The LHC and the CREATIS laboratories in Saint Etienne and Lyon open a PhD fellowship on the topic of computed tomography (CT) reconstruction from truncated projections with patient motion. The position is funded by the ARC6 research community of the Rhone-Alpes region. The PhD fellow will be based in the radiotherapy department of the Léon Bérard center which is a hospital focused on cancer care. The thesis will be co-supervised by Simon Rit (CREATIS) and Rolf Clackdoyle (LHC), with additional guidance available from Laurent Desbat (TIMC-IMAG, Grenoble), Catherine Burnier (LHC), and David Sarrut (CREATIS).
Medical and scientific context
Computed tomography is an established medical tool since the 1980s. In radiation therapy, it is not only used for diagnosis and treatment planning but nowadays it is also used to guide patient treatment. To this end, a cone-beam CT scanner is attached to the treatment linear accelator [1]. Two effects can degrade image quality: truncated projection images and patient motion. Reconstruction from truncated projections, known as region-of-interest (ROI) reconstruction, has been actively investigated during the past 12 years [2], and dynamic CT has been investigated for even longer, although with fewer theoretical results [3, 4].
Objectives and research program
The goal of this project is to apply and extend recent theoretical developments on ROI and dynamic CT to projection data acquired on cone-beam scanners in radiation treatment rooms. The application is the reconstruction of high-quality cone-beam CT images that could be used for adaptive radiotherapy. The PhD fellow will first implement existing ROI and dynamic reconstruction algorithms in the Reconstruction Toolkit (RTK) (open source C++) and evaluate their potential on clinical cases (patient images acquired at the Léon Bérard center). In a second step, these algorithms will be combined to simultaneously cope with dynamic and truncated projection data, and studied using computer simulated data, phantom data, and clinical data. During the PhD, new theoretical solutions may emerge that will also be tested on real clinical data.
Qualifications
- Education: The candidate must hold a Masters degree in mathematics or image processing.
- Scientific interests: mathematics (inverse problems and tomographic reconstruction) and computer sciences (medical image processing).
- Programming skills: C++ and Matlab or IDL.
- Languages: English required, French optional.
- Location: Centre Léon Bérard, Lyon, France.
- Salary (net): 1300 euros/month.
- Period: 3 years starting fall 2013.
Contacts
Send CV and a brief statement of interest by email to:
- Rolf Clackdoyle: rolf.clackdoyle@univ-st-etienne.fr
- Simon Rit: simon.rit@creatis.insa-lyon.fr
References
[1] Jaffray, D.; Siewerdsen, J.; Wong, J. & Martinez, A. (2002), 'Flat-panel cone-beam computed tomography for image-guided radiation therapy', Int J Radiat Oncol Biol Phys 53(5), 1337-1349.
[2] Clackdoyle, R. & Defrise, M. (2010), 'Tomographic Reconstruction in the 21st Century', IEEE Signal Process. Mag. 27(4), 60-80.
[3] Desbat, L.; Roux, S. & Grangeat, P. (2007), 'Compensation of some time dependent deformations in tomography', IEEE Trans Med Imaging 26(2), 261-269.
[4] Rit, S.; Sarrut, D. & Desbat, L. (2009), 'Comparison of analytic and algebraic methods for motion-compensated cone-beam CT reconstruction of the thorax', IEEE Trans Med Imag 28(10), 1513-1525.