Harnessing AI to help the NHS tackle cancer backlogs

Orlando Agrippa, CEO and founder of RwHealth, explains how AI can be key to tackling cancer backlogs.

As COVID-19 becomes better controlled by vaccines, medicines, and other measures, cancer is re-emerging as the UK public’s top health priority.

The effect of the pandemic on cancer treatment has been profound. From patients putting off medical appointments to the ongoing bed shortage, disruptions to cancer diagnosis and treatment continue to threaten the lives and wellbeing of hundreds of thousands of people.

With half of the population expected to be diagnosed with cancer at some point in their lives, it’s crucial to restore pre-pandemic levels of research, prevention, medical and social care as quickly as possible.

Achieving success when health services are still battling with fundamental capacity constraints requires us to harness new technologies aimed at better identifying and treating cancer. So, what are these technologies and how do we support them to deliver the improvements that are so urgently required?

Ensuring early diagnosis and better treatment pathways

Speed is crucial when diagnosing and treating cancer. Early diagnosis leads to more effective therapeutic intervention and, ultimately, better outcomes – and here is where clinicians recognise Artificial Intelligence (AI) as a game-changing technology.

AI algorithms can extract clinically useful knowledge from vast amounts of data from biology, chemistry, pharmacology, structural biology, cellular networks, and clinical annotations. They recognise patterns and identify complex features and characteristics – such as how a disease may progress – that can’t be processed by the human brain. This means a precise diagnosis can be made more quickly, with the subsequent treatment plan easier to define.

Beyond speed and complexity, AI can remove a huge administration burden, freeing up healthcare professionals to dedicate more time to tasks where human judgement is critical. Data from recent clinical trials shows that AI has the potential to reduce the workload of mammography readers by more than 88%, allowing radiologists to examine more images. Ultimately, even once waiting lists have been reduced, we know that the human touch should never be replaced in cancer care. AI gives these healthcare professionals more time for providing care.

Getting the basics rights

At its most sophisticated, AI can help diagnose the rarest of cancers and lead to new treatments. But on an everyday level, the technology can also improve the basic accessibility of the information systems involved in cancer treatment.

Currently, most systems are anything but intuitive, with medical teams grappling with Excel spreadsheets, or forced to use complex software that requires intense, specialised training. By bringing AI-based data analysis to bear on patient records, initial prognoses, and clinical pathways – and making that information available via an easy-to-use interface – it’s possible for patients to be dealt with more quickly and effectively because the healthcare professionals are no longer being hindered by their own tools.

Predicting surges in demand

AI has a significant role in diagnostics and in cutting administrative burden. But there are further benefits in using AI to predict demand and learn what is likely to be needed, and when.

Hospitals that use AI for a more data-driven approach to their operations can also better manage patient volume and bed capacity as well as mitigate potential impacts on the supply chain and other critical areas.

Unlocking the value of datasets

The accuracy of any AI technology always depends on the quality of the data and the way the system is trained.

In the case of healthcare and cancer diagnosis, it’s both quantity and diversity of data which is important. For example, if an image database is limited to white males over the age of 45, then the information gathered by the AI system will only be useful to that demographic. In cancer care, what works for one group of patients may not work for another, and it’s the ability to compare like-for-like cases that will ultimately determine the speed of prognosis and treatment.

The good news is that much of this data is already being recorded by healthcare providers. There are vast quantities of relevant information in GP and hospital databases that are currently only being used discreetly, if at all. More work is needed to extract the data, so that AI can then be deployed to unlock its potential.

The funding imperative

As spending is reviewed and policy makers set budgets for tech provision in healthcare, investing in AI, data extraction tools, and other supportive systems and processes should be top of the priority list if we want to reduce the current backlog of cancer patients.

There are already several AI-based technologies being trialled within the NHS, but it’s what we do next to further its use and hone its effectiveness that will really count in the post-COVID catch-up. If cancer treatment is our number one priority in the UK, then clinicians must have the technology to match it.