Upcoming Workshops (January 2019 - Registration Closed)

 

Radiotherapy datasets are increasing in both size and complexity overtime. On their own these datasets contain information crucial for the treatment of individual patients. Yet aggregate datasets describing hundreds or thousands of patients, can reveal patterns and previously hidden information useful for treating all patients.

The Radiotherapy Machine Learning Network (RTML) aims to bring together a community of experts capable of unlocking this hidden information using state-of-the-art machine learning methods. If you’re a clinician, clinical scientist or a researcher with relevant expertise, why not join us? Together we can improve the lives of patients and perhaps even help beat Cancer. We have three upcoming workshops, and you’re invited. More details to follow. #rtmlnetwork.

Upcoming Workshops (January 2019 - Registration Closed)
Key Dates

Key Dates

 

1st Meeting: October 1st & 2nd 2018 (Done!)
Location: Schuster Annexe, School of Physics & Astronomy, The University of Manchester

Our first two-day event will be held 1st-2nd October 2018, at state-of-the-art facilities located within the University of Manchester. We’re looking for clinicians, clinical scientists to join us, share their expertise and collaborate to solve real-world radiotherapy challenges.

The first event has some modest aims:

  • To familiarise clinicians and clinical researchers with data science, machine learning, and what it takes to deploy cutting-edge methods.
  • To introduce radiotherapy to those unfamiliar with the treatment, it’s importance, and the terminology used by domain experts.
  • To initiate research projects within a simple framework that we, as a community, will work upon at subsequent events.

We aren’t looking for specialists from a domain in particular. Though having prior knowledge of radiotherapy treatment/research is a big advantage. Ideally we hope to engage with experts working with big data, medical data and machine learning more generally. The event programme is TBD at present.

2nd Meeting: January 17th & 18th 2019 (Register now!)
Location: Schuster Annexe, School of Physics & Astronomy, The University of Manchester

Our first event generated 7 projects that aim to apply machine learning to the radiotherapy domain. In our second event, we aim to provide the space to continue working on these projects. We aim to continue building the RTML community. The second event will build on the first, and allow us to continue working together as a community.

3rd Meeting: May 2019 (TBC) 2018
Location: Schuster Annexe, School of Physics & Astronomy, The University of Manchester

The third event will allow us to bring together the outputs of the preceding 2 events, and summarise those in paper format ready for publication.

Workshop format & Outputs

Workshop format & Outputs

Our workshops aren’t going to be completely rigid didactic affairs. Whilst to begin with there will be some didactic delivery (i.e. administration, goals, milestones), we want to provide flexibility which will allow us to work together to produce tangible outputs useful for radiotherapy research. So we’re aiming to foster a comfortable working environment that’s conducive to writing, coding, and thinking. We hope this works for both the introverts and extroverts amongst us, and throughout our events we will listen to the community and dynamically incorporate feedback.

We aim for our meetings to yield high quality publishable outputs. These outputs may take the form of papers, algorithms, software packages, conference proceedings or open source datasets. Whatever their form, these should be linked to Digital Object Identifiers (DOIs) so that participants get visibility for their work. You can get a DOI for your work via a number of tools such as Zenodo. We’ll help with this at our events.

Our work will be made open source too. This is important given that our events our funded via public money. We encourage the use of open source licenses, including those listed at the Open Source Initiative website.

1st Workshop Resources/Documents

Link to the Google document folder with the template presentation / funding request form:

https://drive.google.com/drive/folders/1CqGemD_p6LYTsTXtm45URwb8NSN3ODr4?usp=sharing

2nd Meeting Program

2nd Meeting Program

 

This is an outline programme – it is subject to change!

 

Thursday January 17th

10:00 – 10:30     Arrival, registration, coffee.

10:30 – 10:45     Introduction and aims of the 2nd workshop (Dr Rob Lyon & Dr Tim Rattay)

10:45 – 11:30     Revision of Machine Learning techniques (Dr Rob Lyon)

11:30 – 11:45     Coffee Break

11:45 – 12:45     Talk: Machine Learning for Oncology – Theory and Applications
(Dr John Kang, Rochester, NY, US)

12:45 – 13:45     Lunch

13:45 – 14:45     Group work

14:45 – 15:00     Coffee break

15:15 – 16:00     Group work

16:00 – 16:30     Talk: Decision support system using deep learning in adaptive radiotherapy
(Huan-Hsin Tseng, University of Michigan Ann Arbor, US)

16:30 – 17:00     Talk: Developing an Interpretable, Data-Driven, Model-Based Decision Support System for Personalized Adaptive Radiotherapy in Non-Small-Cell Lung Cancer
(Yi Luo, University of Michigan Ann Arbor, US)

 

Friday January 18th

09:00 – 09:30     Arrival, coffee.

