ESTRO 2024 - Abstract Book

S3939

Physics - Image acquisition and processing

ESTRO 2024

The AI tool successfully identified setup errors, calibration curve errors, and anatomical changes that can compromise treatment quality. The tool also demonstrated a rapid processing ability, analyzing PR images in just six seconds. The retrospective analysis of the PR images demonstrated a good agreement between simulated and measured integral depth doses in anatomically stable areas, supporting its clinical adoption.

Keywords: Proton Radiography, range uncertainties, AI tool

References:

[1] Oria, C. S., Marmitt, G. G., Both, S., Langendijk, J. A., Knopf, A., & Meijers, A. (2020). Classification of various sources of error in range assessment using proton radiography and neural networks in head and neck cancer patients. Physics in Medicine and Biology, 65(23), 235009. https://doi.org/10.1088/1361-6560/abc09c [2] Meijers, A., Oria, C. S., Free, J., Langendijk, J. A., Knopf, A. C., & Both, S. (2021). Technical Note: First report on an in vivo range probing quality control procedure for scanned proton beam therapy in head and neck cancer patients. Medical Physics, 48(3), 1372-1380. https://doi.org/10.1002/mp.14713

[3] Farace, P., Righetto, R., & Meijers, A. (2016). Pencil beam proton radiography using a multilayer ionization chamber. Physics in Medicine and Biology, 61(11), 4078–4087. https://doi.org/10.1088/0031-9155/61/11/4078

2575

Digital Poster

Title: Evaluation of data-driven motion algorithm OncoFreeze AI in a Siemens Biograph PET-CT

Andrés Fernández González 1 , Francisco Mosquera-Pena Sánchez 1 , Jaime Reverter Pérez 1 , Antonio Teijeiro García 1 , Ines Dominguez Prado 2 , Julio Vázquez Rodríguez 1 , Antonio López Medina 1 , Francisco Salvador Gómez 1 , Fernando Carmona García 2

1 Sergas, Radiophysics, Vigo, Spain. 2 Sergas, Nuclear Medicine, Vigo, Spain

Purpose/Objective:

In PET image acquisitions, the patient's breathing movement can cause blurring of the metabolic tumour volume (MTV) which can lead to an inaccurate estimation of the standard uptake value (SUV). Nowadays, new PET/CT scanner includes reconstruction software that minimize this blurring, as is the case of the OncoFreeze AI system from Siemens Healthineers, a data-driven deviceless respiration gating system. This algorithm extracts respiratory waveform signal directly from the acquired PET data and reconstructs respiration-gated image based on those respiratory waveforms. The purpose of this work is to evaluate the OncoFreeze AI reconstruction software by comparing the SUVmax and MTV of reconstructions using OncoFreeze AI with reconstructions without correcting for respiratory motion, both from a phantom that reproduce the patient's breathing and with real patients.