Congratulations on your recent appointment as Director of the CRUK Manchester Institute! How have you found the position and what are the big challenges in taking it on?

What has struck me most so far is the strength and breadth of the science here, and the genuine appetite for collaboration between basic researchers, and clinicians.

The main challenges are ones you would expect for a leading cancer research institute at this moment in time: ensuring we continue to deliver world-class, impactful science while the research and funding landscape becomes more competitive; recruiting and retaining outstanding talent; and making sure that our infrastructure, from core technologies to data platforms, is fit for the next decade of discovery.

Another important challenge is integration: making sure that what we do in the laboratory is tightly connected to the questions being asked in the clinic, and that patients ultimately feel the benefit of the discoveries made here. My focus in the first phase will be on listening, understanding where we are strongest, and then working with teams to define a clear, shared strategy for the Institute’s next chapter.

Liquid biopsies are already making it more realistic to capture tumour evolution over time without repeated invasive procedures. Serial sampling of circulating tumour DNA can tell us when resistance is emerging, which clones are expanding, and whether minimal residual disease remains after treatment. I expect this to become a standard tool in many tumour types but there are some where it is not yet tractable such as kidney cancer.

In the research setting we have been relying on multiregional tumour sampling which is labour and cost intensive. Our lab team have developed tumour homogenisation and representative sequencing as a potential tool to create a more representative sample from left over surgical tissue. We are testing its feasibility at the moment.

Trial design needs to embed heterogeneity and longitudinal sampling from the outset. If we routinely collect multi-region and serial samples in clinical studies, we can build the evidence base that regulators and guideline panels need in order to endorse these approaches.

Over time, I think the language will shift. Instead of saying, “This patient has X cancer with mutation Y,” we’ll talk about tumour trajectories – where the disease is likely to go next – and use that to guide therapy.

As the realm of epigenetics has grown in potential importance, has this changed how you think about classical evolution in tumours?

Epigenetics has broadened, rather than replaced, my view of tumour evolution. Heterogeneity is not just about different mutation profiles; it also includes diverse cell states within the same genotype. This state diversity can act as a reservoir of potential resistance. Classical models focused heavily on genetic mutations as the heritable units of selection. We now appreciate that epigenetic states – patterns of chromatin accessibility, DNA methylation, and histone modification – can also be heritable over many cell divisions and profoundly influence phenotype.

Epigenetic changes can occur and revert more quickly than genetic mutations, allowing tumours to “sample” phenotypic space more rapidly in response to environmental pressures, including therapy. Some resistant phenotypes may be driven less by fixed genetic resistance mutations and more by a plastic, epigenetically mediated state. That suggests we might be able to re-sensitise tumours by targeting the epigenetic machinery rather than only chasing individual mutations.

In the context of the TRACERx Renal study we are currently mapping epigenetic changes through tumour evolution using technologies such as CUT&Tag and ATACseq – watch this space!

As someone who has worked closely with large-scale consortia and data-driven research, how do you think we should approach the ethical and technical challenges of sharing patient-derived data?

Real-time sharing of patient-derived data is essential if we want to accelerate discovery and improve care, but it raises legitimate concerns around privacy, consent, and governance.

We have thought a lot about this in the context of the MANIFEST programme where we are integrating high dimensional data from individual patients to understand how they respond to immunotherapy. I think there are a few core principles we should adhere to: Trust and transparency with patients; Robust de-identification and governance; Interoperable, federated infrastructures.

If we get these elements right, real-time data sharing can become a virtuous circle: patients contribute their data knowing it will be used responsibly; researchers gain access to richer, more timely datasets; and the insights generated can be returned to the clinic more quickly, closing the loop between care and discovery.