Proton radiotherapy Imaging Reconstruction
Lead Research Organisation:
University of Surrey
Department Name: Vision Speech and Signal Proc CVSSP
Abstract
Novel image reconstruction for proton imaging to improve planning during proton beam therapy
Publications
Panagiotidou M
(2019)
Integration of Proton Computed Tomography into the Open Source Software STIR
Panagiotidou M
(2019)
Proton Computed Tomography: A Case Study for Optimal Data Acquisition
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509772/1 | 30/09/2016 | 29/09/2021 | |||
| 1947707 | Studentship | EP/N509772/1 | 30/09/2017 | 31/01/2021 | Margarita Panagiotidou |
| Description | Throughout the entire period of my PhD, extensive knowledge has been acquired for the proton computed tomography field which has led to key findings. Previous work relevant to my research has been thoroughly investigated through the evaluation and validation of existing reconstruction algorithms. Misleadings related to the proton CT data collection have been overcome and discussed with experts of the field. New research networks and collaborations with teams from different Universities and research centers have been established (University of Birmingham, PRaVDA team, Proton CT collaboration, NPL), with an exchange of work and ideas. These collaboarations led to new research questions which are under investigation and further research. The reseach of the 2nd year has been focused more on the data set that has been available to the proton CT group in Surrey from the Proton CT collaboration in USA. A first comparative study has been done among different reconstruction methods for proton CT listmode data, i.e BPF, FBP and DROP-TVS, for different phantoms. This work has been accepted as a conference proceeding for the 2019 IEEE NSS/MIC conference. Another work which was investigating on the side, was the potential integration of the proton computed tomography into an Open Source Software, called STIR. STIR is a software for use in tomographic imaging with the aim to provide a Multi-Platform Object-Oriented framework for all data manipulations in tomographic imaging. Currently, it works for iterative image reconstruction in PET and SPECT. Our aim is to gradually implement the proton CT imaging modality into STIR in order to provide a complete open source software for proton CT data manipulation and reconstruction. An initial study has been accepted as a conference proceeding for the 2019 IEEE NSS/MIC conference. Recently, more focus has been given on new developed iterative recostruction methods with different optimisation schmes and transformation operators for image denoising. The split-Bregman method will be implemented as a convex optimisation method, while the wavelets and shearlets will be examined and compared with the total variation method and the fourier transform for the better denoising of the reconstructed images. |
| Exploitation Route | Considering that the proton CT implementation is yet far from clinical application, and also that this research is still ongoing, I have to say that the outcomes are still in a questioning stage and they need further investigation. However, at the end of this research we aim to provide new knowledge about previous exisiting work in order to avoid techniques and the use of available datasets which seem that they do not work properly for the needs of the proton CT field and also make available a new resonstruction algorithm which hopefully will provide a better reconstruction and improve the accuracy of the treatment planning. |
| Sectors | Healthcare |
| Description | FEPS Faculty Research Support Fund (FRSF) Award |
| Amount | £1,000 (GBP) |
| Funding ID | PGR CG 19-124 |
| Organisation | University of Surrey |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 03/2019 |
| End | 11/2019 |
| Description | São Paulo School of Advanced Science on Learning from Data |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Open presentation of my research using the "one minute strategy" to mixed audience of more than 300 researchers, post-doctoral fellows and graduate students. This session included an interaction among all the attendees and the discussion of our research with a large audience. |
| Year(s) Of Engagement Activity | 2019 |
| URL | https://sites.usp.br/datascience/spsas-learning-from-data/ |