Use this URL to cite or link to this record in EThOS:
Title: Data-driven methods for respiratory signal detection in positron emission tomography
Author: Bertolli, Ottavia
ISNI:       0000 0004 7231 448X
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Positron Emission Tomography (PET) is a nuclear medicine imaging technique which allows quantitative assessment of functional processes, by determining the distribution of radioactive tracers inside the patient body. It is mainly used in oncology. Respiration during PET data acquisition of the chest leads to blurring and other artefacts in the images, lowering their quantitative accuracy. If a respiratory signal is available, these issues can be overcome by splitting the data into different motion states. In current clinical practice this signal is obtained using external devices. However, these are expensive, require prior setup and can cause patient discomfort. This thesis develops and evaluates Data-Driven (DD) techniques based on Principal Component Analysis (PCA) to generate the signal directly from the PET data. Firstly, the arbitrary relation between the sign of the PCA signal and the respiratory motion is addressed: a maximum in the signal could refer either to end-inspiration or end-expiration, possibly causing inaccurate motion correction. A new correction method is proposed and compared with two already existing methods. Subsequently, the methods are extended to Time-of-Flight (TOF) PET data, proposing a data processing step prior to using PCA, in order to benefit from the increased spatial information provided by TOF. The proposed methods are then extensively tested on lower lung patient data (non-TOF and TOF). The obtained respiratory signal is compared with that of an external device and with internal motion observed with Magnetic Resonance Imaging (MRI). Lastly, to investigate the performance of PCA where respiratory motion is minimal, the methods are applied to patient and simulation data of the upper lung, showing that they could potentially be utilised for detecting respiratory-induced density variations in the upper lung. This study shows that the presented methods could replace external devices for obtaining a respiratory signal, providing a simple and cost-effective tool for motion management in PET.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available