Turing 'Spotlight': 14 December
The Digital Futures team are hosting talks by The University of Manchester's new Turing Fellows for 2021-22.
To register for this event, please visit eventbrite.
Marcel van Herk. Marcel is Chair in Radiotherapy Physics and is responsible for developing a programme of international leading radiotherapy physics research and innovation to deliver direct patient benefits with The Christie NHS Foundation Trust. The main focus is on accuracy of radiotherapy including target volume definition, treatment planning, image guidance and treatment follow-up.
Marcel's talk is titled: 'Radiotherapy, a precision treatment for cancer patients - optimisation using big data analysis and AI'.
My research focuses on increasing radiotherapy accuracy to better target the tumour and spare healthy organs. My most impactful work is a system for 3D and 4D CT guided radiotherapy, improving radiotherapy precision for millions of patients. While physicians were initially sceptical of using 3D and 4D imaging in a daily workflow (before only 2D was used), the workflows were found to be highly efficient, and are now standard of care. Given the high accuracy that is achievable, it is recognised that other factors have become limiting. Recently, we analysed a large cohort of lung cancer patients to correlate small residual beam placement errors with patient survival. To our surprise, such correlation was found, and attributed to a higher-than-expected sensitivity of a specific region of the heart to radiation, measurable because these errors are unconfounded. This finding has led cancer centres to re-assess their clinical practice and is the basis for our NIHR-funded rapid learning study. Another aspect is that many cancer patients come to use for treatment with multiple morbidities. I therefore aim to ML-based quantitative imaging biomarkers that can detect conditions such as sarcopenia, cardiac comorbidities, liver cirrhosis, and poor lung function in routine acquired images. These biomarkers can augment and potentially replace qualitative patient assessments used to select the most appropriate treatment for each patient (O'Connor 2016). My research is relevant to the Turing priority research areas optimisation of clinical trials and optimisation and individualisation of treatment.
David Wong. David is Lecturer for AI in Healthcare. His research interests focus on investigating the impact of using continuous signals to assist clinicians in diagnosing and monitoring patients.
David's talk is titled: 'Towards remote and contactless assessment of movement disorders'.
Visual assessment of patients is a core component of clinical medicine. Experienced clinicians often report a gut-feeling of ‘knowing’ when a patient is unwell at first sight. Can we replicate this intuition by extracting information from videos? In this talk, we will explore some of the ways that clinical information has been gleaned from videos over the last 5-10 years. We will then look at some of our recent work in trying to measure and classify some symptoms of Parkinson’s disease. Finally, we will think about some of the practical challenges in developing these methods for real-life clinical practice.