First, congratulations on winning the Early Detection Impact Award! How did it feel to receive that recognition, and what does it mean to you at this point in your career?

I am completely floored and surprised. It means a great deal to get this recognition from the community that is my scientific home.

As a biostatistician in a highly interdisciplinary field – spanning single cells to populations – I appreciate the valuing of the population sciences end of the continuum that this award represents. The focus of my work is on generating the evidence that people, whether they are patients, providers, or policymakers, need to make good decisions about cancer early detection.

Your work has been pivotal in shaping how we model and evaluate cancer screening and early detection. What do you see as the biggest challenges in translating modelling insights into clinical or policy decisions that improve patient outcomes?

Thank you! I think that the main challenge is that models are not by their nature transparent. They typically involve assumptions, mathematical specifications, and structure that can be challenging to explain in detail, so they are much harder to communicate than typical empirical analyses that simply compare survival curves, for example.

This means that modelers have to work hard to describe how their models work so that the audience trusts the results. And while models have become more accepted and more widely used to inform clinical and policy decisions around screening, I think this is still an issue.

Early detection is of course so important, yet it brings complex trade-offs around overdiagnosis and overtreatment. How do you think about striking the right balance between those two sides?

I think we generally don’t get this right in the beginning. In prostate cancer, for example, it took a long time to understand and figure out how to address the fact that we were over diagnosing and over treating too many men.

There may be many ways to achieve the right balance. But key to doing this is to understand the natural history of the disease as this drives both overdiagnosis and benefit. By natural history I am thinking particularly about the opportunity that exists for early detection and interception of the cancer and how this interacts with the performance of the screening test. Knowing this is fundamental to our ability to develop screening programs with a right balance of benefit to harm.

There’s a lot of excitement around new tools for early detection – from liquid biopsies to AI-based screening. what excites you most about where the field is heading?

I think that having a simple test as a first-line screening tool for many cancers is exciting in principle. I am particularly excited about the opportunity that these tests might offer in high-risk populations. Cancer survivors, like my sister, is a great example of population of people who could really benefit.

I think new screening studies will provide information about test performance in the prospective screening setting, and this combined with data for natural history estimation is exciting. Until now we have only had these studies for a relatively few cancers. I am excited about all the new opportunities for modelling that these studies will offer.

I am also excited about exploring new collaborations that will make the data being collected by these studies more widely available. I am excited about working together as a community to figure out the – probably many – right ways to use the new and growing universe of blood tests for early detection of cancer.

What advice do you have for researchers interested in making meaningful contributions to the field of cancer early detection?

Reach out to potential collaborators and mentors, don’t be shy. Be very curious, don’t be afraid to ask questions – there are no wrong questions. Remember that it is hard to have a new idea, so leave no stone unturned when exploring existing work that might be related to your idea.