I was delighted to be awarded the BALR Summer Studentship for 2021. I spent 6 weeks at the UCL Centre for Medical Image Computing (CMIC) working on predicting mortality in idiopathic pulmonary fibrosis (IPF) using computer-derived airway measurements.
Having just completed an intercalated BSc in Mathematics, Computers and Medicine, I was keen to use my newly acquired coding skills in the project, as well as delving into some deep learning algorithms, which I found challenging, but nonetheless fascinating and an important learning experience. Through prior research, I was aware of the importance of quantifying disease severity and progression in IPF. However, this studentship allowed me to dedicate an extended period of time to further develop my understanding of the advancement of the disease, in addition to some of the novel methods currently being explored.
At first, I spent time reading the literature, ensuring I was up to date with the most recent research in the field. I then started to create a deep learning model, working closely with my supervisors. It consisted of coding primarily in Python, but I had the opportunity to use other research software including various 3D segmentation tools to create a ground truth dataset that could be used to train the model. Lobar segmentation proved to be very labour intensive, but it was useful to see so many CT scans, to truly appreciate the specific pathology in IPF. Cox regression was the statistical method of choice for this work, and I had the chance to present this statistical analysis to the rest of the team on my last day. I thoroughly enjoyed preparing and delivering this presentation, and it was a great way to end my summer with the team.
Although the work was computer-based and therefore could be completed remotely, I was grateful to be able to access the lab, where I was both able to work with and meet other members of the team. As most were PhD students, it was extremely valuable to hear about their experiences, and a bit more about their research, which helped me gain insight into their day-to-day lives, which otherwise I would not have been able to experience. In particular, I enjoyed our weekly Friday meetings, where each week we would convene to listen to a short presentation from a member of the team on their recent work, followed by a discussion and Q&A. The work was so varied, and it highlighted to me the importance of sharing ideas, as often suggestions were made during the discussion which provided potential solutions to the problems that some of the researchers had been facing previously.
I would like to thank the BALR committee for giving me the opportunity to work on this project, in addition to enabling me to explore the possibility of a career in research. Additional thanks to Dr Joseph Jacob and Askhan Pakzad for allowing me to work alongside them in the lab, and for providing continual support and supervision throughout the process.
Comments