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 deeply honoured. More than anything, this award recognises our team’s extraordinary dedication.

When we embarked on a screening study involving over 20,000 people, we didn’t know whether plasma DNA could detect cancer before symptoms or if that would improve outcomes. For three years, colleagues went out every weekend to recruit participants. We then rescreened the cohort around four years later, including during the COVID-19 pandemic, when restrictions meant team members visited individual households to collect samples. It was demanding work that rarely appears in a scientific paper.

Seeing the effort of our researchers, clinicians, students and participants recognised at this stage of my career means a great deal to me.

Alot of your work has centred on using circulating DNA to detect cancer before symptoms appear. What have been the advances that have made early detection through a blood test increasingly feasible?

Our non-invasive prenatal testing work showed that clinically useful information could be extracted from tiny amounts of cell-free DNA amid a much larger background. In 2009, we used digital PCR to detect EGFR mutations in the plasma of patients with lung cancer.

Sequencing, methylation analysis and DNA-fragment profiling have since made cancer signals much easier to discern. But much earlier multi-cancer technology development relied on case–control studies comparing diagnosed cancer patients, often symptomatic, with people without cancer. Whether a blood test could find asymptomatic cancers in the general population was uncertain.

Screening studies yield relatively few cases and require enormous effort and expense; undertaking ours some 15 years ago was a gamble. In over 20,000 asymptomatic men, we showed that plasma Epstein–Barr virus (EBV) DNA screening could find nasopharyngeal cancer (NPC) predominantly at an early stage. That prospective evidence helped establish a model for multi-cancer early detection.

What do you see as the biggest biological or technical barriers that remain to be overcome for effective non-invasive cancer detection?

The fundamental challenge is biology: very early cancers may release vanishingly little tumour DNA into plasma and circulating DNA is cleared rapidly. We still do not fully understand the mechanisms governing its release, persistence and clearance.

Shedding also varies greatly between tumour types – and even between tumours of similar size – so a test with good sensitivity for one cancer may perform poorly for another. Simply sequencing deeper cannot recover molecules that never reach the sample. We need better knowledge of those mechanisms, improved ways to capture weak signals and realistic, cancer-specific assessments of sensitivity.

From your experience getting diagnostic discoveries into clinical application, what should cancer researchers consider to progress from the laboratory into routine care?

A promising assay is only the beginning. Regulatory approval requires rigorous analytical validation, large clinical studies in the intended population, a clear pathway for investigating positive results and evidence that benefits justify the harms of false positives. This journey takes years and substantial funding.

Researchers should consider intellectual-property protection early, alongside clinical study design. In our experience, strong patent coverage made it possible to attract commercial partners with the resources and expertise to support development and regulatory submissions.

Over 25 years, our team has filed more than 2,500 patent applications covering more than 160 inventions; protecting those inventions has also meant costly, time-consuming litigation. My advice is to build a multidisciplinary partnership – scientists, clinicians, trialists, regulatory specialists and industry – rather than expect a laboratory alone to take a test into routine care.

Where do you think the greatest opportunity lies in cancer early detection, and what would you most like to see the field achieve over the next five to ten years?

Multi-cancer early detection is gaining momentum. GRAIL’s Galleri test received a favourable FDA advisory-committee recommendation in September 2026, although a final FDA decision remained pending. The question now is how clinicians should use such tests, particularly when a blood signal is detected but imaging and endoscopy find no cancer.

In our NPC cohort, a molecular pattern of cancer-associated plasma EBV DNA was linked to a modelled 87-fold higher risk of NPC being identified around four years later than in EBV-DNA-negative participants. This intriguing result needs independent validation and cannot yet be generalised to other cancers. I hope the next decade brings evidence-based guidance for follow-up and personalised surveillance, alongside trials demonstrating fewer advanced cancers and deaths – not simply more positive tests.