ESTRO 2024 - Abstract Book

S3807

Physics - Image acquisition and processing

ESTRO 2024

In this study, considerable variations in amplitude, in-window duration and end-exhale position deviation were observed, which were strongly correlated to imaging efficiency. Respiratory cycle duration (i.e., breathing frequency) did not correlate with imaging efficiency. Retrospective motion-compensation window positioning would have improved imaging efficiency, while it would have shortened in-window durations. This can be explained by the observation that the window was in 10 out of 11 cases positioned more caudal to compensate for drift and thereby obtain the maximum efficiency. Consequently, parts of the end-exhale phase of the 1D-RNAV signal did fall outside the window, shortening the time between entering and exiting the window. A potential risk of using the retrospectively determined window is that the image data is no longer acquired during the end-expiration phase but during a phase where more respiratory motion occurs, which can result in blurrier images. Therefore, in future, we aim to prospectively regularize the breathing pattern in patients using external biofeedback², which makes imaging more efficient and more reproducible between patients.

Keywords: Motion-compensated MRI, Biofeedback, Lung cancer

References:

1. Pipe, J.G. Motion correction with PROPELLER MRI: application to head motion and free-breathing cardiac imaging. Magnetic Resonance in Medicine. 1999; doi: 10.1002/(sici)1522-2594(199911)42:5<963::aid-mrm17>3.0.co;2-l

2. Keijnemans, K. et al. Effectiveness of visual biofeedback–guided respiratory ‐ correlated 4D ‐ MRI for radiotherapy guidance on the MR ‐ linac. Magnetic Resonance in Medicine. 2023; https://doi.org/10.1002/mrm.29857

829

Poster Discussion

3D CT Reconstruction from biplanar projections with integration of planning CT

Alexandre Cafaro 1,2,3 , Amaury Leroy 1,2,3 , Guillaume Beldjoudy 3 , Pauline Maury 2 , Alexandre Munoz 3 , Charlotte Robert 2 , Vincent Lepetit 4 , Nikos Paragios 1 , Vincent Grégoire 3 , Eric Deutsch 2 1 TheraPanacea, AI Research, Paris, France. 2 Gustave Roussy, Inserm 1030, Paris-Saclay University Department, Villejuif, France. 3 Centre Léon Bérard, Department of Radiation Oncology, Lyon, France. 4 LIGM, Ecole des Ponts, Univ Gustave Eiffel, CNRS, AI Research, Marne-la-Vallée, France

Purpose/Objective:

In radiotherapy, bi-planar projections, inherent in stereotactic treatments, have merits but also limitations. They provide a 2D view of the patient's anatomy with focus on bones, complicating the representation of 3D soft tissue and movement. This limitation can lead to inaccuracies in targeting tumors or sparing structures, particularly when dealing with mobile or anatomical changes throughout treatment. By combining bi-planar X-ray projections, which provide real-time and readily available data, and priors capturing patient’s anatomy at an early stage with the power of generative AI, it becomes possible to create daily 3D volumes that accurately depict the patient's anatomy. This