1st MaThRad Clinical Workshop
9 Feb, 9:00 to 10 Feb, 17:30
Friends House, London, UK
The first MaThRad clinical workshop invited internal and external speakers to present ideas and concepts from their respective disciplines to promote discussions between clinical scientists, nuclear engineers and mathematicians, with the idea that these would be formulated into questions to be addressed through long-term research projects.
The workshop was organised by Dr. Sarah Osman (UCLH), Dr Tristan Pryer (University of Bath), Ana Lourenco (National Physical Laboratory), and Colin Baker (UCLH); and invited speakers from a variety of institutions and companies including: Queen’s University (Belfast), Aarhus University Hospital (Denmark), and the National Institute of Radiological Sciences (Japan).
In between these talks there were numerous rounds of discussion between individuals from multiple different disciplines which would allow for new perspectives on challenging problems. The majority of points discussed (that are within the scope of MaThRad) were related to the following topics:
Robust Optimisation, which involves computing the optimal proton beam positions and intensities, such that an associated cost functional of the dose is minimised with respect to uncertainties in the system. These uncertainties come from many different sources, such as: variations in the patient anatomy, artifacts in the CT scan, and uncertainty in the delination of the tumour.
Bayesian Update Models, which involves using the data from particle detectors surrounding the patient to alter the treatment plan in real time. As a proton beam is fired at a patient, gamma particles are emitted through a proton-nucleaus collision (referred to as non-elastic scattering); through the frequency of emission, the energy and the angle of trajectory, we utilise the methods of Bayesian analysis to create a distribution of the delivered dose and of the structure of the interior of the patient.
Monte Carlo Simulations, which are the current “gold-standard” for treatment planning, utilising random sampling to simulate how a proton beam scatters in a patient. However, current Monte Carlo codes are computationally expensive, which makes them not ideal for computing rapid changes in the treatment planning; thus a mathematical understanding of how small pertubations in the beam position may accelerate such codes.
List of Talks: Day 1
|The Math behind MaThRad and the relation to clinical challenges||Tristan Pryer||University of Bath|
|Particle Therapy: Advantages, disadvantages, current status and challenges||Colin Baker||University College London Hospitals|
|Treatment Planning Photon vs Proton: Differences and Challenges||Vasilis Rompokos||University College London Hospitals|
|Radiation Damage: Radiobiology||Kevin Prise||Queen’s University Belfast|
|The Geant4-DNA project: a Monte Carlo approach for evaluation of DNA damage and cell survival||Dousatsu Sakata||National Institute of Radiological Sciences, QST, Chiba, Japan|
|Robust optimization and probabilistic planning (Photons)||EM Vasquez Osorio||University of Manchester/Christie|
|Robust optimization and probabilistic planning (Protons)||Toni Lomax||Scherrer Institut (PSI) -Switzerland|
List of Talks: Day 2
|MC calculations for RT applications||Ana Lourenco||National Physical Laboratory|
|Clinical TPS dose engines – Raystation||Erik Traneus||RaySearch Laboratories AB|
|Challenges in dose calculation for radiotherapy planning||Anne Vestergaard||Aarhus University Hospital, Denmark|
|Scone: A flexible computational infra–structure for testing and prototyping ideas for structure for testing and prototyping ideas for MC||Eugene Shwageraus||University of Cambridge|
|Dose delivery verification for proton beam therapy and detectors||Simon Jolly||University College London|
|MCBEND Monte Carlo code for radiation shielding||Adam Bird||JACOBS|
Math behind MaThRad related to clinical challenges
Tristan Pryer and Alex Cox (UoB)
The first talk was given by the two lead mathematicians at Bath. In the first half of the talk, Dr. Pryer gave an introduction to the mathematical aspects of MaThRad, with particular focus on the Boltzmann transport equation (BTE). He explained the properties of the BTE, gave an overview of the Fokker-Plank approximation (which introduces a diffusion term to represent the elastic scattering), and then discussed sensitivity analysis of the Bortfeld’s model of the Bragg peak that he, and a former student of his, had conducted.
The second half of the talk, Prof. Cox discussed the application of Bayesian Analysis to the estimation of delivered dose through prompt gamma emission’s. He explained the mathematical principles behind how a new detected gamma emission would update the underlying distribution of the position of the tumour, and elabortated on how this could be used to predict the success rate of a given procedure.
Radiation Damage – Radiobiology
Kevin Prise (Queen’s University Belfast)
Kevin Prise is a professor of radiation biology at the Patrick G Johnston centre for cancer research, and he gave a presentation about how the differences in the relative biological effectiveness (RBE) of the different types of radiotherapy; giving a detailed explanation for how the ionising radiations causes breaks in the DNA strand, which in turn, may lead to the development of secondary cancers. He demonstrated how the Bragg peak of the proton beam was correlated with the linear energy transfer (LET) curve, which describes the rate at which energy is lost by the proton along its track. When substantial amounts of energy is deposited locally this leads to increased biological damage to the surrounding tissue.
SCONE, a flexible computational infra-structure for testing and prototyping ideas for Monte Carlo
Eugene Shwageraus (University of Cambridge)
Professor Eugene Shwageraus is the Nuclear Energy MPhil Programme Director, and he gave a presentation regarding the open-source Monte-Carlo that he and his team developed to simulate the transport of neutrons through a nuclear reactor. This program, called SCONE, is presently the fastest open-source MC neutron-transport simulator available; the code was designed to have an intutive interface (which would allow those unfamiliar with the program to learn quickly) and a modular code design (which would allow the code to be developed and applied into new area’s). One point of future development would be the creation of a proton transport module which would have direct clinical application.