TALK TITLE: COMPUTATIONAL RADIOTHERAPY: OR HOW PHYSICISTS AND DATA SCIENTISTS CAN HELP TREAT CANCER
Cancer affects all our lives, either personally or when a friend or family member falls ill. Many common cancers can now be treated with a relatively high success rate. One of the most effective forms of treatment is radiotherapy: the use of high energy x-rays or particle beams. These are very accurately targeted at a tumour with the intention of killing as many tumour cells as possible, whilst sparing the surrounding healthy tissue. Unfortunately, every patient and every tumour is unique. Therefore planning radiotherapy treatment presents a complex data science problem involving image analysis and a statistical simulation of the radiation beam and its interaction with biological matter to calculate dose.
I will describe a programme of work based at Cambridge involving clinicians, high energy physicists, computer scientists, engineers and mathematicians working together to improve the use of radiotherapy. We have addressed a wide range of practical issues in large scale data processing and computational simulation of tumour growth and treatment, with real impacts on patient care and the teaching of medical students. This uniquely interdisciplinary work is set in the context of the wider HPC and scientific computing environment in Cambridge. We host one of the largest academic supercomputers in the UK, pioneered the early development of cloud computing and run an internationally renowned course training the next generation of HPC and data scientists.
Mark Hayes is based at the Centre for Scientific Computing, part of the Cavendish Laboratory at the University of Cambridge. He has over 20 years experience developing complex, innovative technology solutions in scientific computing, web architecture and sensor networks.
His involvement in scientific computing and HPC goes back to 2001 where he was hired to lead the day-to-day research and IT management at the Cambridge eScience Centre, which was one of the regional centres of the UK-wide eScience grid computing programme. Now that the grid has evolved and matured into the cloud, his research activities centre on a collaboration with the Department of Oncology at Addenbrooke’s Hospital in Cambridge, supporting their work in radiotherapy for cancer patients.
In other recent work, Mark has helped build real time air quality sensor networks and used natural language processing to automate the collection of statistics from hospital patient records. He lectures a course on computer hardware for the annual HPC Academy summer school at Cambridge, co-chaired the European Conference on Python in Science (EuroSciPy0 in 2014 & 2015 and was a founding member of the Cambridge Advanced Imaging Centre.