09:30 – 10:00    Talk: Transfer learning for segmentation – application in sarcopenia measurement
(Dr Andrew Green, University of Manchester)

10:00 – 10.45     Group work

10:45 – 11:00     Coffee break

11:00 – 12:00     Group work

12:00 – 13:00     Lunch

13:30 – 14:45     Group work

14:45 – 15:00     Coffee break

15:00 – 16:30     Group presentations and progress update

Data & Resources

Our workshops will give attendees the chance to interact with newly collected radiotherapy datasets. This includes data collected by colleagues working at the The Christie NHS Foundation Trust, as well as REQUITE data. Both these data sources have high discovery potential, and can be explored in many different ways. Some representative data processing tasks that need to be applied to the data include image segmentation/classification, binary classification, multi-class classification, transfer learning, optimisation, imbalanced learning problems, along with a whole bunch of other analytic and predictive problems.

Data & Resources

Workshop 2 Feedback

Location

Location

Location: Schuster Annexe, School of Physics & Astronomy, The University of Manchester

Our events are being held in building number 54, as shown in the University of Manchester Campus Map.

Getting here: The Schuster Annexe is at the heart of the university campus. It’s a 20 minute walk away from Manchester Piccadilly Station. Alternatively it’s a 20 minute walk away from Manchester Oxford Road Station.

More information on getting here, and other forms of travel you can use can be found here.

Logistics

Logistics

Accommodation / Funding: Accommodation costs for non-Manchester based attendees will be covered by RTML funding. We’ll book you into a nearby hotel for 1 night per event. If we become over subscribed, subsidised accommodation will be prioritised according to need (i.e. students first).

Facilities: We’ll be providing data at our events which we can explore together. We also hope to provide groups with compute resources, so they can “hack” on their problems in real-time. More details to follow.

Code of Conduct

We’re dedicated to ensuring our events are an harassment-free experience for everyone regardless of gender, gender identity and expression, age, sexual orientation, disability, physical appearance, body size, race, ethnicity, nationality, religion, experience or lack of.

We will not tolerate the harassment of participants/members in any form. Sexual language and imagery is not appropriate during out meetings. Participants violating these rules may be sanctioned or expelled without a refund (if applicable) at the discretion of the organisers.

Our full code of conduct can be found here.

Code of Conduct derived from the Hack Code of Conduct , and the Python in Astronomy 2016 Code of Conduct.

About the Network

About the Network

The RTML network is a new community bringing together experts from multiple domains, which aims to improve the lives and treatment outcomes for radiotherapy patients. The network is comprised of experts from multiple domains, coming together to pool their knowledge and experience to help better understand large medical datasets. Our hope is that by improving our understanding, we can improve clinical decision making, ensuring patients receive the best treatment possible.

The network was originally conceived during an STFC funded “sandpit” event. There we realised the radiotherapy community would benefit from collaboration with machine learning experts/statisticians, to help realise ambitious data processing goals. Whilst at the same time we learnt that radiotherapy datasets pose new and interesting learning challenges yet to be overcome for machine learning researchers. For both groups collaboration allows for the exploration of new innovative approaches, and the opportunity to develop methods with real-world value.

At present the RTML network is funded by the STFC to deliver 3 workshops spread across 2018 and 2019. If successful, we’ll have the opportunity to apply for more funding in the future. How will we measure success? Primarily via research outputs resulting directly from our events. Success will also be achieved if we’re able to found a community that persists.

Organisers:

Dr. Rob Lyon (Principal investigator), Machine Learning Researcher, School of physics & Astronomy, The University of Manchester.

Dr. Tim Rattay (Principal investigator), NIHR Clinical Lecturer in Surgery, Leicester Cancer Research Centre, University of Leicester. Honorary SpR in Breast and General Surgery, University Hospitals of Leicester NHS Trust.

Dr. Sarah Osman, Centre for Cancer Research and Cell Biology, Queen’s University Belfast. Radiotherapy Physics, Northern Ireland Cancer Centre.

Dr. Andrew Green, Division of Cancer Sciences, The University of Manchester.

Prof. Nigel Mason OBE, School of Physical Sciences, The Open University.

Partners & Member Institutions

Looking for event resources?