Skip to Main content Skip to Navigation
Theses

Caractérisation, optimisation et simulation des performances d’un TEP-TDM numérique

Abstract : Positons emission tomography (PET) is nowadays the most sensitive functional imaging modality for analyzing molecular interactions in the human body. In constant evolution, this modality has recently benefited from silicon photomultisers (SiPM) to replace conventional photomultipliers (PMT), resulting in an overall performance improvement, especially in the temporal resolution used for time-of-flight (TOF) imaging. The performance of the Vereos PET (Philips) involving digital SiPMs (d-SiPM) was evaluated both on phantoms and in clinical conditions by direct comparison with the Ingenuity PET (Philips), involving PMT and comparable acquisition and reconstruction processes. Thus, compared to its analog predecessor, we demonstrated that the Vereos PET offers an improvement in 1) spatial resolution through 1:1 coupling between scintillators and d-SiPM, 2) detectability and contrast-to-noise ratio (CNR) of small lesions due to a better TOF resolution, and 3) in the stability of performances in the upper routine counting rates, due to higher numbers of photodetectors and trigger circuits, thereby reducing dead time and pile-up effects. While the gain in CNR provided by the TOF information is particularly useful for PET examinations performed with low statistics (reduced injected activities and/or recording times), the stability of performance is particularly advantageous in the dynamic exams for which the recorded activities can be initially high. In addition, the Vereos PET was modeled using the GATE platform and this modeling was validated while providing values very close to those obtained experimentally for the NEMA evaluation of the Vereos PET. Finally, a reconstruction environment specific to the Vereos PET and incorporating all the necessary corrections to obtain quantitative images was developed using the CASToR software. These direct (GATE model) and inverse (CASToR reconstruction) modeling of the Vereos digital PET scanner has enabled us to confirm the influence of certain confounding factors for image quality and they could be useful in the future, especially for optimizing imaging performance and evaluating new methods of image reconstruction and correction.
Complete list of metadata

https://hal.univ-lorraine.fr/tel-03285157
Contributor : Thèses Ul <>
Submitted on : Tuesday, July 13, 2021 - 10:31:01 AM
Last modification on : Thursday, July 29, 2021 - 2:43:21 PM

File

DDOC_T_2020_0294_SALVADORI.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-03285157, version 1

Citation

Julien Salvadori. Caractérisation, optimisation et simulation des performances d’un TEP-TDM numérique. Imagerie médicale. Université de Lorraine, 2020. Français. ⟨NNT : 2020LORR0294⟩. ⟨tel-03285157⟩

Share

Metrics

Record views

46

Files downloads

44