Bio: I am a third year MMath Mathematics student at the University of Edinburgh, with an interest in the applications of mathematics and statistics in medicine – how we can use techniques I’ve learnt on my degree to aid in helping the pursuit of medicine. Specifically, I am interested in how all areas of maths – but especially data science and coding – can be applied in cancer care and cancer research, and from this internship I hope not only to gain valuable skills in research and the work place but also hopefully to make a difference to others and discover if this is something I would like to pursue after graduation.
Project: Cancer care genomics with Dr Maggie Cheang at Institute of Cancer Research.
Genomics is now a part of cancer care. The pressing need in clinics is to identify these patients early, before deadly distant recurrence developed, and to track treatment resistance so that optimal treatment plans directed by the tumour biology can be offered. Meanwhile, omics assays are not easily accessible economically in low- and middle- income countries (LMIC); consequently, there is limited adoption of such technologies despite the obvious added health economics value. This highlights the need for a pragmatic tool to prioritise samples for omics analyses within LMIC region.
The machine learning methods will include a series of semantic segmentation approaches to automatically identify the area of tumour on a H&E slide and use this as the basis for patch generation. We would apply state of the art deep learning/neural network methodologies-based pipelines on a pilot of 600 digitised pathological images to integrate selected tumour morphological features, to identify novel image features associated with good/poor quality of samples for downstream biomarker analyses, to integrate with omics data to predict patient outcome and finally implement digital pathology as a routine workflow for new biomarkers in clinical trials.
