You cannot apply for this job anymore (deadline was 13 Sep 2020).
Browse the current job offers or choose an item in the top navigation above.
The difficulty of detecting or characterizing less common breast cancer subtypes from medical images has long been recognized. The combination of advances in X-ray imaging and artificial intelligence (AI) open up new avenues for this problem. This project will build and validate a combined deep learning and handcrafted radiomics solution for CEM. The new tool will offer decision support for detection and characterization using the quantitative perfusion patterns uncovered by CEM.
The large number of training images required for building and testing AI models is challenging for a relatively new technique such as CEM, where large clinical trials are still absent. This data poverty will be overcome with the creation of an abundant amount of synthetic, virtual cases, including in particular difficult to detect and less common cancer subtypes.
In parallel, eReaders (or model observers) will be tuned to simulate human reader performance. This is a new approach to create relevant ground truth data. Ultimately, it will allow the execution of virtual clinical trials to investigate the clinical impact and cost-effectiveness of using the AI model in the radiological practice.
Our ultimate goal is to offer decision support for detection and characterization of breast lesions. We believe that developing AI models, based partially on synthetic data and tested in virtual clinical trials will support this ultimate goal. In this research project, we investigate the following hypotheses:
This project is a collaboration between the Department of Imaging and Pathology of Leuven University (KUL) and the Department of Precision Medicine of Maastricht University (UM).
During the first two years the PhD student will be employed at KU Leuven. The Primary Investigator of the project, Hilde Bosmans (Scopus Hirsch-index 45, 243 peer reviewed publications), is professor in both the faculty of medicine and the faculty of sciences of the KUL and in the University of Liège, Belgium. Together with her team, she performs the medical radiation physics services in the University Hospital and in several Flemish hospitals, and for a network of 102 mammography units of the Flemish breast cancer screening. A large part of her activities is therefore devoted to breast cancer imaging techniques, and in particular to the development of new and better quality assurance protocols. They have developed the concept of virtual clinical trials to study several aspects of the medical imaging chain, such as the impact of image processing on detectability of lesions. In addition, prof. Nicholas Marshall, medical physicist, Lesley Cockmartin, PhD, biomedical scientist, and prof. Chantal Van Ongeval, breast radiologist, will be involved in supervision of the student with regard to synthetic data creation and clinical correctness and relevance.
During year 3 and 4 of the PhD trajectory, the PhD student will be employed at Maastricht University in the Dpt of Precision Medicine. Prof. Philippe Lambin (ERC advanced & ERC PoC grant laureate, 490 peer reviewed scientific papers, Hirsch Index: 94) co-invented the concept of radiomics (three Nature papers published on the concept, and several of his papers on this topic are cited over 500 times, 2> 1000 times). The group excels at applying machine learning methods such as deep learning & handcrafted radiomics on vast amounts of medical data. Cary Oberije, PhD, senior clinical data scientist, has extensive experience in designing and analysing clinical studies and is responsible for daily supervision of the student. There will be a close collaboration with Dr. Marc Lobbes, MD, PhD, who is breast radiologist in the University Hospital. He is co-PI on this project for the University of Maastricht and world specialist for CEM. He shares the clinical supervision of the project with prof. Chantal Van Ongeval. He will provide all necessary radiological information and experience regarding CEM and studies using CEM. In addition, Bram Ramaekers, PhD, senior researcher on health economics, will be involved on behalf of Maastricht university to guide the health economical aspects.
The applicant has completed a Master in (bio)medical engineering, technical medicine, physics, machine learning, computer science, biomedical sciences or equivalent, with an interest for quantitative imaging, Deep Learning, radiomics and synthetic data.
We are looking for a scientist with a positive attitude, motivated to learn new approaches and ready to work hard to build a scientific career. The candidate should have a sociable personality with good communication skills, a problem-solving attitude, learn fast to plan his own workload effectively and to delegate when necessary and have conceptual ability.
Additional requirements:
Fixed-term contract: 4 years.
You will be appointed and paid as PhD student. We offer full-time employment for a PhD-researcher in an international PhD training program taught in English. The appointment will be for a period of 2 years and will be extended for another 2 years after positive evaluation.
The terms of employment of Maastricht University are set out in the Collective Labour Agreement of Dutch Universities (CAO). Furthermore, local UM provisions also apply. For more information look at the website www.maastrichtuniversity.nl > Support > UM employees.
The School for Oncology & Developmental Biology (GROW) focuses on research and teaching of genetic and cellular mechanisms, as well as environmental and life-style factors that underlie normal (embryonic and fetal) and abnormal (cancer) development. The emphasis is on basic and translational research, aiming at innovative approaches for individualizing prevention, patient diagnosis, and treatment for genetically determined diseases and cancer.
https://www.maastrichtuniversity.nl/research/school-oncology-and-developmental-biology
We like to make it easy for you, sign in for these and other useful features: