Harnessing the power of data for paediatric cancer research

Dr Emma Woodward, a clinical geneticist, works with patients who have hereditary conditions that make them more likely to develop cancer. “Just 30 years ago, a lot of children with these conditions wouldn’t survive their first primary tumour. Modern medicine is fantastic,” she says. “It’s also unleashed a lot of challenging clinical questions that we need to get better at answering.”

The seed of a new research project idea was planted during an appointment with a young adult who, after surviving a primary tumour as a child, had been diagnosed with a second primary tumour. “They had questions I couldn’t answer about their risk and survival. I was sure the answers would be in a database somewhere – but I had no way to access that information.”

Emma’s team recently received a Data for Children’s and Young People’s Cancer (Data4CYP) Pilot Award: £250k over two years to develop new, data-driven solutions to common challenges in children’s and young people’s cancers.

Together with the research community, CRUK has developed a programme that channels the ambitions of our Research Data Strategy to support data-intensive CYP research. This programme secured a £5 million investment to be spent over the next five years. Eight new projects have been funded by CRUK, with four of those jointly funded through partnerships – two with Great Ormond Street Hospital Children’s Charity and two with Children with Cancer UK.

“At CRUK, we have a dedicated research strategy to drive increased survival and long-term quality of life for children and young people with cancer. And it’s clear that an enabler of this research is data,” explains Dr Laura Danielson, CRUK’s research lead for children’s and young people’s cancer. “Advances in data sciences mean we’re now able to answer questions, like Emma’s, that we’ve never been able to even ask before.”

Speaking with Emma, Laura and other recipients of the Data4CYP award, there is a sense of momentum and hope: as if the research community is at a tipping point in overcoming some of the challenges that continue to frustrate progress for children affected by cancer and their families. There’s a wealth of data at our fingertips – and we’re on the cusp of unlocking its potential.

The Data4CYP awardees come from a breadth of computational, clinical and experimental backgrounds. We caught up with four of them to explore what they hope to achieve and where their pilot awards could take us.

The team will explore whether it’s possible to link  individuals’ pseudonymised genomic data – captured through diagnostic and predictive tests and held by the NHS – with clinical records held by the National Cancer Registry and Analysis Service, including layers of family history, personal history and other risk factors. Initially, they’ll focus on TP53 and RB1 – two genes they’re confident could prove the principle of their idea.

Thanks to advances in genomic technology, and increasing survival rates for children and young people with cancer, data often now exists that spans from before a child is diagnosed with cancer, through their treatment journey and on to long-term survival, relapse or new primary tumour. Emma is hopeful that linking these data is the key to supporting clinicians to calculate personalised risk for their patients, so they can put in appropriate steps to support prevention or early cancer detection.

“I truly believe risk prediction and early detection are the most effective ways we can improve cancer outcomes – but only if they come hand in hand,” she says. Understanding risk in children and young people offers unique challenges, but also opportunities when it comes to keeping them healthy for the long-term. “We’re looking at life-long risk prediction and clinical decision making to support the kids in our clinics for the rest of their lives.”

Most children with cancer in the UK receive hub-and-spoke model care, with treatment coordinated by an oncology centre and delivered by district and community teams. “This means diagnostic tests, blood work, toxicity data… it’s all sitting in different places,” describes Andi. “The only way to know if we’re treating our patients in the best way is to pull all this data together into one place.”

The team hopes to use their pilot funding to create a national database that can longitudinally collect and store this data. In parallel, they’re performing tests on patient samples to enrich the dataset – such as immunophenotyping, single-cell RNA sequencing and multi-omics assays.

Beyond the clinic and the lab, Andi and the team want to understand the impact of treatment from a family’s perspective – qualitative data that isn’t usually captured. “We want to know about the days the children are too sick for school or nursery, or their sleep is affected, or their parents couldn’t go to work,” she says. Working with parent advocates and data strategists, they plan to develop a portal for parents to submit this information in a way that maximises interactive usage while preserving patient privacy.

Eventually, machine learning will be used to analyse this database of clinical, molecular and real-world data, looking for clues that could help build a holistic understanding of iALL – ultimately hoping to improve patient stratification and treatment to improve disease outcomes.

