Dean Fennell, MD, PhD

”I don’t have the vision of somebody who can see a piece of basic science and know whether it’s going to change the world 20 years down the line: what I do know is that patients need to have their world changed as soon as possible. I look for the kind of discoveries and innovations that will make a difference for the patients who are alive today.

“At the Leicester Mesothelioma Research Programme, our goal is primarily to benefit patients through our research, through the development of new drugs, which can improve outcomes, and particularly survival rates, for patients with mesothelioma—a rare, asbestos-related cancer.

“As a physician-scientist, I focus predominantly on clinical trials and translational research linked to those clinical trials. […] We sequence patients’ tumors not only for DNA and the whole exome, but also for RNA, so we have the entire transcriptome as well. In addition to conventional DNA/RNA sequencing, we’re looking at the immune microenvironment using multiplex immunofluorescence. With this, we’re basically trying to scope out the landscape of the immune cells in the tumor microenvironment, and that’s an area that is moving very quickly—spatial phenotyping—in that we can now see more and more using this type of technology. [...]

“Certainly, the amount of information we’re getting from patients really is quite extraordinary–this is five-dimensional data—and we’re getting a lot of data around the various mutations and architecture of tumors. In all this we want to know one thing: whether the patient responds or doesn’t. So, we are leveraging machine learning quite extensively now in our work, as a tool to help us rapidly analyze reams of data and look for molecular and histological features that we think have some importance.

“Artificial intelligence is very much embedded in what we do. Currently we use basic machine learning techniques more prominently than more advanced methods such as deep learning. This is because deep learning is best suited to very large datasets, which we don’t have yet for things like immunotherapy. However, now that immunotherapy is embedded in practice everywhere in the world, that’s changing very rapidly; there is potential to have very large-scale datasets that we can learn from and employ more advanced AI approaches on. We work very closely with professors of machine learning and computational informatics, and I have to tell you, the speed at which this whole field of data science is evolving is incredibly exciting.”

As told to Christine Parry.