Prof Louis Chesler leads the Paediatric Oncology Experimental Medicine (POEM) Centre at the Institute of Cancer Research (ICR). He tells us about some of the exciting digital innovations that have the potential to transform childhood cancer treatment.
At the POEM Centre, we have two core goals: to deliver transformative research direct to hospitals so that best practice improves, and to educate the next generation of young academic clinicians with the same goal.
There has never been a more promising time for what can potentially be achieved using the complex tools now available. Many of these advances have been inspiring and will transform the way we treat cancer in children. Finding funding for technology-driven clinical research is a challenge and we’re grateful for the charities, organisations and patient advocates who help us.
Using multi-gene testing and liquid biopsies
Technology allows us to probe a small amount of tumour tissue or blood in better detail, showing us the structure and function of DNA and other cells. We potentially can now analyse blood droplets within hours and generate both point-of-care test results for doctors as well as research data. This means we could perform these tests and report results at the same clinic visit.
These results give a detailed view – at the molecular level – of how the cancer is responding to treatment, including anything that may be stopping the treatment from working, as well as alerting doctors to any disease that has spread but can’t yet be picked up by scans. We’re keen to roll out this technology in all children’s cancer hospitals.
We can now rapidly sequence tumour tissue in better detail and return test data back to doctors in time for them to select molecularly-guided drugs for children who relapse with cancer. This sequencing programme was a collaborative success and delivered four molecular-based tests to the NHS. The NHS will use this knowledge to expand whole-genome and high-depth sequencing across the UK, a major success for children with cancer.
Digital pathology and artificial intelligence (AI)
Technology is also transforming the work of pathologists. In the past, pathologists looked at the shape and appearance of normal cells and cancer cells using stained slides, and then added more ‘stains’ to detect specific markers of cancer type. They can now reconstruct the data digitally, integrate it with clinical scans done on the same tumour tissue, and even train a computer to recognise patterns and cancer signatures, or identify drug targets via AI algorithms.
This gives pathologists a visual and functional 3D ‘Google map’ of the cancer material at the level of individual cells. Integrating these systems so that we can almost instantaneously diagnose and quickly start treatment for children with newly-diagnosed cancers is on the horizon.
Assigning mechanistically-guided treatment using AI
Google DeepMind recently announced AlphaFold, an AI algorithm that accurately predicts the 3D structures of proteins.
Proteins are the target molecules that cancer hijacks in normal cells making them turn into cancerous cells. Targeting them using small-molecule ‘next-generation’ cancer drugs, which bind to proteins and silences them, represents perhaps the most significant advance in cancer treatment in the past 25 years. This has led to longer and better quality of life for adult cancer patients and, increasingly, children with cancer. The significant task of manually examining each protein structure to identify docking sites for drugs used to take months or years in the past, but now takes just hours.
Using AI for data sharing, collaboration and learning models
Organising, sharing and analysing data emerging from these technological advances is now even more important. It’s essential that anonymised clinical data on patient outcomes and characteristics is studied together with research data, results of clinical tests and procedures and is shared across all areas to generate the most benefit to patients. It’s still an aspiration that AI could potentially synthesise all this data to identify treatment approaches for individual patients, but it probably isn’t far off!
From Contact magazine issue 93 - December 2021