It might sound ambitious for a pilot project – but Andi’s eyes are firmly on the bigger picture. And with other childhood cancers and rare diseases presenting similar challenges of disjointed data, the team hopes their efforts could be easily scaled to other conditions. “We’ll spend these first two years building the pipelines, but if we’re going to improve outcomes for our patients, we’ve got to think long-term,” she says. “These are rare diseases, and too often they’re neglected. But for the families affected, they’re everything.”

– fuelled by an interest in using computational tools to understand how mutations cause cancer through physical and behavioural changes to a cell. The award represents Ben’s first foray into the field of children’s and young people’s cancer, sparked by a new collaboration with joint lead applicant Professor Tariq Enver.

Together, the team is using computational modelling to understand how mutational and epigenetic changes influence cell behaviour in childhood B cell precursor acute lymphoblastic leukaemia (BCP-ALL).

“BCP-ALL is a genetically well-defined disease – but we still don’t know why some children relapse or respond differently to treatment,” Ben explains. “By understanding the interaction networks of genes and proteins, and layering this with mutational data, we might work out how to stratify patients to make treatment more effective or identify new therapeutics for children who relapse.”

These integrated layers of data will be presented as an interactive web dashboard, so that anyone can use the team’s code to interrogate the data, such as looking at their own genes of interest or hoping to unlock new insights for other genetically well-defined conditions. This dashboard approach was inspired by a similar output on a recent metabolism study – which, “full disclosure, was the suggestion of the reviewer, but made so much sense when we saw it in action,” he says.

FAIR (findable, accessible, interoperable, reusable) data principles are an important part of all Data4CYP data sharing plans. But for Ben, the most exciting thing is when open access combines with meaningful collaboration to spark new avenues of exploration. “Data doesn’t exist as an island. It should be interrogated and challenged from multiple angles to really make the most of it,” he describes. “That’s what really leads to transformative research.”

“When Burkitt lymphoma comes back, there’s a complete change in its behaviour,” Simon describes. “It switches from a disease we’re very good at treating, to one we’re extremely poor at treating.” With largely ineffective second-line therapies, the community is working hard to provide early phase clinical trial opportunities for those children. The recently opened Glo-BNHL trial is a great example.

Simon’s research is focused on understanding the fundamental cause of that switch, which could identify new therapeutics for those whose cancer has returned. It’s only recently that advances in technology have caught up with our ambitions to make progress against Burkitt lymphoma. “Just three to four years ago, we’d have said it wasn’t possible to stratify patients into molecular subgroups – that it was a pretty homogenous disease,” Simon says.

In 2022, with access to advanced sequencing technologies, the team identified TP53 status as the disease’s first risk stratifier. Now, with all paediatric malignancies eligible for whole genome sequencing, “there’s a lot of rich data available to us”.

The team’s Data4CYP award makes use of high quality, single-cell molecular, functional and transcriptomic data, collected from patient biopsies, the VIVO Biobank and healthy patients. By linking these datasets – and considering how further analyses, such as epigenetics, could be layered in the future – they hope to identify a core set of Burkitt lymphoma-specific pathways to explore further.

“There’s an inordinate amount of information stored in that data. The challenge is knowing how to interrogate it,” he says. “It’s thanks to our bioinformatics colleagues that we’ve identified a way to make sense of that data. There’s real value in stepping out of your comfort zone and looking beyond your field to find the right step forward.”

The team hope this could provide the starting point to bring our understanding of Burkitt lymphoma closer to that of more common childhood cancers, such leukaemia. This could also provide proof-of-principle for what can be achieved through multidisciplinary collaboration and a small number of high-quality samples, to help advance progress in other poorly understood conditions.

“For children who have relapsed, an early phase clinical trial may be the last glimmer of optimism in overcoming their illness,” Simon describes. “It’s paramount we’re deepening our understanding of Burkitt lymphoma, so we’re building those clinical trials on the highest quality preclinical evidence we can generate.”

What’s next?

The eight Data4CYP recipients start their pilot projects this summer, running for two years. As well as funding, they’ll receive a range of wraparound support and form a network that shares their learnings and resources with each other and the wider research community.

“We received fantastic applications for this round of pilot funding. The ideas that came through really demonstrate the impact that could be made for children and young people, if we can harness the power of data,” says Dr Melissa Lewis-Brown, Head of Research Data Strategy at CRUK. “We’re excited about where these awards could take the field – not just the direct outputs of their pilot funding, but their plans for a wider legacy.”

Larger-scale awards are expected to be announced in 18-24 months’ time, though details of what the scope of this call will look like is still under development.