FLY4PET · Fly Algorithm in PET Reconstruction for Radiotherapy Treatment Planning

Fly Algorithm in PET Reconstruction for Radiotherapy Treatment Planning

During this Marie Curie Career Integration Grant (CIG) the fellow has established a new research topic at Bangor University. He has led his research in an independent manner. So far 10 articles have been published during the duration of this Fly4PET project. At least another 8 are expected. The fellow has organised the Eurographics Workshop on Visual Computing for Biology and Medicine in 2015 (http://www.vcbm.org/). Fly4PET is about Positron Emission Tomography (PET) reconstruction using Artificial Evolution (AE). PET is one of the two kinds of tomographic modality in nuclear medicine. The core principle in nuclear medicine is to administer a radioactive substance (called radiotracer) to patients. Radiotracers are absorbed by tissue in proportion to a physiological process, e.g. the growth of tumour cells in oncology. The reconstruction enables the recovery of the three-dimensional distribution of the tracer through the body. AE is a stochastic optimisation tool that relies on Darwin's principles. It makes use of `genetic operators' based on the biological mechanisms of natural evolution. Our method is called `Fly algorithm'. Each fly represents a point in space and mimics a positron emitter. Their positions are iteratively optimised using genetic operators to closely match the data measured by the imaging system. We have designed a website (http://fly4pet.fpvidal.net/) dedicated to the project. Our algorithm is designed to run on fast supercomputers. To engage with the general public, the fellow and his PhD student adapted the Fly algorithm to produce artistic images (http://tinyurl.com/ho5kfvb). Another research topic investigated in Fly4PET is about fast and realistic simulation of the patient respiration. A software library has been implemented for this purpose. In addition the fellow has developed an OpenSource library, gVirtualXRay, to simulate X-ray attenuation on graphics cards (http://gvirtualxray.sourceforge.net/). It can be used in various applications, such as tomography reconstruction in medical physics, or in virtual reality training simulators. The library provides visually realistic rendering methods of inner organs (e.g. liver, diaphragm, and lungs). Since the start of his position the fellow has contributed to grant applications, 3 of them have been successful. His research network is growing. Most grant applications are in partnership with collaborators that he did not know before the start of this project. He has supervised 1 research assistant and 2 postdoctoral researchers. The fellow is fully integrated within the School of Computer Science at Bangor University. His current obligations include teaching and supervision (23 undergraduate, 9 masters students, and 3 PhD students) as well as administrative duties. He acts as personal tutor. He is Emergency First Aider for the school. He is a member of the staff-student liaison committee. Since then the fellow has been offered a permanent position. He is the school's examination director (inc. preparation and quality assurance of exam papers). He is also the disability tutor of the school, and he now shadows the senior tutor who is retiring next year. As a summary we can say that this Marie Sklodowska Curie CIG has positively contributed to the integration and career development of the fellow.

Data: CORDIS, © European Union