmann 2 , Rene Werner 1

1 University Medical-Centre Hamburg Eppendorf, Institute for Applied Medical Informatics, Hamburg, Germany. 2 Siemens Healthcare, -, Forchheim, Germany. 3 University Medical-Centre Hamburg Eppendorf, Department of Radiotherapy and Radiation Oncology, Hamburg, Germany

Purpose/Objective:

Respiratory signal-guided 4DCT forms the basis of 4D SBRT safety margin and target definition. However, compared to 3DCT, the patient is exposed to a higher radiation dose. Recently, dose modulation (DM) was proposed for step and-shoot 4DCT [1]: Based on online analysis of the respiratory signal, the CT tube current is reduced during extended end-exhalation (EE) phases to prevent acquisition of redundant projection data with high imaging dose. A breathing signal-based analysis promised a reduction of the imaging dose by 10\%. Although it is crucial for clinical application, the impact of DM on 4D CT image reconstruction has not been assessed so far. In the present study, the DM-impact on 4DCT images is investigated for the first time.

Material/Methods:

The study includes 37 clinical 4DCT scans of patients with thoracic lesions acquired with sequence-scanning 4DCT without DM (i4D, Siemens SOMATOM go.Open Pro; image resolution 0.97x0.97x2 mm$^3$; breathing signal recorded with Varian RGSC system). The 4DCT images were re-reconstructed using EE DM as described in [1]. DM was simulated as reduction of the tube current to 0 mAs, i.e., projection data acquired during DM periods were not used for image reconstruction. Instead, the affected images close to EE were reconstructed based on the projection data not covered by DM and closest to the projection data used for original clinical image reconstruction (Figure 1 top) .The clinical images (no DM) served as ground truth (GT). In addition to a qualitative assessment (visual inspection), the impact of EE DM on image reconstruction was quantified by deformable image registration: The DM-affected phase images were registered to the corresponding GT images. The resulting vector field magnitudes represented the impact of DM on the image reconstructed and were separately evaluated for specific regions-of-interest (ROI; upper/middle/lower lung, heart). Mean displacements and the ratio of ROI voxels with displacements of at least one voxel were evaluated.